Skip navigation

U.S. Dept of Justice-Restrictive Housing in the U.S. Issues, Challenges, and Futures Directions, Nov. 2016

Download original document:
Brief thumbnail
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
U.S. Department of Justice
Office of Justice Programs
National Institute of Justice

Restrictive Housing
in the U.S.
Issues, Challenges, and Future Directions

U.S. Department of Justice
Office of Justice Programs
810 Seventh St. N.W.
Washington, DC 20531
Loretta E. Lynch
Attorney General
Karol V. Mason
Assistant Attorney General
Nancy Rodriguez, Ph.D.
Director, National Institute of Justice

This and other publications and products of the
National Institute of Justice can be found at:
National Institute of Justice
Strengthen Science • Advance Justice
http://www.NIJ.gov
Office of Justice Programs
Innovation • Partnerships • Safer Neighborhoods
http://www.ojp.usdoj.gov

The National Institute of Justice is the research, development and evaluation agency of the U.S.
Department of Justice. NIJ’s mission is to advance scientific research, development and evaluation
to enhance the administration of justice and public safety.
The National Institute of Justice is a component of the Office of Justice Programs, which also
includes the Bureau of Justice Assistance; the Bureau of Justice Statistics; the Office for Victims
of Crime; the Office of Juvenile Justice and Delinquency Prevention; and the Office of Sex
Offender Sentencing, Monitoring, Apprehending, Registering, and Tracking.
Opinions or conclusions expressed in this volume are those of the authors and do not necessarily
reflect the official position or policies of the U.S. Department of Justice.
Photo Source: © Masterfile

Restrictive Housing
in the U.S.
Issues, Challenges, and Future Directions

November 2016
NCJ 250315

CONTENTS
Introduction .................................................................................. iii
About the Authors ........................................................................ ix
Chapter 1: Administrative Segregation In U.S. Prisons................. 1
Chapter 2: The Use of Administrative Segregation and
Its Function in the Institutional Setting ........................................ 49
Chapter 3: The Conditions of Confinement in
Restrictive Housing ..................................................................... 85
Chapter 4: Gang Affiliation and Restrictive
Housing in U.S. Prisons ............................................................ 117
Chapter 5: The Relationship Between Inmate Misconduct,
Institutional Violence, and Administrative Segregation ............. 165
Chapter 6: Mental Health Effects of Restrictive Housing.......... 199
Chapter 7: Critical Research Gaps in Understanding
the Effects of Prolonged Time in Restrictive Housing
on Inmates and the Institutional Environment ........................... 233
Chapter 8: The Effect of Administrative Segregation on
Prison Order and Organizational Culture................................... 297
Chapter 9: Toward an Understanding of
“What Works” in Segregation.................................................... 331
Chapter 10: Restricted Housing and Legal Issues.................... 367

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • i

ii • National Institute of Justice | NIJ.gov

Introduction
Nancy Rodriguez, Ph.D.

F

or almost 50 years, the National Institute of Justice (NIJ) has invested
in high-quality, high-impact research across disciplines to build
knowledge that serves the needs of criminal justice professionals.
Put simply, our mission as the science agency within the Department of
Justice (DOJ) is to solve real-world crime and justice problems.
Given the current environment, it is clear that the work being carried out by
officials in law enforcement, courts, and corrections is changing rapidly. This
makes our work at NIJ even more critical and vital to the criminal justice
community. As Director of NIJ, I’ve had the privilege of outlining and supporting
research in key priority areas. Since arriving at NIJ, my commitment has
remained the same: to make strategic research investments in areas that directly
address the needs of the criminal justice field.
Mass incarceration and its impact on individuals, families, and communities will
continue to be a focus of inquiry as we revisit its role and function in our society.
NIJ is proud to have supported the National Academies’ report, The Growth of
Incarceration in the United States: Exploring Causes and Consequences, which

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • iii

presents the blueprint for expanding an evidence base in areas such as the impact
of incarceration on justice-involved individuals, on their children and families, and
on how the incarceration experience shapes their way of life and re-entry process.
Consistent with this line of inquiry, institutional corrections, and more
specifically restrictive housing and other strategies that facilities use to manage
and control incarcerated individuals, have become a national priority for
President Obama, DOJ, and corrections administrators at the federal, state,
and local levels. Restrictive housing, commonly known as solitary confinement
or administrative segregation, is a common practice in corrections. A recent
national estimate by the Bureau of Justice Statistics reveals that as many as one
in five individuals has spent time in restrictive housing while in jail or prison.
Despite its use throughout facilities nationwide, we lack the scientific evidence
to convey how corrections administrators use this strategy and its impact on
incarcerated individuals, staff, and the organizational climate. While there are
claims that this correctional strategy increases the safety and well-being of staff
and inmates, there is increasing concern about its potential over-use and its
effects on incarcerated individuals, especially those with mental illness.
To launch NIJ’s dedicated strategic investment in this area, we held a two-day
convening on October 22 and 23, 2015, composed of a diverse group of more
than 80 experts from federal, state, and local corrections agencies, advocacy
groups, academia, and research organizations. This group convened to discuss
(1) what we know and don’t know about the inmates who are put into this type
of housing; (2) the relationship between institutional violence and restrictive
housing; (3) issues related to the mental health of inmates, officer and inmate
safety and wellness, civil rights, and safe alternatives to restrictive housing; and
(4) the gaps in data collection efforts and the existing empirical literature.
Throughout the two days, attendees discussed the research gaps in restrictive
housing and debated the multiple policy and practice concerns that currently
exist. NIJ greatly appreciates these experts’ participation as they shared their
individual perspectives and contributed to identifying how best to move forward
in developing restrictive housing policies and practices that are grounded in
science. Certainly, the most comprehensive understanding of restrictive housing
can only come when we consider the various facets that characterize its use
and impact and consider how these issues affect our theoretical and practical
understanding of this correctional practice.

iv • National Institute of Justice | NIJ.gov

With this goal in mind, this volume includes 10 chapters on restrictive housing,
each with a distinct focus and written by leading experts from various disciplines
including criminology, psychology, sociology, and law. The volume represents the
most comprehensive review to date of emerging issues and concerns surrounding
restrictive housing, including the roles that gangs, violence, and mental health
play in the management of individuals in restrictive housing. Most importantly,
readers of this volume will also find a strong focus on the conceptual and
empirical challenges we face in addressing restrictive housing.
One critical conceptual challenge that readers will notice throughout the
volume is the way authors use different, sometimes contradictory, terms to
define and discuss this practice. Some authors use terms such as administrative
segregation and restrictive housing interchangeably, while other authors carefully
differentiate such terms to highlight critical nuances regarding this practice.
As a whole, these chapters offer an innovative perspective for guiding future
research in this area and ensuring that our efforts have a strong scientific
foundation. Individually, the chapters present an in-depth review of the
important features that characterize restrictive housing.
The volume begins with Natasha Frost and Carlos Monteiro presenting a
historical overview of the use and effects of administrative segregation. Frost and
Monteiro discuss how terminology used to describe administrative segregation
policies and practices varies greatly across jurisdictions, making it difficult to
compare and monitor the use and impact of this strategy. The authors point
to how the increased use of solitary confinement, as opposed to more general
segregation, has brought the greatest legal and ethical concerns from the
field, especially as it involves youth, gang members, and the mentally ill. They
outline the challenges in conducting research on restrictive housing and how to
overcome them to conduct research in this space.
The next chapter in the volume, written by Ryan Labrecque, focuses on two
integral elements of administrative segregation: its use and function within
correctional institutions. Labrecque conveys that in order to understand whether
administrative segregation is an effective strategy for reducing crime and violence
and increasing safety, we need to better understand how this practice is used to
manage and control individuals in correctional facilities. His chapter presents the
limitations of the current empirical research and makes recommendations for
future research.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • v

In the following chapter, Holly Foster presents a thorough synthesis and
critique of the literature on the conditions of confinement in restrictive
housing. Foster discusses the role that litigation has played in uncovering and
addressing concerns raised over the conditions of confinement in restrictive
housing. She outlines that additional information about the conditions across
different types of restrictive housing practices (i.e., punitive, administrative,
and protective segregation) will better inform just and humane policies
and practices. Foster discusses how the general population of incarcerated
individuals, their families, and society at large may benefit from further
research in this area given the spillover effects of restrictive housing as well as
individuals’ eventual return to communities.
The volume also includes a chapter by David Pyrooz who places theoretical and
practical focus on an important element of prison life: gangs. Pyrooz outlines
the role that gangs play in prison violence and misconduct, and establishes that
restrictive housing is the most common practice used to remove gang affiliates
from the general population. While various strategies have been implemented to
maintain control and order and develop programming as it relates to gangs and
gang affiliates, Pyrooz notes that there is no evidence that any of these strategies
are effective. The author highlights the role that hunger strikes at Pelican Bay,
the Ashker v. California class-action lawsuit in 2012, and the resulting settlement
in 2015 have played in bringing attention to this important subpopulation of
incarcerated individuals.
In Benjamin Steiner and Calli Cain’s chapter, the focus moves to institutional
violence and misconduct, that is, how acts of violence and misconduct threaten
the safety and order of an institution and how administrative segregation is used
to address institutional violence and misconduct. To date, few scholars have
examined whether administrative segregation is overwhelmingly used to address
violent or at-risk individuals and whether administrative segregation actually
achieves its goal of reducing individuals’ problematic behavior. The authors
present a comprehensive review of the research on the use and behavioral effects
of administrative segregation. Importantly, they highlight the leading theories
used to explain why administrative segregation is used and the most common
elements that predict its use.
Reena Kapoor and Robert Trestman’s chapter continues the discussion by
addressing another important element of restrictive housing, that is, the role
of mental health and the availability of mental health treatment. Kapoor and

vi • National Institute of Justice | NIJ.gov

Trestman present the latest scientific knowledge on the relationship between
mental health and restrictive housing from a variety of disciplines, including
the medical, legal, and social and behavioral sciences. At the core of their review
is establishing whether restrictive housing causes psychological harm. The
authors focus on various behavioral outcomes including suicides, psychiatric
hospitalizations, institutional misconduct, and post-traumatic stress disorder.
The authors also place this work within the context of consensus statements of
major mental health professional organizations and provide insight on future
research directions.
In the next chapter, Daniel Mears and Mark Stafford review research on the
effects of prolonged time in restrictive housing on individuals and on the
institutional environment of correctional systems. The authors argue that any
policy discussion on restrictive housing centers on understanding (1) under
which circumstances it is appropriate to use restrictive housing, (2) whether its
use leads to improved outcomes for individuals and the broader institution, (3)
and when it is inappropriate and produces adverse effects. Mears and Stafford
conclude that the evidence base that establishes the effectiveness of restrictive
housing is very thin. In fact, the lack of consistent findings from existing research
has shown that there may be benefits to its use, including improved behavioral
outcomes, as well as harms such as worsened mental health or increased
recidivism. The authors conclude by identifying prominent gaps in research that
need attention.
The chapter by Jody Sundt applies an organizational perspective to examining the
effects and the use of administrative segregation. The author presents research on
the effectiveness of administrative segregation as a management strategy as well
as its effect on prison organizational culture. Sundt outlines a research agenda
to develop knowledge around how to use this correctional strategy to maintain
secure and humane institutions. The organizational and management framework
provides an important theoretical perspective to better understanding the effects
of this practice.
Paula Smith’s chapter discusses the availability of programming and re-entry­
focused services in restrictive housing. The author presents the challenge in
balancing the need to provide effective treatment to disruptive individuals in
crowded settings with maintaining safety and control within the institution.
Restrictive housing is presented as one of several options in a continuum
of placement options within institutions. In the end, identifying the most

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • vii

effective ways to ensure that programming and a long-term treatment plan for
all individuals who are incarcerated is vital. Guidelines and best practices for
correctional agencies, such as the one presented in the DOJ’s 2016 Report and
Recommendations Concerning the Use of Restrictive Housing, provide a start.
This volume concludes with Fred Cohen’s legal analysis of the use of restrictive
housing in correctional facilities throughout the country. The analysis includes
discussions on the due process implications of disciplinary segregation,
administrative segregation, and protective custody, as well as the Eighth and 14th
Amendments and how they apply to the use of restrictive housing. Cohen also
presents the legal implications of the conditions and duration of confinement,
youth in restrictive housing, and the confinement of the mentally ill. This
particular chapter thoroughly presents the significant court decisions in this area
and outlines the impact they have had on the current practice.
In sum, these chapters provide a comprehensive look at what we currently know
about the use of restrictive housing in U.S. correctional facilities and the effects
of this practice on incarcerated individuals, corrections staff, and the institution
as a whole. Just as important is what this work tells us about what we do not know
about this practice, its effects, and potential alternatives. As a collection, these
chapters enable us to develop a future research agenda to further expand our
knowledge of this important correctional strategy and identify evidence-based
solutions to the challenges currently presented by restrictive housing.
A volume of this breadth would not be possible without the concerted effort of
many people. I want to thank the attendees of NIJ’s two-day restrictive housing
convening. These experts were generous with their time and knowledge, and
they helped build the foundation of this effort. My sincere appreciation goes
to the contributing authors because of their insight, expertise, and dedication
to this important issue. Finally, this volume would not exist without the steady
guidance of Marie Garcia. This work was her inspiration, and her hard work and
dedication made it possible.
NIJ remains committed to investing in high-quality, multidisciplinary science to
address the challenges faced by policymakers and criminal justice professionals
working within corrections. I hope readers find this volume both informative
and helpful in providing potential ways forward on the use and impact of
restrictive housing.

viii • National Institute of Justice | NIJ.gov

About the Authors
Marie Garcia
Marie Garcia, Ph.D., the editor of this volume, is a social science analyst in the
Justice Systems Research Division at the National Institute of Justice (NIJ). Dr.
Garcia earned her doctorate in criminal justice from Temple University, where
her research focused on the impact of disorganization and crime on levels of
trust in Philadelphia, communities and crime, and institutional corrections. Dr.
Garcia’s NIJ research portfolio has focused on justice systems, special offender
populations, offender re-entry, and institutional and community corrections.
Prior to NIJ, Dr. Garcia worked as a mental health counselor at Valley State
Prison for Women in Chowchilla, California. Her writing has appeared in Justice
Quarterly, Criminology & Public Policy, The Journal of Research on Crime and
Delinquency, The Journal of Family and Marriage, and Criminology.

Calli M. Cain
Calli M. Cain is a graduate student in the School of Criminology and Criminal
Justice at the University of Nebraska at Omaha. Her primary research interests
include women and crime, with an emphasis on human trafficking, corrections,
and juvenile delinquency.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • ix

Fred Cohen
Fred Cohen is a graduate of Yale Law School and professor emeritus at the State
University of New York at Albany, School of Criminal Justice, for which he is one
of the original founders. He is the executive editor of the Correctional Law Reporter
and the Correctional Mental Health Reporter, and he is the author of A Practical
Guide to Correctional Mental Health and the Law. Mr. Cohen served as a reporter
to the American Bar Association (ABA)–Institute of Judicial Administration’s
Juvenile Justice Standards Project and assisted in drafting the current ABA
Standards for Corrections. He provided testimony on solitary confinement at
U.S. Senate hearings in 2012 and 2014 and has testified before other federal
commissions on solitary confinement. Mr. Cohen is plaintiff ’s expert-monitor in
Rudisill v. Ryan (a civil rights case in Arizona), overseeing racial and ethnic prison
integration, and he works with the Jesuit Social Research Institute on bringing
human rights norms and training to private correction facilities.

Holly Foster
Holly Foster, Ph.D., is an associate professor in the Department of Sociology
at Texas A&M University. Her research focuses on the effects of parental
incarceration on children, women’s imprisonment, and children’s exposure to
violence. Dr. Foster’s research interests include crime and deviance, the life
course, social inequality, families, and children and youth. She teaches courses in
criminology and gender and crime, as well as a graduate seminar on crime and
the life course.

Natasha A. Frost
Natasha A. Frost, Ph.D., is associate dean of graduate studies in the College
of Social Sciences and Humanities, and is an associate professor in the School
of Criminology and Criminal Justice at Northeastern University in Boston,
Massachusetts. Dr. Frost holds a Ph.D. in criminal justice from the City
University of New York’s Graduate Center. Her most recent research on mass
incarceration has focused on the effects of incarceration on individuals, families,
and communities, and on the well-being of those who work in correctional
facilities. She is the co-author of The Punishment Imperative: The Rise and Failure
of Mass Incarceration in America.

x • National Institute of Justice | NIJ.gov

Reena Kapoor
Reena Kapoor, M.D., is associate professor of psychiatry at Yale School of
Medicine’s Law and Psychiatry Division, and serves as associate program
director for the Yale Forensic Psychiatry Fellowship program. She earned her
M.D. at Northwestern University’s Feinberg School of Medicine, completed her
residency in psychiatry at Harvard Medical School, and was the recipient of a
forensic psychiatry fellowship at Yale. Dr. Kapoor’s clinical work and scholarship
is focused on the intersection between serious mental illness and the criminal
justice system, and she has expertise in correctional psychiatry, community
treatment of persons with criminal justice involvement, and management of
problematic sexual behaviors. She has lectured widely on forensic psychiatry
and holds leadership positions in several professional organizations, including
serving as Connecticut’s representative to the American Psychiatric Association
Assembly and president of the Connecticut Psychiatric Society. Dr. Kapoor
co-founded the Community Forensics committee of the American Academy of
Psychiatry and the Law and is president-elect of the International Association for
Forensic Psychotherapy.

Ryan M. Labrecque
Ryan M. Labrecque, Ph.D., is an assistant professor in the Criminology and
Criminal Justice Department at Portland State University. He earned his Ph.D.
in criminal justice at the University of Cincinnati and was the recipient of a 2014
Graduate Research Fellowship award from NIJ for his dissertation, “The Effect of
Solitary Confinement on Institutional Misconduct: A Longitudinal Evaluation.”
His research interests focus on the evaluation of correctional interventions,
effects of prison life, development of risk and needs assessments for community
and institutional corrections settings, and transfer of knowledge to practitioners
and policymakers. Dr. Labrecque has published a number of articles, book
chapters, and conference presentations. Prior to undertaking his doctoral studies,
Ryan was a probation and parole officer in Maine.

Daniel P. Mears
Daniel P. Mears, Ph.D., is the Mark C. Stafford Professor of Criminology at
Florida State University’s College of Criminology and Criminal Justice. He

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • xi

conducts basic and applied research on crime, criminal justice, and corrections,
including studies funded by foundations and state and federal agencies. His
work has appeared in a wide variety of journals and in American Criminal Justice
Policy, which won the Academy of Criminal Justice Sciences’ outstanding book
award. He also co-authored Prisoner Reentry in the Era of Mass Incarceration.

Carlos E. Monteiro
Carlos E. Monteiro, Ph.D., is a senior research associate at the Institute of
Race and Justice at Northeastern University. Dr. Monteiro earned his Ph.D.
in criminology and justice policy at Northeastern University. With combined
expertise in education and criminal justice policy, Dr. Monteiro’s research
interests center on the factors affecting access to, and quality of, education for
young adults of color. His scholarly interests are tied to race, ethnicity, and
educational access, and, in particular, how those factors interact to produce
disparate outcomes across the criminal justice system.

David C. Pyrooz
David C. Pyrooz, Ph.D., earned his doctorate in criminology and criminal
justice at Arizona State University. He is the recipient of a 2011 graduate research
fellowship award from NIJ and is the author of Confronting Gangs: Crime and
Community. His current research projects involve tracking the patterning of gang
membership across the life course, meta-analyzing the consequences of gang
membership, and understanding the victim-offender overlap.

Paula Smith
Paula Smith, Ph.D., is an associate professor at the University of Cincinnati
School of Criminal Justice. She received her doctorate in psychology from the
University of New Brunswick, Saint John. Dr. Smith’s research interests include
meta-analysis, assessment of offender treatment and deterrence programs,
development of risk and need assessments for clinicians and managers in prisons
and community corrections, effects of prison life, and the transfer of knowledge
to practitioners and policymakers. She has published numerous articles, book
chapters, and conference presentations on the above topics and has directed
numerous federal- and state-funded research projects — including studies

xii • National Institute of Justice | NIJ.gov

of prisons, community-based correctional programs, juvenile drug courts,
probation and parole departments, and mental health services.

Benjamin Steiner
Benjamin Steiner, Ph.D., is an associate professor of criminology and criminal
justice at the University of Nebraska at Omaha. He earned his doctorate at the
University of Cincinnati. Dr. Steiner’s research interests focus on issues related
to juvenile justice, and institutional and community corrections. Dr. Steiner has
published more than 50 journal articles and book entries related to these topics
and is a co-author of Criminal Justice Case Briefs: Significant Cases in Juvenile
Justice. He has consulted with the Iowa Department of Corrections, the Kentucky
Department of Corrections, the Nebraska Department of Correctional Services, the
Ohio Department of Rehabilitation and Correction, the Bureau of Justice Statistics,
and NIJ. Dr. Steiner has testified on evidence-based practices in corrections and the
deterrent effects of direct file transfer laws to various government entities.

Jody L. Sundt
Jody L. Sundt is an assistant professor of criminal justice at Indiana University.
She earned her Ph.D. at the University of Cincinnati. Dr. Sundt has published
in the areas of correctional policy, religion in prison, offender change, public
attitudes toward punishment, and white collar crime. Her work has appeared in
several journals, including Criminology, Justice Quarterly, Crime and Delinquency,
Criminal Justice and Behavior, the Journal of Criminal Justice, and The Prison
Journal. Her current research focuses on the occupational experiences of
correctional officers, the effectiveness of supermax prisons, and religious coping
and attribution.

Robert L. Trestman
Robert L. Trestman, M.D., Ph.D., is a professor of medicine, psychiatry, and
nursing and is executive director of Correctional Managed Health Care at
UConn Health. Dr. Trestman earned his Ph.D. in psychology and his M.D. at
the University of Tennessee, and trained in psychiatry and neurobiology at the
Icahn School of Medicine at Mount Sinai. He has served as clinical vice-chair
of psychiatry at the Mount Sinai Medical Center in New York and at UConn

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • xiii

Health. Dr. Trestman has studied the neurobiology and treatment of people with
severe mood and personality disorders, and conducts translational research on
correctional health. He has published more than 100 articles and book chapters;
serves as consultant to the National Institute of Mental Health, the Substance
Abuse and Mental Health Services Administration, and several branches of the
U.S. Department of Justice; and is chair of the American Academy of Psychiatry
and the Law research committee.

xiv • National Institute of Justice | NIJ.gov

CHAPTER 1

Administrative Segregation
In U.S. Prisons
Natasha A. Frost, Ph.D., & Carlos E. Monteiro, Ph.D.
Northeastern University

Introduction

O

n September 1, 2015, newspapers across the country announced that
a settlement agreement had been reached between the California
Department of Corrections and Rehabilitation (CDCR) and
inmates incarcerated at Pelican Bay State Prison, one of the most well-known
supermaximum (supermax) security facilities in the country (St. John, 2015).
The settlement agreement, which should result in the return of close to 2,000
inmates from supermax confinement back to the general prison population, is
expected to end CDCR’s practices of indefinitely housing inmates in supermax
confinement and of routinely incarcerating those with suspected gang
affiliations in solitary confinement. Although California’s practice of confining
gang members in administrative segregation is certainly not the norm around
the country, long-term segregation in restrictive housing is more common.
The California settlement was announced amid a more general and growing
concern about the practice of solitary or near-solitary confinement through
administrative segregation.
In a July 2015 speech before the National Association for the Advancement of
Colored People (NAACP), President Barack Obama questioned the practice of

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 1

solitary confinement by calling for a Department of Justice investigation into its
use across the United States:
I’ve asked my Attorney General to start a review of the overuse of solitary
confinement across American prisons. The social science shows that an
environment like that is often more likely to make inmates more alienated, more
hostile, potentially more violent. Do we really think it makes sense to lock so many
people alone in tiny cells for 23 hours a day, sometimes for months or even years
at a time? That is not going to make us safer. That’s not going to make us stronger.
And if those individuals are ultimately released, how are they ever going to adapt?
It’s not smart. (White House Office of the Press Secretary, 2015)1
President Obama is not alone in his reservations about the practice. Supreme
Court Justice Anthony Kennedy has repeatedly made clear his concern about
solitary confinement across several venues, including by using largely unrelated
cases to question the policies of long-term solitary confinement (Liptak, 2015).
In May 2015, the United Nations (U.N.) passed the Mandela Rules, which
represent the first modification to the U.N.’s standards for the treatment of
prisoners in 60 years (United Nations, 2015a). Rule 43 of the Mandela Rules
prohibits both indefinite solitary confinement and prolonged solitary confinement
(defined as lasting more than 15 days) (United Nations, 2015b). Human Rights
Watch, Amnesty International, and the American Civil Liberties Union (ACLU)
have each published reports condemning the use of solitary confinement for both
juvenile and adult correctional populations (American Civil Liberties Union,
2014; Amnesty International, 2012; Human Rights Watch, 2000; Human Rights
Watch & American Civil Liberties Union, 2012). Individual state ACLU chapters
have published fairly scathing critiques of more localized practices, for example,
in Colorado and Texas (Butler & Simpson, 2015; Wallace, 2013). The perspective
of these advocacy organizations is clear and unapologetic: They seek an end
to the practice of solitary confinement in juvenile correctional settings and
extensive restrictions on its use among adult correctional populations.
There is growing concern across the political spectrum about the efficacy
and utility of administrative segregation practices — particularly those that
involve extended solitary confinement — and growing support for finding
ways to safely reduce its use across correctional systems. In 2006, the bipartisan
Commission on Safety and Abuse in America’s Prisons, co-chaired by thenchief judge of the Third Circuit Court of Appeals, John J. Gibbons, and the
former U.S. Attorney General, Nicholas de B. Katzenbach, noted that the rapid
1

Although President Obama has called for a Department of Justice review of solitary confinement practices
nationally, two recent and substantial inquiries were made into the federal use of administrative segregation
(Baker & Goode, 2015). In May 2013, the U.S. Government Accountability Office issued its report on the use
of segregated confinement across the federal prison system (United States Government Accountability Office,
2013). After the publication of that report, the CNA Institute for Public Research, with the cooperation of the
Federal Bureau of Prisons, conducted an independent assessment (McGinnis et al., 2014).

2 • National Institute of Justice | NIJ.gov

increase in the use of solitary confinement across the country had outpaced the
remarkable growth in overall correctional populations (Gibbons & Katzenbach,
2006). The Commission deemed solitary confinement both expensive and
counterproductive, and recommended limiting its use. After the publication of
the Commission’s Confronting Confinement report, research organizations also
turned their attention to solitary confinement. Researchers at the Vera Institute
of Justice recently published a report on solitary confinement, identifying
what they describe as 10 common misperceptions about solitary confinement
(Shames, Wilcox, & Subramanian, 2015). Among those misconceptions are the
common beliefs that segregated housing deters violence and misbehavior and
that segregation helps keep prisons and jails safer. The Vera Institute of Justice
(Shames, Wilcox, & Subramanian, 2015) has also launched the Segregation
Reduction Project, partnering with four states (Illinois, Maryland, New Mexico,
and Pennsylvania) to assess the criteria for placement in segregation with the
explicit goal of reducing the use of segregation across those states. Several other
states, including Arizona, California, Colorado, Indiana, Michigan, Nebraska,
New Jersey, New York, Ohio, and Wisconsin, have already begun working to
reduce the number of inmates in administrative segregation, with more states
passing reforms related to the use of solitary confinement in 2014 than in the
previous 16 years combined (Hager & Rich, 2014).
Although the spotlight seems to be shining especially brightly at this moment,
the practice of solitary confinement has a long and storied history in corrections.
Some of the earliest American correctional facilities — the early Quaker-inspired
penitentiaries in Pennsylvania — were built on a model of extended solitary
confinement intended to bring about penitence (Rothman, 1971/1990). Although
the “Pennsylvania model” was abandoned relatively quickly in favor of a model
based on the more congregate style of confinement that is still prominent, the
use of solitary confinement — usually for behavioral control and management
— never went away. All correctional systems (including those for men, women,
and juveniles) have cells or units and, in some cases, entire facilities designed
to isolate some inmates in more restrictive housing units for administrative
purposes. Segregated confinement in restrictive housing units is sometimes
solitary. Whether they involve complete solitary confinement or not, restrictive
housing units are intended to offer a more secure housing alternative for those
who cannot be safe toward others, kept safe, or adequately controlled in the
traditional congregate correctional setting.
Within correctional contexts, the terms used to describe segregation policies and
practices vary greatly across jurisdictions. Although they represent conceptually
distinct practices, it is difficult to separate the literature on disciplinary
segregation from the literature on administrative segregation because researchers
have tended to study solitary confinement without carefully distinguishing
the various types of restrictive housing units. As a result, this paper does not
use “administrative segregation” as an umbrella term, instead opting for either
segregation or segregation in restricted housing. Where possible, the paper

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 3

distinguishes between solitary confinement through disciplinary segregation and
solitary confinement through administrative segregation. The former refers to
short-term confinement after a specific infraction, and the latter refers to long­
term classification to a supermax unit or facility within a correctional system.
Most of the paper’s early discussion focuses on administrative segregation
(rather than disciplinary segregation), but when the paper begins discussing the
empirical research, it refers more broadly to the practice of solitary confinement
(whether in disciplinary or administrative segregation units). Although not all
units and facilities used for disciplinary and administrative segregation follow
a strict regimen of solitary confinement, this paper primarily describes the
empirical research that has been conducted in settings that do follow such a
regimen, as it is clearly solitary confinement that most troubles those who have
expressed grave concerns about correctional segregation policies.

Brief History of Administrative Segregation
Developed as a strategy for separating problematic inmates from the general
population, administrative segregation is one of two dominant behavioralcontrol models used by correctional administrators to address any number
of challenges that accompanied the rapid growth of prison populations
(Hershberger, 1998; Riveland, 1999). The dispersion and consolidation models
represent contrasting strategies for handling inmates who are perceived to pose
significant security challenges to the correctional system. Administrators using
the dispersion model manage inmates through a divide-and-conquer approach,
attempting to limit the impact of problematic inmates by dispersing them
throughout the correctional system. Dispersion avoids the concentration of
inmates classified as disruptive or unruly in one location, thereby allowing staff
to control disorder more effectively throughout the system (Pizarro & Stenius,
2004). Conversely, consolidation, an approach more aligned with contemporary
administrative segregation practices, consolidates disruptive or unruly inmates
in highly restrictive settings. The presumed benefits of the consolidation model
lie in its efficiency in directing resources toward a central location, be it a unit or
stand-alone facility that can house individuals and groups identified as a threat to
institutional security (Hershberger, 1998).
Historical accounts indicate that correctional administrators have alternated
between these two approaches. The federal prison at Alcatraz, for example,
operated under the consolidation model, housing some of America’s most
notorious and disruptive offenders for most of the early 1900s (Pizarro & Stenius,
2004). When Alcatraz shut its doors in 1963 with no viable alternative location
for consolidation, the inmates from Alcatraz were dispersed throughout the
federal prison system. The federal system’s return to the dispersion model,
however, was short-lived because of increasing violence within the system
between 1970 and 1980. After the rate of assaults escalated throughout the
Federal Bureau of Prisons (BOP) facilities, the federal prison in Marion,

4 • National Institute of Justice | NIJ.gov

Illinois, was modified for increased security and became the first level-6
supermax facility in the United States (Ward & Werlich, 2003). Intended as a
replacement for Alcatraz, the high-security prison at Marion gradually became
the preferred facility not only for BOP’s most problematic inmates but also for
inmates perceived to represent a grave threat to institutional security across
state correctional systems. In other words, Marion rapidly became the go-to
institution for housing the “worst of the worst” (Richards, 2008).
Although the increase in prison violence troubled correctional administrators,
correctional historians often point to the 1983 killing of two correctional
officers at Marion as the trigger for the revival of total lockdown units and
facilities (King, 1999; Pizarro & Stenius, 2004). In the immediate aftermath of
those killings, Marion administrators rapidly reintroduced highly restrictive
procedures, beginning with the immediate removal of inmates’ personal
property from individual cells, followed by the placement of severe restrictions
on inmates’ movements within the prison, the use of handcuffs whenever an
inmate was not in the cell area, and increased use of solitary confinement (King,
Steiner, & Breach, 2008). Although the conditions at Marion sparked immediate
pushback from prisoner rights groups, the use of control units received judicial
endorsement when, in Bruscino v. Carlson, a federal court opined that BOP had
not violated inmates’ constitutional rights (Olivero & Roberts, 1987). Eventually,
the U.S. Supreme Court’s denial of the petition for writ of certiorari in Bruscino
only strengthened the sense among correctional administrators that the courts
had formally sanctioned the use and expansion of control units similar to those
at Marion. In the aftermath of these court decisions, supermax-style facilities,
such as the security housing unit at Pelican Bay that opened in 1989, became
models for correctional jurisdictions across the country (Bosworth, 2004; King,
1991; Romano, 1996).

Contemporary Use of Administrative Segregation
Since the 1980s, entire facilities in both the state and federal correctional
systems have been constructed with isolation and segregation as their central
purposes. Commonly referred to as supermax facilities, these units offer
enhanced security and control, allowing for only minimal contact between
inmates and staff. Where construction of a new facility was either not necessary
or not feasible, entire sections of existing facilities were repurposed to segregate
the inmates deemed the worst of the worst (Butler, Griffin, & Johnson, 2013).
Although the clear general consensus is that supermax facilities are designed
to isolate offenders who require the highest and most restrictive security
classification, there is no universally accepted definition (Fellner & Mariner,
1997; Henningsen, Johnson, & Wells, 1999; Riveland, 1999).
The lack of definitional consensus has made collecting information on this type
of custody (including data on prevalence, goals, objectives, and associated effects)
difficult. The National Institute of Corrections (NIC) attempted to provide clarity
Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 5

about the practice of supermax incarceration through its 1997 national survey
of state departments of corrections that focused on supermax-style housing
(Riveland, 1999). Although the survey results from all 50 state departments
of corrections offered a rich source of information by identifying more than
55 functioning supermax facilities or units in 1997, it also demonstrated the
significant variation across jurisdictions. Some facilities, for example, were
stand-alones, whereas others were sections or units within existing correctional
facilities that had been repurposed and retrofitted to meet the strict control needs
of the supermax model.
Moreover, across correctional systems, units classified as supermax might be
referred to as administrative maximum units, administrative segregation units,
special housing units, secure housing units, segregation units, isolation units,
close custody units, control units, management units, and adjustment centers, to
name but a few (Kupers et al., 2009; Naday, Freilich, & Mellow, 2008; National
Institute of Corrections, 1997). A national survey of state wardens (Mears &
Castro, 2006) found that more than 95 percent of the state prison wardens
surveyed agreed that the following modified definition put forth by the NIC was
accurate: “[A] supermax is a stand-alone unit or part of another facility and is
designated for violent or disruptive inmates. It typically involves up to 23-hour­
per-day, single-cell confinement for an indefinite period of time. Inmates in
supermax housing have minimal contact with staff and other inmates” (p. 40).
Supermax units and facilities might house inmates being segregated for both
administrative and disciplinary purposes (Browne, Cambier, & Agha, 2011).
The lack of definitional clarity gives rise to the first of many challenges
to conducting research (or interpreting existing research) on segregation
and solitary confinement in correctional contexts: the use of overlapping
terminologies for what are sometimes distinct correctional practices. As an
umbrella term, “segregation” refers to placement in restricted housing for
disciplinary segregation or protective custody and temporary or long-term
supermax housing (McGinnis et al., 2014, 2008b; Shames et al., 2015). The
primary purpose of this practice is to separate and isolate an inmate or certain
groups of inmates from the general population primarily for security and safety
within the facility or across the correctional system.
There are at least three distinct types of segregation: administrative segregation,
disciplinary segregation, and protective custody (see Shames et al., 2015, p. 4).
Some inmates are segregated because they are identified as being at high risk
for victimization. Inmates in protective custody are segregated for their own
protection — their placement in segregation is sometimes voluntary. Solitary
confinement for a specified period to punish misbehavior is generally referred
to as disciplinary or punitive segregation. Disciplinary segregation is typically
imposed as a sanction following a disciplinary hearing related to a specific
instance of misconduct. It is crucial to note that disciplinary segregation is a
form of punishment, so inmates subjected to it are afforded due-process rights
(O’Keefe, 2008). Administrative segregation is used to separate inmates deemed
6 • National Institute of Justice | NIJ.gov

to pose a significant threat to institutional security from the general population.
Inmates are often classified or transferred to administrative segregation based
on patterns of disruptive behavior, security threat group identifications, or
designation as high-risk inmates.
In a recent review of segregation policies, Metcalf and colleagues (2013) noted
that jurisdictions tend to invoke the safety of inmates and staff as well as overall
institutional security as the primary criteria for placement in administrative
segregation. In addition to safety and security, many states included more specific
placement criteria (typically tied to either the offense that triggered the inmate’s
initial incarceration or the accumulation of disciplinary infractions). Unlike
disciplinary segregation, which is time-limited, length of stay in administrative
segregation is typically indefinite and imposed largely at the discretion of
correctional administrators. Although placement into this often much longerterm form of segregation does not trigger the same due-process rights and
protections as does placement in disciplinary segregation (O’Keefe, 2008),
according to a recent review of correctional policies across the United States,
almost all correctional systems have procedures for review of placements into
administrative segregation (Metcalf et al., 2013). With several pathways into the
various restrictive housing units and different trajectories once there, one could
anticipate substantial variations not only in prevalence but also in psychological
and behavioral effects on inmates across restrictive housing types.

Solitary Confinement Versus Administrative Segregation
Isolation through solitary confinement is prevalent across both administrative
and disciplinary segregation. Solitary confinement practices vary across
correctional systems, but a defining feature of current practice is the isolation
of inmates for 22-24 hours per day in small cells, with minimal contact with
others, in areas of the facility designed for the purpose of restricting inmates’
movement. Other distinct features include reduced natural light; limited artificial
lighting; little or no access to programming, classes, reading materials, or radio
and television; and restrictions on visits from friends and family (American Civil
Liberties Union, 2014). Although researchers tend to be most interested in this
type of custody, correctional administrators rarely refer to solitary confinement
in any context, perhaps seeking to avoid the controversy the phrase often
invokes.2 Scholars studying psychological and behavioral effects focus almost
exclusively on solitary confinement, often with little regard for the varying
contexts in which it occurs.

2

A good example is BOP, which has refused to acknowledge that it uses solitary confinement. The recent U.S.
Government Accounting Office report on the federal use of administrative segregation included the following note:
“According to BOP officials, the BOP does not hold anyone in solitary confinement because BOP staff frequently
visit inmates held in single-bunked cells alone” (United States Government Accountability Office, 2013, p. 12).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 7

Prevalence of Administrative Segregation
Given the paucity of data on solitary confinement through administrative or
disciplinary segregation, both types are difficult to quantify with any precision.
Estimates from studies in individual states may not be representative of trends
more generally. Prevalence estimates vary widely across sources, and many of
those estimates are dated. Early estimates suggested that somewhere between
1 percent and 3 percent of the total correctional population was incarcerated
in highly restrictive administrative segregation units (King, 1999; O’Keefe
et al., 2011). These data, however, have been criticized as underestimates, as
prison systems have been accused of failing to report or of underreporting as a
strategy for avoiding the controversy associated with solitary confinement and
administrative segregation policies (Naday, Freilich, & Mellow, 2008).3
As a result of several recent comprehensive reports on the use of administrative
segregation in federal prisons, some of our most current and best estimates
of prevalence and cost come from analyses of its use in the federal system
(McGinnis et al., 2014; United States Government Accountability Office, 2013).
At the federal level, administrative segregation covers three distinct types of
restrictive housing (see Appendix Table A1) which BOP refers to as special
housing units (SHUs), special management units (SMUs), and administrative
maximums (ADXs). All three types of administrative segregation share the same
purpose: to separate inmates identified for their disruptive or violent behavior
in a controlled setting that emphasizes the safety, security, and orderly operation
of BOP facilities. In the first report, the U.S. Government Accounting Office
(GAO) (2013) reviewed the use of administrative segregation across the federal
system. According to the report, 7 percent of all federal inmates in BOP facilities
are held in administrative segregation. The report notes that most segregated
inmates in the federal system were in SHUs (81 percent), slightly less than 2,000
inmates (approximately 16 percent) were in SMUs, and the ADX facility housed
approximately 450 inmates. The GAO report found that per-capita cost estimates
for housing inmates in segregation were higher than in nonsegregated or general
population housing. Specifically, the GAO report found that, for fiscal year
2012, the total cost of housing 1,987 inmates in SMUs was $87 million (it would
have cost approximately $42 million to house those same inmates in a mediumsecurity facility; $50 million in a high-security facility).4
In addition to prevalence estimates from reviews of the federal system, staff of
the Liman Program and the Association of State Correctional Administrators
(ASCA) recently collaborated on a survey of correctional systems directors
3

The BOP again serves as an example, reporting to the American Correctional Association in 2008 that it had no
inmates in administrative segregation (cited in O’Keefe et al., 2011).

4

After the release of the critical GAO report, BOP commissioned an independent evaluation of its operation
of administrative segregated housing and sought strategies for improving policies, operations, and services
(McGinnis et al., 2014).

8 • National Institute of Justice | NIJ.gov

across 46 jurisdictions (Liman Program & Association of State Correctional
Administrators, 2015). In fall 2014, the survey specifically asked jurisdictions
to account for the number of prisoners held in any form of segregated housing,
including disciplinary segregation, protective custody, and administrative
segregation. The recent report to the Department of Justice, Time-In-Cell, offers
a detailed, current assessment of the prevalence of the use of restrictive housing
across the country. Only 34 of the 46 responding jurisdictions provided counts
of inmates across all forms of restricted housing, which included approximately
66,000 people (Appendix Table A2).5
The Time-in-Cell report also reported a relatively stable trend in the numbers of
inmates housed in administrative segregation, noting an average decrease of less
than 1 percent (0.59) in the percentage of prisoners in administrative segregation
between 2011 and 2014. With regard to time spent in administrative segregation,
32 jurisdictions reported no fixed minimum period, and 42 jurisdictions reported
no maximum duration after which prisoners must be released into the general
population. The survey also focused on the number of consecutive days spent in
administrative segregation. Of the 24 jurisdictions reporting system wide data on
length of stay, 11 reported that most prisoners held in administrative segregation
were there for fewer than 90 days (Liman Program & Association of State
Correctional Administrators, 2015).
A recently published Bureau of Justice Statistics (BJS) special report also
provided more recent estimates of the extent of the use of restrictive housing
across U.S. prisons and jails (Beck, 2015). Using 2011-2012 data from the
National Inmate Survey, Beck (2015) provided a detailed accounting of the
housing status of inmates, including measures of time spent in restrictive
housing during their past 12 months. Beck notes that roughly 20 percent of
prison inmates and 18 percent of jail inmates had spent time in restrictive
housing during the past 12 months and that restrictive housing rates correlate
with common predictors of prison misconduct (such as age and prior criminal
history). One of the most important findings in this special report is that facilities
with higher rates of restrictive housing tended to have higher levels of facility
disorder and a larger proportion of vulnerable inmates (inmates with mental
health problems; lesbian, gay, and bisexual inmates; and younger inmates) in
their populations.

Issues Related to the Use of Solitary Confinement
Some of the most controversial issues related to the use of solitary confinement
involve its use among special populations, most notably young people but also
suspected or known security threat group members and mentally ill inmates.
5

It should be noted that California, one of the largest prison systems in the country, did not respond and, therefore,
is not included in the Liman Program/ASCA report.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 9

Juveniles and Solitary Confinement
If research on solitary confinement through administrative segregation among
adult correctional populations can best be described as scarce, the research on
the use of solitary confinement among juvenile correctional populations is almost
nonexistent. Moreover, the limited data we have on the solitary confinement of
young persons come from just a handful of sources. Within juvenile corrections
there has been even less empirical research, but more determined efforts have
been made to end the practice of punitive isolation for young inmates. In juvenile
corrections, sometimes a distinction is drawn between punitive confinement and
nonpunitive solitary confinement, where the latter is described as confinement
for the protection and safety of others (Weiss, Kraner, & Fisch, 2013). Most
of what is known about juvenile solitary confinement comes from either the
national Survey of Youth in Residential Placement (SYRP) or reports authored
by advocacy organizations such as Human Rights Watch and the ACLU (Human
Rights Watch & American Civil Liberties Union, 2012; Sedlak & McPherson,
2010). In the recent SYRP, almost one-third of all young persons in custody
reported having spent time in solitary confinement, with more than half of those
reporting having spent more than 24 hours there (Sedlak & McPherson, 2010).
In recent years, legislatures have been particularly active in the issue of
solitary confinement of young persons (Therolf, 2015). West Virginia became
the first state to ban the solitary confinement of youth in custody in 1998.
Alaska, Colorado, Mississippi, and West Virginia followed with bans on
solitary confinement of young persons in 2012; Nevada and Oklahoma passed
restrictions on the solitary confinement of young persons in 2013; and a slew of
states, including New York, banned solitary confinement of young persons in
2014 (Hager & Rich, 2014). After several scathing critiques of the treatment of
inmates in New York City’s Rikers Island jail in early 2015, the New York City
Department of Corrections announced, in one of the most sweeping decisions,
that it would no longer allow solitary confinement of anyone 21 years of age or
younger (Winerip & Schwirtz, 2015).

Solitary Confinement to Control Gangs
In some jurisdictions, inmates have been isolated in administrative segregation
simply because of a suspected or known gang affiliation. As noted in the
Introduction, a settlement agreement between a group of inmates incarcerated at
Pelican Bay and CDCR is expected to bring about the gradual end to this practice
in California (St. John, 2015). Although it is based on evidence suggesting that
prison gangs bear responsibility for much of the prison violence experienced
in some correctional systems (Griffin & Hepburn, 2006), segregation of
confirmed or suspected gang members is among the most criticized of practices
because there is no clear endpoint to the isolation. Some inmates classified to
administrative segregation for known or suspected gang affiliations have spent

10 • National Institute of Justice | NIJ.gov

decades in isolation units (Baker & Goode, 2015). The decision to release almost
all gang members currently held in solitary confinement in California will result
in the release of almost half of inmates incarcerated in secure housing across
the state. It is not yet clear what impact, if any, the settlement will have on other
jurisdictions that routinely incarcerate known or suspected gang members in
administrative segregation.

Mental Illness and Solitary Confinement
Although precise numbers are hard to come by, some have argued that most
inmates who are placed into solitary confinement are mentally ill (Toch, 2001).
By virtue of their illness, these individuals may have trouble conforming to
institutional rules and accrue more disciplinary misconduct sanctions (Kurki &
Morris, 2001). The Bureau of Justice Statistics (BJS) has reported that close to 45
percent of federal inmates and more than half of all jail and state prison inmates
suffer from mental health problems as measured by diagnosis and treatment
of symptoms (James & Glaze, 2006). Earlier estimates, which focused narrowly
on mental illness, tended to be much lower: approximately 16 percent of prison
inmates were classified as suffering from serious mental illness (Ditton, 1999;
Osher et al., 2012).
Some prevalence estimates of mental illness across populations in administrative
segregation have been derived from the empirical research (O’Keefe, 2007).
Lovell and colleagues (2008) have produced several estimates of the levels of
psychosocial impairment (a construct that includes several measures of mental
health deterioration) and serious mental illness among the supermax population
in Washington state. Lovell (2008) randomly sampled inmates from all three of
Washington’s supermax housing units and found that 45 percent of supermax
inmates were suffering from serious mental illness. Lovell and colleagues (2000)
earlier used similar methods to find serious mental illness among approximately
13 percent of general population inmates (Lovell et al., 2000). Cloyes and
colleagues (2006) similarly reported that almost 30 percent of inmates in
supermax units meet the criteria for serious mental illness, and in their study of
the supermax unit in Washington, they found that 22 percent of the inmates were
actively experiencing high levels of “psychosocial distress.” Many researchers
have used similar statistics to argue that inmates with mental illness are more
likely to be placed in solitary confinement.
The courts have been particularly active in the issue of confining mentally
ill inmates in administrative segregation, based in part on the accumulated
evidence that confinement under such restrictive and isolating conditions is
especially harmful for this already vulnerable population.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 11

Court Decisions and Consent Decrees
It is beyond the scope of this paper to comprehensively review all court cases
and legal decisions relative to the use of solitary confinement; therefore, a review
of only a small sampling of the more significant and directly relevant court
decisions and consent decrees are presented (for a comprehensive review, see
Collins, 2004).
The most notable litigation that addresses administrative segregation has focused
on supermax confinement or solitary confinement in administrative, rather than
disciplinary, segregation.6 State and federal cases related to the use of solitary
confinement have tended to focus on the overall conditions of confinement in
supermax settings, and several have focused on the placement of mentally ill
inmates into facilities that use solitary confinement for extended periods.
An early federal case signaled that the courts were not likely to get involved in
the administration of facilities used for administrative segregation. In 1984, a
group of inmates housed at the Marion Federal Penitentiary filed a Section 1983
complaint alleging that federal prison officials were violating their constitutional
rights. The primary complaints Bruscino v. Carlson (1988) were related to
arbitrary placements, conditions of confinement, extended solitary confinement,
use of force, and cavity searches. The U.S. Supreme Court, in considering both
the context and the complaint, found that the “ghastly” conditions at Marion,
although “depressing in the extreme,” did not amount to cruel and unusual
punishment. After describing Marion’s inmates as “among the most dangerous
people in this nation,” the Court was “not persuaded that any relaxation in the
controls instituted in the fall of 1983 [is] constitutionally required, given the
extraordinary security problems at the prison.” The Court went on to note,
“The controls are a unitary and integrated system for dealing with the nation’s
least corrigible inmates; piecemeal dismantling would destroy the system’s
rationale and impair its efficacy.” The Bruscino decision signaled that the federal
courts would be unlikely to interfere with the management of administrative
segregation units. As Feeley and Rubin (1999) noted in the comprehensive
overview of correctional cases, “Whatever the reasons for the judiciary’s positive
response to Marion, the correctional establishment chose to interpret it as
validating the concept of a supermaximum-security prison.”
A few years later (1990), a group of inmates in Pelican Bay State Prison filed a
Section 1983 claim against CDCR. In the 1995 Madrid v. Gomez decision, the
court found in favor of the inmates, ruling that:
6

Placements into disciplinary segregation for infractions are subject to the landmark U.S. Supreme Court case
Wolff v. McDonnell (1974). The Wolff decision established the minimal procedural due-process rights that must
be afforded to inmates during prison disciplinary hearings, and these rights have since been reaffirmed in Dixon
v. Goord (2002). The minimum procedural due-process rights during disciplinary hearings include advance
written notice of charges, an advance written statement of evidence on which the determination will be made,
and the right to call witnesses and present evidence.

12 • National Institute of Justice | NIJ.gov

In particular, defendants have failed to provide inmates at Pelican Bay with
constitutionally adequate medical and mental health care, and have permitted
and condoned a pattern of using excessive force, all in conscious disregard of the
serious harm that these practices inflict. With respect to the SHU, defendants
cross the constitutional line when they force certain subgroups of the prison
population, including the mentally ill, to endure the conditions in the SHU,
despite knowing that the likely consequence for such inmates is serious injury to
their mental health, and despite the fact that certain conditions in the SHU have
a relationship to legitimate security interests that is tangential at best (p. 1280).
Although its concern for inmates with mental illness was clear, the Court was
less definitive when it came to inmates not suffering from mental illness; it
noted, “while the conditions in the SHU may press the outer bounds of what
most humans can psychologically tolerate, the record does not satisfactorily
demonstrate that there is a sufficiently high risk to all inmates of incurring a
serious mental illness from exposure to conditions in the SHU to find that the
conditions constitute a per se deprivation of a basic necessity of life” (p. 1267).
As a result of the Madrid decision, a special marshal was appointed to work with
CDCR to develop a plan to remedy the conditions at Pelican Bay (Fathi, 2004). In
many ways, the Madrid case provided the first in-depth look into the conditions
of confinement in supermax settings. The Court continued to monitor the
case for more than a decade until, finally satisfied that the conditions had been
remedied, it dismissed the case in 2011 (Simon, 2014).
Other prison systems have drawn the attention of the courts as well. In Ruiz v.
Johnson (1999), the administrative segregation units of the Texas Department
of Corrections came under judicial scrutiny, and the federal courts came
close to declaring that solitary confinement in administrative segregation is
unconstitutional per se. After describing the conditions of confinement in
administrative segregation units across the Texas Department of Corrections, the
court in Ruiz declared:
Before the court are levels of psychological deprivation that violate the United
States Constitution’s prohibition against cruel and unusual punishment. It has
been shown that defendants are deliberately indifferent to a systemic pattern
of extreme social isolation and reduced environmental stimulation. These
deprivations are the cause of cruel and unusual pain and suffering by inmates in
administrative segregation, particularly in Levels II and III (pp. 914-915).
The Ruiz decision was, in many ways, more sweeping than that of Madrid several
years earlier. Most subsequent cases have resulted in settlements or consent
decrees, but it is worth noting that conditions of confinement in administrative
segregation have been challenged in Connecticut, Indiana, New Mexico, Ohio,
and Wisconsin, and other states (see Fathi, 2004).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 13

Although the court in Ruiz came close to ruling solitary confinement
unconstitutional, it can be said with confidence that this country is moving
toward a general consensus (as illustrated by various court decisions, consent
decrees, and settlement agreements) that these environments are not appropriate
for inmates with mental illness and might constitute cruel and unusual
punishment for this subset of the inmate population. The U.S. Supreme Court has
regularly denied certiorari in administrative segregation cases, but with Justice
Anthony Kennedy repeatedly expressing his concern about solitary confinement
in recent months, it seems likely that there could be some U.S. Supreme Court
movement on this front in the coming years (Hananel, 2015).
In addition to court cases and consent decrees, there have been several notable,
recent congressional and legislative hearings related to the use of solitary
confinement.7 In June 2012 and February 2014, the U.S. Senate Judiciary
Subcommittee on the Constitution, Human Rights, and Civil Rights held two
hearings on the use of solitary confinement.8 Several notable legislative hearings
have also been held in California, where conditions of confinement in general
and administrative segregation in particular have been the focus of ongoing
litigation. The California State Assembly and Senate Public Safety Committee
held two hearings on CDCR’s use of solitary confinement.9 As noted, CDCR
recently reached a settlement agreement to end a class action lawsuit on behalf of
inmates at Pelican Bay (500 of whom had been held in solitary confinement for
more than 10 years at the time the suit was filed). The lawsuit was preceded by a
series of well-publicized hunger strikes across CDCR facilities that also triggered
the legislative hearings on the issue.

The Utility and Effects of Administrative Segregation
Proponents of administrative segregation and the supermax model argue that
solitary confinement is necessary for maintaining the safety and security of
the entire correctional system. Some inmates, it is argued, are so disruptive
to the orderly running of a facility that they simply cannot or should not be
maintained among the general population (O’Keefe, 2008; Pizarro & Narag,
2008; Pizarro & Stenius, 2004; Pizarro, Stenius, & Pratt, 2006). This view, however,
is far from universal. The most obvious division is between scholars (who
study either incarceration in general or solitary confinement in particular) and
7

Written and oral testimony, as well as video recordings of the hearings discussed in this section, can be accessed
on the advocacy group Solitary Watch’s website (http://solitarywatch.com/).

8

The hearings, “Reassessing Solitary Confinement: The Human Rights, Fiscal, and Public Safety Consequence,”
each included oral and written testimony from heads of departments of corrections, researchers who have
studied solitary confinement, and activists against the practice (Solitary Watch, 2015). See https://www.judiciary.
senate.gov/imo/media/doc/CHRG-112shrg87630.pdf

9

Both hearings before the California Assembly focused on CDCR’s proposed reforms to inmate segregation
policies and to the use of secure housing units.

14 • National Institute of Justice | NIJ.gov

the correctional administrators responsible for drafting and enforcing official
correctional policies and practices.
Recent reviews of policies related to placement in administrative segregation also
emphasize the official view that segregation minimizes threats to institutional
security (Butler et al., 2013; O’Keefe, 2008). A recently published study that
reviewed official policies across 42 states reported that 98 percent of the states
identified the catch-all term “threats to institutional security” as a justification
for placement in administrative segregation (Butler et al., 2013). The most
commonly noted threats to institutional security that trigger administrative
segregation included repeated violent behavior (78 percent), escape risk (67
percent), riotous behavior (45 percent), and security threat group membership
(36 percent). Concerns remain about the specific criteria used to determine that
an inmate represents such a threat, particularly in the absence of full due-process
rights that typically accompany placement in disciplinary segregation after a
discrete incident of violent or disruptive behavior (O’Keefe, 2008).
Despite the dearth of empirical evidence demonstrating effectiveness, those
charged with running correctional facilities overwhelmingly believe that
administrative segregation achieves its aims (particularly as related to increasing
the safety and security of the correctional system). In one of the largest studies
of practitioner’s views, Mears and Castro (2006) noted that prison wardens,
who maintain primary responsibility for running the nation’s correctional
facilities, were not only “largely unanimous in saying that supermax prisons
serve to increase safety, order, and control throughout the prison system and
to incapacitate violent and disruptive inmates” (p. 407) but also “strongly
believe that supermax prisons are effective in achieving these four goals”
(Mears & Castro, 2006, pp. 407, 409). Those who argue for the effectiveness
of administrative segregation units and supermax facilities often base their
arguments on appeals to the self-evident rather than on an evidence base (Sundt,
Castellano, & Briggs, 2008, p. 115). Noting that these units and facilities house
the worst of the worst, proponents of the practice argue that these facilities will
reduce violence in prison systems and assaults on other inmates and staff because
those who are most likely to engage in such conduct have been isolated and
further incapacitated (Butler et al., 2013; Lanes, 2011; Mears et al., 2013).
The public, at least as gauged by opinion in Florida, also overwhelmingly
supports supermax prisons, even when their utility is less than clear. Mears and
colleagues (2013) reported that 80 percent of the public supported supermax
incarceration in general and that 60 percent maintained that support even if
there was no associated public safety benefit. Moreover, there appears to be
little public support for the notion that these facilities are inhumane: 70 percent
of those surveyed said that they did not consider supermax facilities to be
inhumane (Mears et al., 2013).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 15

Evaluation Research
Mears and Watson (2006) identified several of the most frequently cited
justifications for (and goals of) solitary confinement through administrative
segregation, including increasing prison safety, increasing systemwide order
and control of prisoners, improving the behavior of violent and disruptive
prisoners, reducing gang influence, punishment (of violent and disruptive
prisoners), increasing public safety, and improving the efficiency of correctional
system operations. Although some are easier to operationalize than others,
any of these intended impacts could be evaluated. In their 2006 article,
Mears and Watson even offered measurable indicators for each “performance
measure,” (see Appendix Table A3), which may prove useful to those seeking
an evaluation framework. Mears and Watson also offered observations about
some unintended potential impacts (both positive and negative). Finally, they
raised questions related to the mechanisms by which supermax confinement is
expected to achieve these goals; identified barriers to achieving them; and offered
some moral, political, and fiscal dimensions that should be considered in any
comprehensive assessment of supermax prisons.
When Mears (2008) attempted to apply an evaluation framework to supermax
incarceration, he struggled to find the research base necessary to answer key
questions across five domains, ultimately concluding that:
[T]here is (a) minimal indication that supermax prisons were needed as long­
term solutions to any of a range of problems (e.g., order, safety, escapes, public
safety); (b) no strong or consistent theoretical foundation for anticipating that
they would exert any substantial effect on a range of outcomes, and, to the
contrary, strong theoretical grounds to anticipate a worsening of these outcomes;
(c) minimal documentation of their implementation (including the procedures
and adherence to these procedures) for admitting and releasing inmates,
monitoring of inmate behavior, or compliance with state and federal laws as
well as constitutional requirements, juxtaposed against accounts showing that
mentally ill and other inmates inappropriate for extended solitary confinement
reside in supermaxes; (d) minimal evidence of any positive impact on any of a
range of outcomes, with considerable evidence of harmful, unintended effects;
and (e) no evidence that they are cost-efficient (Mears, 2008, p. 61).
If dangerous and violent inmates represent a real threat to others within the
correctional environment, the options for containing them without resorting to
isolation in restrictive housing seem to be few. Correctional administrators often
feel that they are left with no other option than to isolate inmates who represent a
threat to themselves, other inmates, or to staff. Moreover, the most ardent critics
of solitary confinement often have little to say about alternatives to the practice
if and when solitary confinement were to be eliminated on evidence-based
grounds or outlawed on constitutional grounds. Those charged with running
prisons — even those who have argued that there are problems with the practice

16 • National Institute of Justice | NIJ.gov

of confining inmates in highly restrictive environments for extended periods
— lament the lack of options at their disposal for those inmates who are truly
dangerous to both the prison staff and other inmates. Rick Raemisch, executive
director of the Colorado Department of Corrections, spent 20 hours in solitary
confinement to understand the experience and has worked to significantly reduce
its use in the Colorado system but still acknowledges the need for the practice in
some instances (Goode, 2014; Raemisch, 2014). Raemisch recently argued, “If
someone has committed a violent assault … until you can solve that problem,
that person is going to need to be isolated.” He went on to note, “There are those
who say this is bad, but when you look around for an alternative, people have left
the room” (Baker & Goode, 2015, p. A16).

Violence in Correctional Institutions
It is difficult to determine with any degree of precision the prevalence of the use
solitary confinement through administrative or disciplinary segregation across
U.S. correctional systems (Naday et al., 2008). Perhaps not surprisingly, there are
few accurate estimates of either the levels of disruptive behavior among those
sent to administrative segregation or of the impact of administrative segregation
on reducing levels of violence in prisons.
Some of the estimates of violence within correctional institutions come from
victimization surveys and research. Wolff and colleagues, for example, reported
in their study of inmates incarcerated in 1 of 14 institutions in a mid-Atlantic
state that more than 20 percent of inmates reported being the victim of physical
violence in the previous six months (Wolff et al., 2007). It is important to note
that although rates of victimization were roughly equal for male and female
inmates, prevalence rates of victimization varied substantially across facilities
(even within this state’s correctional system). Although there are little hard data
on the subject, some evidence shows that serious assaults against correctional
officers are rare but tend to be more — not less — common in administrative
segregation units (Sorenson et al., 2011).
One study that provided a profile comparing inmates in administrative
segregation with general population inmates in Colorado suggested that those
housed in administrative segregation incurred significantly more disciplinary
infractions, were significantly more likely to have been previously placed in
punitive segregation, and were significantly more likely to have been identified as
having a known or suspected security threat group affiliation (O’Keefe, 2008). In
his study of supermax inmates in Washington that focused on inmates suffering
from mental illness, Lovell (2008) found that these inmates had substantially
higher infraction rates and that many of those infractions were indeed indicative
of disruptive institutional behavior:

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 17

These 60 inmates had committed 135 assaults: 45 aggravated and 65 (including
five aggravated) on staff. Four of them had infractions for homicide. Lessviolent forms of disruptiveness included 220 infractions for threatening, 168 for
throwing objects (often urine or feces), 83 for destroying property, and 28 for
flooding cells. Twelve men had been infracted for mutilating themselves, usually
two or three times (p. 990).
In addition to the work on levels of violence committed by inmates sent to
segregated housing, some effort has been made to assess the impact of SHUs on
levels of correctional system violence.

Institutional Violence and Administrative Segregation
In the late 1980s and 1990s, a series of authors offered some (mostly speculative)
evidence that the expanded use of administrative segregation resulted in
lower levels of violence across correctional systems. In their discussions of
the effects of court-ordered changes in the Texas correctional system in the
wake of the sweeping Ruiz v. Estelle (1980) decision, Marquart and colleagues
tied reductions in prison violence and inmate murders, as well as increases in
inmates’ perceptions of safety, to the extensive use of administrative segregation,
particularly of gang members, across the Texas Department of Corrections
(Crouch & Marquart, 1989, 1990; Ralph & Marquart, 1991). Austin and Irwin
(2001) similarly tied declines in prison violence in California’s prison system to
increased use of segregation. The authors of both studies looked back at declines
in violence and speculated that increased reliance on segregation might be a
cause. More recently, researchers have assessed the impact of administrative
segregation on levels of prison violence by using more sophisticated research
designs; those researchers have typically reported mixed support for the thesis
that the increased use of segregation has resulted in decreases in prison violence.
In a national study of inmate violence that used multilevel data from more than
4,000 inmates across 185 correctional institutions, Huebner (2003) found that the
use of solitary confinement for disciplinary purposes, measured as “the percent
of the total inmate population that received solitary confinement as a disciplinary
response to the most recent rule infraction,” was unrelated to levels of inmate
assaults (Huebner, 2003, p. 110). Using data from three states (Arizona, Illinois,
and Minnesota) and one control state (Utah), and a multiple interrupted timeseries design, Briggs, Sundt, and Castellano (2003) found no evidence that the
introduction of a supermax facility had any effect on inmate-on-inmate violence
in any state. Their findings related to inmate-on-staff violence were mixed, with
no effect found in Minnesota, decreased inmate-on-staff assaults in Illinois, and
a counterintuitive, temporary increase in staff injuries in Arizona (Briggs, Sundt,
& Castellano, 2003). Briggs and colleagues (2003) concluded that “the bulk of
the evidence presented here suggests that supermax is not effective at reducing
system wide levels of prison violence” (p. 1368).

18 • National Institute of Justice | NIJ.gov

Sundt and colleagues (2008) used a quasi-experimental interrupted time-series
design to examine the effect of the opening of a supermax facility on subsequent
levels of inmate-on-inmate and inmate-on-staff violence in prisons in Illinois.
They found that the opening of the supermax facility did not significantly affect
the number of inmate-on-inmate assaults but did result in a “gradual permanent
reduction in assaults against staff ” and an “abrupt, permanent reduction in the
use of system wide lockdowns” (Sundt et al., 2008, pp. 115, 117). These findings
lend credence to the notion that supermax facilities might increase the safety
of the entire correctional system. However, because of the unique (and volatile)
political context in which the Illinois supermax facility was built, the results are
not particularly generalizable to other states (Sundt et al., 2008). Over all, there
is little evidence that the introduction of supermax facilities has reduced levels
of violence across the correctional system. Given the paucity of research in this
area, however, it seems fair to say that, like so many factors related to the use of
administrative segregation, this remains an open, empirical question.
More recently, Wooldredge and Steiner’s (2015) research examined the extent
to which misconduct levels in prisons are driven by the effects of inmate
population composition (concentrating the highest risk inmates in particular
facilities) versus by the prison’s organizational context (tighter institutional
security and control). In their macro-level analysis of 247 state-operated prisons
across the country, using both official and self-reported data collected through
inmate surveys, Wooldredge and Steiner’s (2015) multi-level models revealed
that compositional effects were more influential than contextual effects in
shaping behavioral outcomes. The authors reported that inmate population
composition characteristics (demographics, social demographics, and criminal
histories) were more important than administrative and contextual controls
(classification, security, supervision, expenditures, and punishment philosophies)
for understanding differences in prison misconduct levels. In essence, the results
suggest that coercive controls do not necessarily promote lower levels of violence
within facilities, and that the concentration of high-risk populations typically
drives violence levels. Although this was only a partial test of the authors’
full thesis, the results are consistent with other studies that have identified
compositional effects as more significant indicators of violence levels.

The Effects of Solitary Confinement
Psychological Effects of Solitary Confinement
Two types of studies of the psychological effects of isolation have been used:
qualitative studies that focused on providing rich descriptions of the effects of
the experience on inmates who have typically spent considerable time in solitary
confinement (often having spent many years incarcerated in isolation), and

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 19

quantitative studies of such effects among larger groups of inmates, sometimes
using matched comparison or control groups, but typically focusing on inmates in
disciplinary units or those serving shorter terms in administrative segregation.10
The only clear statement that can be made about the body of literature that
assesses the psychological effects of solitary confinement is that researchers
using different methods to study different populations have come to different
conclusions about the psychological effects on inmates (Arrigo & Bullock, 2008;
Gendreau & Labrecque, 2015). A fair summary statement would say that there
is a collection of scholars who have been studying solitary confinement for
many years; these scholars strongly believe that the experience can have lasting
and substantial damaging psychological effects. The most well-known findings
about the impact of extended solitary confinement on the mental health of
prisoners have come from the accumulated work of Stuart Grassian and Craig
Haney, both of whom have testified extensively on behalf of inmates in class
action lawsuits (Grassian, 1983; Grassian & Friedman, 1986; Haney, 2002, 2003,
2008; Haney & Lynch, 1997). In one of his earliest studies, Grassian conducted
extensive interviews with 14 prisoners who were in the process of challenging
the conditions of their confinement in a lawsuit against the Massachusetts
Department of Corrections. He documented a long list of damaging
psychopathological effects, including difficulty concentrating and thinking,
perceptual distortions and affective distortions changes, and problems with
impulse control (Grassian, 1983). Grassian concluded that the psychopathology
he documented “strongly suggests that the use of solitary confinement carries
major psychiatric risks” (p. 1454).
Many other studies employing a case-study approach across a variety of settings
have similarly documented far-reaching and long-lasting psychological effects
(for more comprehensive reviews of studies of the psychological effects, see
Gendreau & Labrecque, 2015; Labrecque & Smith, 2013; Shalev, 2008; Smith,
2006). In summarizing the psychological literature in 2002, Haney reported that:
[D]ocumented negative psychological consequences of long-term solitary-like
confinement include: an impaired sense of identity; hypersensitivity to stimuli;
cognitive dysfunction (confusion, memory loss, ruminations); irritability, anger,
aggression, and/or rage; other directed violence, such as stabbings, attacks on
staff, property destruction, and collective violence; lethargy, helplessness and
hopelessness; chronic depression; self-mutilation and/or suicidal ideation, impulses,
and behavior; anxiety and panic attacks; emotional breakdowns, and/or loss
of control; hallucinations, psychosis and/or paranoia; overall deterioration of
mental and physical health (Haney, 2002, pp. 85-86).
10

Although rarely acknowledged, the research on psychological or psychiatric effects frequently relies on a large
body of literature on the effects of sensory deprivation (Gendreau & Labrecque, 2015). Reviewing that literature
is beyond the scope of this review of administrative segregation as a practice, but suffice it to say that it is often
taken for granted that isolation will have severe and lasting detrimental effects on the psychological well-being
of all those exposed to it, even though the evidence in this area does not always bear out this assumption (for a
comprehensive review, see Gendreau & Labrecque, 2015).

20 • National Institute of Justice | NIJ.gov

A substantial body of work has shown that solitary confinement can have
damaging psychological effects, particularly when that confinement involves
near-complete isolation and sensory deprivation or when the term of such
confinement is extended. Moreover, despite the methodological limitations of
some studies, most agree that extended confinement under extreme conditions
of isolation in some segregation units is indeed harmful and should be avoided
where possible. More questionable, however, is whether the populations who
were the subjects of study across this body of research are representative of
inmates experiencing administrative segregation more generally. Although
Haney is confident in his assertions about the psychological effects of long-term
solitary confinement, other research involving in-depth case studies has reached
the opposite conclusion in terms of the psychological effects of shorter-term
solitary confinement.
In an early study of the experience of just four inmates, Suedfeld and Roy (1975)
argued that short-term solitary confinement (ranging from one week to 30 days)
was beneficial to those inmates and resulted in a lower incidence of violence,
aggression, and self-injurious behavior, as well as improved adjustment. In later
work involving a much larger group of inmates incarcerated in five facilities
across the United States and Canada, Suedfeld and colleagues again found
“no support [for] the claim that solitary confinement … is overwhelmingly
aversive, stressful, or damaging to the inmates” (Suedfeld et al., 1982, p. 335).
Several decades later, Zinger and colleagues (2001) compared inmates held in
administrative segregation in Canada with those in the general population and
found that the “segregated prisoners had poorer mental health and psychological
functioning. There was no evidence, however, that, over a period of 60 days, the
mental health and psychological functioning of segregated prisoners significantly
deteriorated” (Zinger et al., 2001, p. 48). Other respected scholars have also been
less than convinced by the accumulated evidence pertaining to psychological
effects of segregation. Bonta and Gendreau (1990), for example, argued that there
is little evidence of deteriorating mental health among inmates, emphasizing
that “long-term imprisonment and specific conditions of confinement such as
solitary, under limiting and humane conditions, fail to show any sort of profound
detrimental effects” (Bonta & Gendreau, 1990, p. 364).
In another prospective longitudinal study, albeit for a more extended period (one
year, as opposed to 60 days), O’Keefe and colleagues (2011) sought to improve
understanding of the psychological effects of solitary confinement on inmates
in administrative segregation in Colorado. By relying primarily on assertions
in earlier psychological research about the effects of solitary confinement, the
authors hypothesized that inmates who had served time in administrative
segregation would experience aggravated psychological symptoms while in
administrative segregation, exhibit deteriorating mental health over time (with
that effect exacerbated among those with preexisting mental health issues), and
experience greater mental health deterioration than would those in comparison
groups of inmates who had not served time in segregation.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 21

The researchers collected data through a series of interviews with and assessments
of inmates in administrative segregation, the general population, and a specialized
mental health unit. To their apparent surprise, none of their initial hypotheses was
borne out by their data (O’Keefe et al., 2011). Inmates in administrative segregation
exhibited more mental health issues than did “normative adult samples,” but there
were few differences between inmates in administrative segregation and those
in comparison groups. In other words, the researchers documented elevated
levels of mental health problems across both groups, but those problems were
not significantly more pronounced among those in administrative segregation.
Moreover, although inmates in administrative segregation initially exhibited
signs of elevated psychological distress, those signs tended to dissipate over time.
Psychological problems tended to decrease between the first and the second testing
period, which was true across both groups, not just for the group in administrative
segregation. Finally, the researchers reported that inmates with mental illness
demonstrated greater psychological impairment across all groups, not just in the
segregated setting, and did not deteriorate more rapidly than initially predicted
(O’Keefe et al., 2011).
Perhaps the most interesting finding of O’Keefe and colleagues’ Colorado study
was not that time in administrative segregation had no deleterious effect but
that the deleterious effects experienced by those in administrative segregation
were no different from those experienced by prison inmates in general. In some
ways, this finding is equally, if not more, disturbing. O’Keefe and colleagues’
findings could as easily be interpreted as demonstrating that incarceration in and
of itself has damaging effects on the mental health of individuals subjected to it,
especially initially. Of course, because these findings undercut some of the earlier
research and call into question the extent to which administrative segregation
has distinct and distinguishable damaging psychological effects, the research has
come under enhanced scrutiny (Bulman, Garcia, & Hernon, 2012).
Shortly after it was published, the Colorado study was subjected to a series of
methods critiques that called into question the validity of its results, and several
attempts were made to undermine its impact (Grassian & Kupers, 2011; Rhodes
& Lovell, 2011; Shalev & Lloyd, 2011). The various critiques had some merit but,
as Gendreau and Theriault (2011) noted, the type of work represented in the
Colorado study (with controversial or unpopular findings in a highly contested
domain) is frequently attacked on methods grounds, adding that “none of the
work we are aware of that has been cited by those who contend that prisons
produce serious psychological trauma comes close to the Colorado study in
terms of its methodological rigor (e.g., repeated measures, comparison group
design, and the choice of constructs to assess psychopathology)” (Gendreau &
Theriault, 2011, p. 1). It is also worth noting that the findings in the Colorado
study are not antithetical — other researchers have reported similar findings on
the basis of previous research (Suedfeld et al., 1982; Zinger et al., 2001).

22 • National Institute of Justice | NIJ.gov

It should be noted that a key distinction among the separate bodies of
work related to psychological effects is that those who have found marked
psychological distress among inmates subjected to solitary confinement have
tended to study those held in solitary confinement for extended periods; whereas
researchers who find no convincing evidence of distress have tended to study
inmates held in solitary confinement for far more abbreviated periods (30, 60 or
90 days, for example). It should also be noted that those finding excessive harm
tended to employ an intensive, qualitative, case-study approach, conducting
extensive interviews with (and assessments of) inmates held, often indefinitely,
in solitary confinement. In some instances, the researchers have been experts,
retained on behalf of inmates who were filing claims against departments of
corrections, preparing for written or oral testimony for cases to be heard in state
and federal courts.

Behavioral Effects of Solitary Confinement
Some researchers have sought to measure the effects of segregation on behavioral
(rather than psychological) outcomes. Researchers seeking to better understand
behavioral effects typically rely on deterrence theories to argue that, if effective,
solitary confinement should reduce levels of institutional misconduct (both
individual and systemwide) and should be expected to have some effect on
post-release behavior (where one could argue for effects in either direction).
These researchers have tended to focus on one of two behavioral outcomes —
institutional misconduct or post-release recidivism — and have typically studied
inmates who have experienced solitary confinement, irrespective of the type of
restrictive housing unit. For the most part, researchers have found that inmates
who spent time in solitary confinement through administrative or disciplinary
segregation fare no better or worse than inmates never exposed to it.

The effects of solitary confinement on institutional misconduct
Relatively few studies have focused on the effect of solitary confinement on
subsequent institutional misconduct, a somewhat surprising finding given that
a primary claim about administrative segregation is that it should restore order
and lead to greater safety and security in correctional facilities.
Labrecque (2015) recently provided an assessment of the impact of solitary
confinement on subsequent institutional misconduct among inmates who were
incarcerated for at least one year between 2007 and 2010 and experienced at
least one instance of solitary confinement in an Ohio Department of Corrections
facility. The study focused solely on inmates who had experienced solitary
confinement. Those inmates represented approximately 21 percent of the inmates
in the overall sampling frame covered by the period. Labrecque (2015) employed
a pooled time-series panel design to “assess whether [the solitary confinement]

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 23

experience has an influence on being found guilty of subsequent institutional
misconduct” (p. 76).
Labrecque (2015) included in his assessment all of the various types of
misconduct, finding that an experience of solitary confinement had no effect on
subsequent levels of misconduct, and that the duration of solitary confinement
also had no effect on subsequent institutional misconduct. Labrecque concludes
“that neither the experience of [solitary confinement] nor the number of days
spent in [solitary confinement] had any effect on the prevalence or incidence
of the finding of guilt for subsequent violent, nonviolent, or drug misconduct”
(Labrecque, 2015, p. 122). Labrecque’s findings related to institutional outcomes
suggest that such confinement neither decreases nor increases subsequent
institutional misconduct and add to the growing body of literature that suggests
that solitary confinement has few demonstrable effects on behavioral outcomes
for those exposed to it.
Morris (2015) recently studied the effect on violence of short periods of
solitary confinement after misconduct. Acknowledging that short-term solitary
confinement might be beneficial (reducing subsequent violence through
deterrence), harmful (exacerbating problem behavior among inmates exposed
to it), or inconsequential, Morris used a propensity score matching (PSM)
technique to create treatment and control groups in which the primary difference
between the groups was exposure to solitary confinement. As Morris noted,
“PSM approximates the conditions of an experiment by establishing ‘synthetic’
treatment and control groups, which are balanced across all available variables
known or believed to potentially confound the effect of exposure to [solitary
confinement]” (p. 6).11 Morris hypothesized that inmates exposed to punitive
solitary confinement (typically of up to 15 days) after an initial act of violent
misconduct would be more likely to engage in subsequent violence, engage in
violence more quickly than would those not exposed, and exhibit more antisocial
tendencies than would those in the control group. Morris’ results indicated
that “on average, the initial experience with [solitary confinement] alone (as a
direct and independent effect) may not play a causal role in subsequent physical
violence, its timing, or its downstream effect on misconduct development. These
findings suggest neither a positive nor negative relationship between solitary
confinement and subsequent violent behavior, nor for misconduct in general
(following initial violence) (Morris, 2015, p. 17).
Although certainly not without limitations, among studies of solitary
confinement (and like O’Keefe et al., 2011), Morris’s study represents an
analytically sophisticated effort to isolate the effect of solitary confinement on
subsequent behavior (in this case, subsequent violence). The study overcame
11

The PSM technique is increasingly used when random assignment to treatment and control groups is not
possible. The technique results in a quasi-experimental counterfactual design and is generally considered to be
the next-best analytical approach when experimental designs are not feasible.

24 • National Institute of Justice | NIJ.gov

many of the methods limitations of earlier work by including a relatively large
sample from a single state measured over time, incorporating treatment and
control groups (with controls for selection bias), and assessing effects across
multiple outcome measures. Of course, Morris’s study focuses on only inmates
sentenced to solitary confinement in disciplinary segregation for a relatively
short period; it cannot speak to behavioral effects of longer-term confinement
or to confinement for purposes other than punishment. That research has yet
to be conducted.

The effects of solitary confinement on recidivism outcomes
Some effort has been made to understand the impact of solitary confinement
on post-release outcomes, usually recidivism rates. In one of the few studies of
supermax incarceration in federal prisons, Ward and Werlich (2003) examined
post-release outcomes for inmates who had been incarcerated at both Alcatraz
and Marion, and reported the extent to which inmates who had been released
from restrictive housing were later returned to such custody. They found that
only 3 percent of prisoners who had returned to the general prison population
from Alcatraz and 16 percent of prisoners returned to the general prison
population from Marion were returned to higher custody. Less than half of those
who returned to higher custody were returned for violent or assaultive behavior.
The post-release outcomes for inmates incarcerated in federal supermax facilities
suggest that the reincarceration rate (for inmates who had spent time in Alcatraz
or Marion) was approximately 49 percent. This study, although interesting,
was largely descriptive and included no control or comparison group (Ward &
Werlich, 2003).
In a more recently published study of recidivism outcomes, Pizarro, Zgoba, and
Haugebrook (2014) examined the covariates of recidivism among inmates in
supermax custody in a northeastern state in 2004. They found that the covariates
for these inmates were almost identical to those for inmates in general. Supermax
inmates who recidivated tended to be younger, have more extensive criminal
histories, and were more likely to have histories of disciplinary misconduct
than were supermax inmates who did not recidivate (Pizarro, Zgoba, &
Haugebrook, 2014). Although interesting, this study compared supermax
inmates with other supermax inmates, so it said little about how inmates who
have served time in supermax facilities compare with inmates never exposed to
extended solitary confinement.
Other research on recidivism outcomes has attempted to overcome the effects
of selection bias inherent in much of the research on the effects of solitary
confinement in general. In one such study, using a retrospective matched
control design, researchers examined recidivism outcomes for prisoners
in Washington state and reported that, although no statistically significant
differences were found in the recidivism rates of supermax inmates (compared
with their matched controls), significantly higher recidivism rates were found

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 25

among inmates released directly from supermax confinement to society (Lovell,
Johnson, & Cain, 2007). In other words, inmates held in supermax confinement
until the day of their release offended more quickly and more often than did
their counterparts who had either never served time in supermax confinement
or had been returned from supermax confinement to the general population at
least three months prior to their release. The supermax inmates released directly
to society also differed from the others in that they tended to be younger and to
have more extensive criminal histories (beginning at an earlier age). Because age
and criminal history are among the most significant predictors of recidivism,
Lovell and colleagues (2007) matched by age and criminal history direct-release
inmates with those who were returned to the general population at least three
months prior to release. Although there were differences between the groups,
when matched on age and criminal history, these differences were not statistically
significant. The authors attribute the nonsignificance in part to the small sample
sizes of the two groups. It is equally as plausible that the finding may not be an
artifact of sampling size, but instead may reflect the lack of an effect of supermax
exposure on recidivism outcomes when one controls for other relevant factors.
Although provocative, these findings cannot be considered to be definitive in
any way because of the nonexperimental retrospective research design and small
sample size when a matched control group was incorporated.
A second study that examined the impact of supermax confinement on
recidivism in Florida added to the evidence that supermax incarceration might
have negligible effects on post-release recidivism (Mears & Bales, 2009). Mears
and Bales examined recidivism outcomes for more than 1,200 inmates who had
been released after having served time in Florida’s supermax prisons and found
that, when compared with all Florida inmates, those who had served time in
supermax confinement were much more likely to recidivate were those who
had not. However, and crucially, when they compared supermax inmates with a
PSM control group, almost all of the differences in recidivism disappeared (only
violent recidivism remained elevated), which suggests that the initial differences
were likely an artifact of selection bias. Like Lovell and colleagues (2007), Mears
and Bales also examined the potentially differential effects of the duration
and recency of supermax confinement at the time of release. They concluded
that “neither the duration nor the recency of supermax incarceration seems
to be consequential for recidivism” (p. 1153). In other words, although Lovell
and colleagues had offered some preliminary evidence that inmates who were
released directly to society might be at elevated risk for recidivism, Mears and
Bales found no such evidence among Florida supermax populations.

Meta-Analyses
Two teams of scholars have recently conducted meta-analyses. Meta-analyses
quantitatively synthesize research evidence by taking findings from existing
research and calculating overall effect sizes across studies. In their recent

26 • National Institute of Justice | NIJ.gov

meta-analysis, Smith, Gendreau, and Labrecque (2015) used inclusion criteria
that required a measure of solitary confinement as an independent variable, a
research design that included either randomized selection or comparison and
control groups, and sufficient data to calculate an effect size (Smith, Gendreau,
& Labrecque, 2015). Of the 150 studies of solitary confinement identified by the
authors, 70 percent had been published in the past 15 years and only 14 could be
included in the meta-analysis (in other words, approximately 90 percent of the
studies did not meet the inclusion criteria). Their meta-analytic review found
only weak effects of solitary confinement on inmate outcomes (most of which
were psychological). The team concluded that their meta-analytic review did not
find support for the long-argued contention that solitary confinement has lasting
psychological effects on those subjected to it (Smith et al., 2015). Significantly,
the studies that employed weaker research designs produced stronger effects than
those employing more rigorous research designs (Gendreau & Labrecque, 2015).
A second meta-analysis (Morgan et al., 2016)reported similar results, leading
Labrecque (2015) to conclude that the findings from recent meta-analyses “cast
some doubts about [solitary confinement] being as devastating to inmates as
has often been portrayed in the media and by some human rights organizations,
activists, and scholars who vehemently oppose the practice on moral/ethical
grounds,” adding, “these findings serve as a caution to reviewers about
making judgments regarding the effects of [solitary confinement] too hastily,
especially when they are based on qualitative rather than quantitative evidence”
(Labrecque, 2015, p. 6).

The Future of Administrative Segregation
It may come as a surprise that the research described in the previous three
sections of this paper represents the bulk of the published empirical research
related to the use of administrative segregation in U.S. prisons. The research
findings can be described in a relatively short white paper, which suggests that
considerable room exists for further research in this area.
At least three distinct perspectives emerge from a thorough review of the
literature. Some researchers strongly believe that segregation, with its focus on
isolation through solitary or near-solitary confinement, is incredibly damaging,
and not only exacerbates inmates’ existing mental health problems but also
may create mental illness where there previously was none. These scholars
find the practice to be morally and ethically objectionable, argue that it causes
excessive harm, and frequently call for its immediate and absolute abolition.
Most evidence supporting this perspective has emerged from research
involving in-depth case studies of a few inmates incarcerated for long periods,
usually in supermax settings.
Correctional administrators charged with day-to-day prison operations represent
a second perspective. Tasked with ensuring safety and security for all of those

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 27

who live or work in correctional facilities (as well as contributing to the public
safety mission more broadly), these experts tend to argue that restrictive
housing is necessary for the safety and security of the entire correctional system.
Although the evidence supporting enhanced institutional or systemwide
safety and security as a result of the proliferation of restrictive housing units
is thin, many correctional administrators strongly believe that segregation has
had positive effects. The public tends to agree. Recently, as states have begun
looking for ways to reduce reliance on incarceration more generally, correctional
administrators have begun taking a critical look at their segregation policies
and are increasingly seeking ways to safely reduce the use of administrative
segregation across their systems as well.
A third group of correctional experts argues that we do not know enough about
the utility and effects of administrative segregation to conclude with any degree
of confidence that it is either a harmful or necessary approach. These scholars
typically point to the lack of an evidence base, noting that contradictory findings
can largely be explained by differences in methods. They tend to argue that the
most analytically sophisticated studies, although certainly not perfect, have failed
to document damaging long-term psychological effects or worse behavioral
outcomes among those exposed to restrictive housing in all its varieties. It
is crucial to note that these scholars do not, therefore, argue that solitary
confinement through administrative or disciplinary segregation is sound, wise, or
worthwhile correctional policy. Indeed, they argue that evidence that the practice
achieves its intended or stated goals is equally lacking.
Although all three perspectives were given consideration, this paper focused on
describing the empirical evidence, regardless of the perspective from which the
evidence emerged. To be clear, the authors of this paper were not asked to draw
conclusions about the harmfulness of the practice of solitary confinement; we
were asked to review and describe the research on the practice of administrative
segregation in restrictive housing units across the United States more broadly.
Throughout this white paper, the authors have remained cognizant of the
numerous types of confinement that fall under the administrative segregation
umbrella and have not focused exclusively on the most extreme versions of the
practice. After a thorough review of the extant literature, it is clear that even
today, the questions continue to be many and the answers few. It is equally clear
that when researchers have disagreed — and in this area, they have tended to
disagree passionately — they have not always been speaking the same language
or conducting research with equivalent populations. Moreover, for many
researchers studying solitary confinement, the practice raises not only empirical
questions but also moral and ethical concerns that will persist regardless of the
breadth or depth of the evidence base. Across a body of literature replete with
highly charged emotions, interpreting the evidence and separating research
evidence from strongly held beliefs have become exceptionally difficult.

28 • National Institute of Justice | NIJ.gov

This white paper turns now to some recommendations for future research. These
recommendations are offered not to provide definitive answers but to launch an
important conversation about the future of research on segregation in restrictive
housing units. Developing an agenda for future research is important, given that
the practice is increasingly facing scrutiny on the national and international stages.

What We Know — The Empirical Evidence
This review of the empirical evidence reinforces what many have been arguing
for many years: We know surprisingly little about the use of solitary confinement
and its effects, particularly given the speed and extent to which the practice of
segregating those deemed to be the worst of the worst in restrictive housing units
has proliferated across correctional systems.12 Indeed, the most concise summary
would likely say that, although the jury is still out, there may be some potentially
devastating psychological effects on inmates in restrictive housing (especially for
those who already suffer from mental illness and for those who are subjected to
lengthy or indefinite terms of confinement in administrative segregation), but
that most research suggests that fewer negative effects of solitary confinement
have been demonstrated when the term of confinement is relatively short.
Critically, almost no literature documents the utility of the practice or
demonstrates that the use of these administrative segregation has demonstrably
achieved specific aims (Mears & Watson, 2006; Mears 2008). It is equally
important to note that despite a growing body of literature on correctional officer
stress and wellness, we know almost nothing about the potentially differential
effects of working in restrictive housing units on health and safety outcomes for
correctional employees.

What We Still Don’t Know — Gaps in the Knowledge Base
There are relatively few well-designed quantitative studies of the effects of
administrative segregation, and those have tended to produce either null or
inconclusive findings. It is difficult to design and carry out empirically sound
research in the restrictive housing environment — the basic issues of access and
feasibility are difficult to overcome in this context. In an era of evidence-based
policy and practice, where the gold standard for validating a practice would be the
implementation of an experimental design, no existing evidence-based research

12

Although evidence of psychological effects is limited, the lack of a solid, empirical evidence base leads even
the most highly respected scholars to make assertions based on anecdotal evidence, with many agreeing that
the experience is likely psychologically damaging in profound ways. For example, after reviewing the research,
Kurki and Morris (2001) acknowledged that, “[a]lthough hard data and controlled clinical studies are lacking, we
find it difficult not to believe that prolonged supermax conditions would cause serious psychological and social
problems for anyone, whether mentally strong, weak, or something between” (p. 415).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 29

reaches any definitive conclusions about the wisdom or utility of the practice of
solitary confinement through administrative (or disciplinary) segregation.
Therefore, the gaps in the knowledge base are many. Most agree, at least in
principle, that long-term solitary confinement in administrative segregation
— for periods of years or decades — is likely detrimental to the individual,
but it is not as clear that short-term solitary confinement through disciplinary
or administrative segregation is equally detrimental. The qualitative research
tells us that long-term segregation in solitary confinement seems to have some
profound psychological effects, particularly on inmates with mental illness, but
there is far less evidence that short-term segregation in solitary confinement has
pronounced or lasting negative effects, either psychologically or behaviorally.
We do not know how general insights that can be gleaned from the literature
might apply to different types of inmates across different contexts. Do inmates
placed in solitary confinement for their protection fare better or worse than those
placed there for the protection of others? Growing evidence seems to show that
those who enter solitary confinement with a serious mental illness do not fare
well, but are those who are not suffering from mental illnesses equally likely to
deteriorate under such conditions, as some contend? Is it only extended isolation
that results in negative effects? Are shorter periods of solitary confinement not
damaging and, perhaps, even beneficial? If differential effects exist, how long is
too long? Can some inmates develop survival strategies that might help them
become more resilient, mitigating the potentially damaging effects?13 The list of
potential questions is endless. The key to the future of research on administrative
segregation involves generating agreement on the right questions to ask and the
most appropriate methods for answering those questions.

Future Directions — Research and Funding Priorities
Given that some inmates spend months, years, or even decades in administrative
segregation, it is shocking how little we know about the contemporary use of this
practice. The solitary confinement model associated with the early American
penitentiaries was quickly abandoned, in part, as a result of concerns about the
potential effects on inmates (but also, in no small part, to the far lower cost of
keeping prison populations in congregate settings). Although concerns about
costs and effects persist, the most that can be said is that researchers disagree
about the effects of administrative segregation; they disagree about effects on
violence levels within institutions and across correctional systems, on individuals
psychologically, and on inmate outcomes behaviorally.

13

See O’Donnell (2014, 2016).

30 • National Institute of Justice | NIJ.gov

Establish agreed-upon definitions
With increasing calls for a hard look at solitary confinement and segregation
in restrictive housing, initial steps toward formulating a research agenda for
the future should, at a minimum, include coming to some agreement on terms
and definitions. It is crucial that the field settle on generally agreed-upon terms
and definitions so that scholars and practitioners can speak to each other in
a common language. A review of the voluminous literature makes it clear
that many of the apparent contradictions in the literature can be attributed
to differences in what is being characterized, described, and counted as
administrative segregation. Rarely are those with opposing viewpoints describing
the same thing.
Scholars almost exclusively refer to solitary confinement and use the term
“supermax incarceration” as shorthand to characterize the restrictive housing
units that use solitary confinement. Practitioners, including those who run
supermax facilities, rarely use the term “solitary confinement” in either official
policy or day-to-day practice, preferring to refer to the types of units that
typically involve 22- to 24-hour isolation. In the correctional world, however,
the term “administrative segregation” is sometimes used as an umbrella that
covers many types of confinement — some of which bear little resemblance to
the solitary confinement that is generally described by scholars who conduct
research in this area. These issues are not trivial. We invite those who run
correctional facilities, and those who fund or conduct correctional research, to
begin a dialogue on how best to characterize and distinguish the various types of
confinement in restrictive housing units, and to develop definitions that would
distinguish practices from units and allow for a more refined understanding of
the effects of varied correctional practices across contexts.

Collect and analyze data to establish reliable prevalence estimates
Once definitions are formulated, BJS might be encouraged to begin routinely
collecting data on the prevalence of confinement in the various types of
restrictive housing around the country — whether through its annual effort to
produce statistics related to the prevalence of incarceration or through a separate
program dedicated to the collection of data related to restrictive housing. State
correctional systems should be encouraged to develop data systems that could
track and distinguish placements in disciplinary segregation, administrative
segregation, and protective custody. Despite a few recent reports offering some
baseline estimates, we know little about the prevalence of any of these three types
of confinement. Consider, for example, the federal system. Most in the field are
familiar with the ADX-Florence facility, the federal supermax facility that houses
some of the most notorious prisoners in the federal system. Through a review
of the two recent reports on administrative segregation in the federal system, it
becomes clear that inmates being held in ADX-Florence, while likely spending

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 31

the most time in solitary confinement, hardly represent the bulk of inmates
serving time in highly restrictive administrative segregation settings (McGinnis
et al., 2014; United States Government Accountability Office, 2013).

Distinguish differential effects of short-term versus long-term exposure to
solitary confinement
Debates over the effects of administrative segregation tend to be so divisive that
it is difficult to find an objective assessment of the evidence. A hard look at the
empirical evidence, though, makes clear from the limited studies conducted to
date that there is little good empirical evidence that time spent in isolation has
demonstrable negative effects on psychological or behavioral outcomes for most
inmates subjected to it. As this debate continues, however, a crucial distinction
must be drawn between time-limited segregation imposed for an infraction or
series of infractions (typically referred to as disciplinary segregation) and long­
term segregation for management of prison populations (typically referred to as
either supermax incarceration or administrative segregation).
It seems imperative that future research distinguishes the findings related
to the effects of short-term solitary confinement (as Morris, 2015, has done
recently) from those related to long-term solitary confinement in supermax
units or facilities (as Haney and Grassian have done for many years). There are
both qualitative and quantitative differences between incarceration in a cell for
23 hours per day for one to 90 days and being held under such conditions for
months and years on end. Those who decry the inhumane character of solitary
confinement typically point to the latter. They present devastating portraits
of the relentless anguish and serious deterioration suffered by some inmates
incarcerated for years on end in supermax facilities. However, those effects might
not be representative — and certainly might not be reflective — of those confined
for short periods as a more immediate behavioral management approach. That is
not to say that the short-term solitary confinement of inmates is not harmful, but
if it is harmful, we do not know to whom it is harmful or the circumstances in
which it is harmful. Little work has been done in this area, especially for distinct
populations or in ways that would allow us to assess differences, controlling
for what might be key explanatory variables such as variations in time spent in
solitary confinement or variations in the conditions of the confinement.

Establish standards for research access to populations in segregated housing units
This paper would be remiss if it did not note that a major reason for the lack of
an evidence base related to solitary confinement is issues of access. Research
within prisons is notoriously difficult. Until relationships are formed and trust is
established between an external research team and correctional administrators,
independent research is difficult to pursue. Much of what we know about
supermax prisons and administrative segregation, therefore, comes from official

32 • National Institute of Justice | NIJ.gov

agency reports and statistics or research internal to corrections departments. If
access to general prison populations has proven difficult for all but a few, access
to those in isolation is almost, by definition, impossible. Inmates isolated in
administrative segregation units or supermax facilities are permitted little contact
with the inside world, let alone with the outside one.14
Without question, the use of administrative segregation has expanded, with
limited evidence of its impacts on the operation of prisons or on the inmates
subjected to it. Many researchers have argued that, given the extreme conditions
sometimes associated with administrative segregation as it is currently
practiced, the onus is on those advocating for the maintenance or expansion of
administrative segregation and on those running such restrictive correctional
units and facilities to prove their benefit, especially given their substantially
greater cost. The authors of this paper tend to agree.
If we are to learn more about administrative segregation, whether it be its
costs, benefits, or impacts, it is incumbent on the research-funding agencies to
provide financial support for what will be costly research to conduct, and on the
departments of corrections around the country to permit the access that would
facilitate such research. Doing so will require allowing independent researchers
access not only to data pertaining to use of these units but also unfettered access
to the inmates being housed within them. This is easier said than done. There are
legitimate safety and security concerns that will likely be raised in the context of
facilitating such access; nonetheless, an evidence base will never be established in
the absence of empirically sound research that requires access.

Prioritize funding for research that can overcome the methods’ shortcomings
Access is not the only problem. The only way to escape the criticism that is sure
to face research on controversial issues is to conduct research that gets as close
as possible to experimental designs. Unfortunately, experimental designs are
hard to come by and almost impossible to carry out in correctional settings, for
both practical and ethical reasons (Clear, 2010). Several more recent quantitative
studies of administrative segregation have employed PSM because it is not
possible, nor would it be ethical, to assign inmates randomly to administrative
segregation. Studies that have used PSM have demonstrated that effects that exist
with unmatched samples (e.g., prior to creating a matched comparison group)
tend to diminish (and even disappear) when the potential selection effects are
controlled for through PSM (Lovell et al., 2007; Mears & Bales, 2009; Morris,
2015). Therefore, it is highly recommended that funding be reserved for research
that can control for selection effects.

14

As Kurki and Morris (2001) concluded at the end of their review of the thin research base for confinement in
supermax prisons, “like so much else about a supermax prison, the walls of exclusion of knowledge are here,
too, so much higher” (p. 418).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 33

Evaluation research, particularly research that includes cost-benefit analyses,
should also be prioritized. In 2008, when Mears applied an evaluation research
framework to the emergence and growth in the use of supermax prisons, he
concluded that there was minimal evidence that such facilities were necessary,
no evidence that they were designed on a sound theoretical base or were costeffective, and minimal evidence that they were implemented in a consistent,
principled manner or achieved their intended goals. In other words, the evidence
was scant. His summary assessment remains relevant seven years later.

Conclusion
Few researchers would question that some prisoners being held in isolation are
exceptionally dangerous and violent and might require some type of segregation.
At the same time, few researchers would believe that all (or even most) of
those held in isolation require the type of solitary confinement that is typical of
such settings, especially for extended periods. Many have more fundamentally
questioned whether administrative segregation requires the extreme isolation
and sensory-deprivation characteristics of some of these environments. Virtually
all agree that any harm associated with extended solitary confinement could and
should be avoided.
Notwithstanding the many gaps in the research base, the most important
research going forward will be that which can lead to a substantial reduction
in the need for solitary confinement through administrative segregation. It is
incumbent upon researchers and correctional administrators to work together to
identify viable alternatives that can ensure institutional and public safety without
compromising the occupational well-being of correctional employees or the
psychological well-being of inmates in the care of departments of corrections.

3 4 • National Institute of Justice | NIJ.gov

References
American Civil Liberties Union. (2014). The dangerous overuse of solitary
confinement in the United States. New York: ACLU Foundation.
Amnesty International. (2012). Cruel isolation: Amnesty International’s concerns
about conditions in Arizona maximum security prisons. London, England:
Amnesty International.
Arrigo, B. A., & Bullock, J. L. (2008). The psychological effects of solitary
confinement on prisoners in supermax units: Reviewing what we know and
recommending what should change. International Journal of Offender Therapy
and Comparative Criminology, 52(6), 622-640.
Baker, P., & Goode, E. (2015, July 21). Critics of solitary confinement are buoyed
as Obama embraces their cause. New York Times. Retrieved from http://www.
nytimes.com/2015/07/22/us/politics/critics-of-solitary-confinement-buoyed-as­
obama-embraces-cause.html.
Beck, A. J. (2015). Use of restrictive housing in U.S. prisons and jails, 2011–12.
Washington, DC: U.S. Department of Justice, Bureau of Justice Statistics.
Bonta, J., & Gendreau, P. (1990). Reexamining the cruel and unusual punishment
of prison life. Law and Human Behavior, 13(4), 347-372.
Bosworth, M. (Ed.). (2005). Encyclopedia of Prisons & Correctional Facilities.
Thousand Oaks, CA: Sage Publications.
Briggs, C. S., Sundt, J. L., & Castellano, T. C. (2003). The effect of supermaximum
security prisons on aggregate levels of insitutional violence. Criminology, 41(4),
1341-1376.
Browne, A., Cambier, A., & Agha, S. (2011). Prisons within prisons: The use of
segregation in the United States. Federal Sentencing Reporter, 24(1), 46-49.
Bulman, P., Garcia, M., & Hernon, J. (2012). Study raises questions about
psychological effects of solitary confinement. NIJ Journal. Washington, DC:
National Institute of Justice.
Butler, B., & Simpson, M. (2015). A solitary failure: The waste, cost, and harm of
solitary confinement in Texas. Houston, TX: American Civil Liberties Union of
Texas & The Texas Civil Rights Project-Houston.
Butler, H. D., Griffin, O. H., & Johnson, W. W. (2013). What makes you the
“worst of the worst?” An examination of state policies defining supermaximum
confinement. Criminal Justice Policy Review, 24(6), 676-694.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 35

Clear, T. R. (2010). Policy and evidence: The challenge to the American Society of
Criminology. 2009 Presidential Address to the American Society of Criminology.
Criminology, 48(1), 1-25.
Cloyes, K. G., Lovell, D., Allen, D. G., & Rhodes, L. A. (2006). Assessment of
psychosocial impairment in a supermaximum security unit sample. Criminal
Justice and Behavior, 33(6), 760-781.
Collins, W. C. (2004). Supermax prisons and the Constitution: Liability concerns in
the extended control unit. Washington, DC: National Institute of Corrections.
Crouch, B. M., & Marquart, J. W. (1989). An Appeal to justice: Litigated reform of
Texas prisons. Austin, TX: University of Texas Press.
Crouch, B. M., & Marquart, J. W. (1990). Resolving the paradox of reform: Litigation,
prisoner violence, and perceptions of risk. Justice Quarterly, 7(1), 103-123.
Ditton, P. (1999). Mental health and treatment of inmates and probationers.
Washington, DC: U.S. Department of Justice.
Fathi, D. C. (2004). The common law of supermax litigation. Pace Law Review,
24(2), 675-690.
Feeley, M. M., & Rubin, E. L. (1999). Judicial policy making and the modern state:
How the courts reformed America’s prisons. New York: Cambridge University Press.
Fellner, J., & Mariner, J. (1997). Cold storage: Super-maximum security
confinement in Indiana (Vol. 2156). New York: Human Rights Watch.
Gendreau, P., & Labrecque, R. M. (2015). The effects of administrative
segregation: A lesson in knowledge cumulation. In J. Wooldredge & P. Smith
(Eds.), Oxford handbook on prisons and imprisonment. New York: Oxford
University Press.
Gendreau, P., & Theriault, Y. (2011). Bibliotherapy for cynics revisited:
Commentary on “One Year Longitudinal Study of Psychological Effects of
Administrative Segregation.” Corrections and Mental Health: An Update of the
National Institute of Corrections.
Gibbons, J. J., & Katzenbach, N. de B. (2006). Confronting confinement: A report
of the Commission on Safety and Abuse in America’s Prisons. New York: Vera
Institute of Justice.
Goode, E. (2014, March 16). After 20 hours in solitary, Colorado’s prisons chief
wins praise. New York Times, p. A16.
Grassian, S. (1983). Psychopathological effects of solitary confinement. American
Journal of Psychiatry, 140(11), 1450-1454.

36 • National Institute of Justice | NIJ.gov

Grassian, S., & Friedman. (1986). Effects of sensory deprivation in psychiatric
seclusion and solitary confinement. International Journal of Law and Psychiatry,
8(1), 49-65.
Grassian, S., & Kupers, T. (2011). The Colorado Study vs. the reality of solitary
confinement. Correctional Mental Health Report, 13(1), 1-4.
Griffin, M. L., & Hepburn, J. R. (2006). The effect of gang affiliation on violent
misconduct among inmates during the early years of confinement. Criminal
Justice and Behavior, 33(4), 419-466.
Hager, E., & Rich, G. (2014, December 23). Shifting away from solitary. Retrieved
from https://www.themarshallproject.org/2014/12/23/shifting-away-from­
solitary.
Hananel, S. (2015, August 11). Justice Kennedy provides hope to solitary
confinement foes. Retrieved from www.businessinsider.com/ap-justice-kennedy­
provides-hope-to-solitary-confinement-foes-2015-8.
Haney, C. (2002). The Psychological impact of incarceration: Implications
for post-prison adjustment. Paper presented at the From Prison to Home
Conference, Washington, DC.
Haney, C. (2003). Mental health issues in long-term solitary and “supermax”
confinement. Crime and Delinquency, 49(1), 124-156.
Haney, C. (2008). A culture of harm: Taming the dynamics of cruelty in
supermax prisons. Criminal Justice and Behavior, 35(8), 956-984.
Haney, C., & Lynch, M. (1997). Regulating prisons of the future: The
psychological consequences of supermax and solitary confinement. New York
University Review of Law and Social Change, 23, 477-570.
Henningsen, R. J., Johnson, W. W., & Wells, T. (1999). Supermax prisons:
Panacea or desperation? Corrections Management Quarterly, 3, 53-59.
Hershberger, G. (1998). To the max. Corrections Today, 59(1), 54-56.
Huebner, B. M. (2003). Administrative determinants of inmate violence: A multilevel
analysis. Journal of Criminal Justice, 31, 107-117.
Human Rights Watch. (2000). Out of sight: Supermaximum security confinement
in the United States. New York: Human Rights Watch.
Human Rights Watch & American Civil Liberties Union. (2012). Growing up
locked down: Youth in solitary confinement in jails and prisons across the United
States. New York: Human Rights Watch.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 37

James, D. J., & Glaze, L. E. (2006). Mental health problems of prison and jail
inmates. Washington, DC: Bureau of Justice Statistics.
King, K., Steiner, B., & Breach, S. R. (2008). Violence in the supermax: A self-fulfilling
prophecy. The Prison Journal, 88(1), 144-168.
King, R. D. (1991). Maximum security custody in Britain and the USA: A study
of Gartree and Oak Park Heights. British Journal of Criminology, 31(2), 126-152.
King, R. D. (1999). The rise and rise of supermax: An American solution in
search of a problem? Punishment & Society, 1(2), 163-186.
Kupers, T. A., Dronet, T., Winter, M., Austin, J., Kelly, L., Cartier, W., ... McBride,
J. (2009). Beyond supermax administrative segregation: Mississippi’s experience
rethinking prison classification and creating alternative mental health programs.
Criminal Justice and Behavior, 36(10), 1037-1050.
Kurki, L., & Morris, N. (2001). The purposes, practices, and problems of
supermax prisons. Crime and Justice: A Review of Research, 28, 385-424.
Labrecque, R. M. (2015). The effect of solitary confinement on institutional misconduct:
A longitudinal evaluation. Washington, DC: U.S. Department of Justice.
Labrecque, R. M., & Smith, P. (2013). Advancing the study of solitary confinement.
In J. Fuhrmann & S. Baier (Eds.), Prisons and prison systems: Practices, types, and
challenges (pp. 57-70). Hauppage, NY: Nova Science Publishers.
Lanes, E. C. (2011). Are the “worst of the worst” self-injurious prisoners more
likely to end up in long-term maximum-security administrative segregation?
International Journal of Offender Therapy and Comparative Criminology, 55(7),
1034-1050.
Liman Program & Association of State Correctional Administrators. (2015).
Time-In-cell: ASCA-Liman 2014 National Survey of Administrative Segregation in
Prison. New Haven, CT: Yale Law School.
Liptak, A. (2015, June 19). Supreme Court permits new hearing for mentally
disabled inmate. New York Times, p. A13.
Lovell, D. (2008). Patterns of disturbed behavior in a supermax population.
Criminal Justice and Behavior, 35(8), 985-1004.
Lovell, D., Cloyes, K., Allen, D., & Rhodes, L. (2000). Who lives in supermaximum custody? A Washington State study. Federal Probation, 64(2), 33-38.
Lovell, D., Johnson, L. C., & Cain, K. C. (2007). Recidivism of supermax
prisoners in Washington State. Crime & Delinquency, 53(4), 633-656.

38 • National Institute of Justice | NIJ.gov

McGinnis, K., Austin, J., Becker, K., Fields, L., Lane, M., Maloney, M., ..., Felix, T.
(2014). Federal Bureau of Prisons: Special Housing Unit Review and Assessment.
Arlington, VA: CNA.
Mears, D. P. (2008). An assessment of supermax prisons using an evaluation
research framework. The Prison Journal, 88(1), 43-68.
Mears, D. P., & Bales, W. D. (2009). Supermax incarceration and recidivism.
Criminology, 47(4), 1131-1166.
Mears, D. P., & Castro, J. L. (2006). Wardens’ views on the wisdom of supermax
prisons. Crime & Delinquency, 52(3), 398-431.
Mears, D. P., Mancini, C., Beaver, K. M., & Gertz, M. (2013). Housing for the
“worst of the worst” inmates: Public support for supermax prisons. Crime &
Delinquency, 59(4), 587-615.
Mears, D. P., & Watson, J. (2006). Towards a fair and balanced assessment
of supermax prisons. Justice Quarterly, 23(2), 232-270.
Metcalf, H., Morgan, J., Oliker-Friedland, S., Resnik, J., Spiegel, J., Tae, H.,
... Holbrook, B. (2013). Administrative segregation, degrees of isolation,
and incarceration: A national overview of state and federal correctional policies.
Hartford, CT: Liman Public Interest Program, Yale Law School.
Morgan, R. D., Gendreau, P., Smith, P., Gray, A. L., LaBrecque, R. M., MacLean,
N., Van Horn, S. A., Bolanos, A. D., Batastini, A. B., Mills, J. F. (2016, Aug 8).
Quantitative syntheses of the effects of administrative segregation on inmates’
well-being. Psychology, Public Policy, and Law. Advance online publication.
http://dx.doi.org/10.1037/law0000089.
Morris, R. G. (2016). Exploring the effect of exposure to short-term solitary
confinement among violent prison inmates. Journal of Quantitative Criminology,
32(1), 1-22.
Naday, A., Freilich, J. D., & Mellow, J. (2008). The elusive data on supermax
confinement. The Prison Journal, 88(1), 69-93.
National Institute of Corrections. (1997). Supermax housing: A survey of current
practice. Washington, DC: National Institute of Corrections.
O’Donnell, I. (2014). Prisoners, solitude, and time. Oxford, U.K.: Oxford
University Press.
O’Donnell, I. (2016). The survival secrets of solitaries. The Psychologist, 29(3),
184-187.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 39

O’Keefe, M. L. (2007). Administrative segregation for mentally ill inmates.
Journal of Offender Rehabilitation, 45(1-2), 149-165.
O’Keefe, M. L. (2008). Administrative segregation from within: A corrections
perspective. The Prison Journal, 88(1), 123-143.
O’Keefe, M. L., Klebe, K. J., Stucker, A., Sturm, K., & Leggett, W. (2011). One year
longitudinal study of the psychological effects of administrative segregation. Final
report to the U.S. Department of Justice. Washington, DC: U.S. Department of
Justice.
Olivero, J. M., & Roberts, J. B. (1987). Marion Federal Penitentiary and the
22-month lockdown: The crisis continues. Crime and Social Justice, 27/28, 234-255.
Osher, F., D’Amora, D. A., Plotkin, M., Jarrett, N., & Eggleston, A. (2012).
Adults with behavioral needs under correctional supervision: A shared
framework for reducing recidivism and promoting recovery. New York: Council
of State Governments.
Pizarro, J. M., & Narag, R. E. (2008). Supermax prisons: What we know, what we
do not know, and where we are going. The Prison Journal, 88(1), 23-42.
Pizarro, J. M., & Stenius, V. M. K. (2004). Supermax prisons: Their rise, current
practices, and effect on inmates. The Prison Journal, 84(2), 248-264.
Pizarro, J. M., Stenius, V. M. K., & Pratt, T. C. (2006). Supermax prisons: Myths,
realities, and the politics of punishment in American society. Criminal Justice
Policy Review, 17(1), 6-21.
Pizarro, J. M., Zgoba, K. M., & Haugebrook, S. (2014). Supermax and recidivism:
An examination of the recidivism covariates among a sample of supermax exinmates. The Prison Journal, 94(2), 180-197.
Raemisch, R. (2014, February 21). My night in solitary. New Tork Times, p. A25.
Ralph, P. H., & Marquart, J. W. (1991). Gang violence in Texas prisons. The
Prison Journal, 71(2), 38-49.
Rhodes, L. A., & Lovell, D. (2011). Is “adaptation” the right question? Addressing
the larger context of administrative segregation. Corrections and Mental Health:
An Update of the National Institute of Corrections.
Richards, S. C. (2008). USP Marion: The first federal supermax. The Prison
Journal, 88(1), 6-22.
Riveland, C. (1999). Supermax prisons: Overview and general considerations.
Washington, DC: National Institute of Corrections.

4 0 • National Institute of Justice | NIJ.gov

Romano, S. M. (1996). If the SHU fits: Cruel and unusual punishment at
California’s Pelican Bay State Prison. Emory Law Journal, 45, 1089.
Rothman, D. J. (1971/1990). The discovery of the asylum: Social order and disorder
in the new republic (Rev. ed.). Boston: Little, Brown, and Company.
Sedlak, A. J., & McPherson, K. S. (2010, May). Conditions of confinement:
Findings from the Survey of Youth in Residential Placement. Bulletin.
Washington, DC: U.S. Department of Justice, Office of Justice Programs,
Office of Juvenile Justice and Delinquency Prevention.
Shalev, S. (2008). A sourcebook on solitary confinement. London, England:
London School of Economics, Mannheim Centre for Criminology.
Shalev, S., & Lloyd, M. (2011). Though this be method, yet there is madness in’t:
Commentary on One Year Longitudinal Study of the Psychological Effects of
Administrative Segregation. Corrections and Mental Health: An Update of the
National Institute of Corrections.
Shames, A., Wilcox, J., & Subramanian, R. (2015, May). Solitary confinement:
Common misconceptions and emerging safe alternatives. New York: Vera Institute
of Justice.
Simon, J. (2014). Mass incarceration on trial: A remarkable court decision and the
future of prisons in America. New York: The New Press.
Smith, P., Gendreau, P., & Labrecque, R. M. (2015). The impact of
solitary confinement on inmate behavior: A meta-analytic review. Paper
presented at the North American Correctional and Criminal Justice
Psychology (N3) Conference, Ottawa, Canada. http://media.wix.com/
ugd/7fc458_2efc5654e7ea4d27a9d45a64f331fec5.pdf.
Smith, P. S. (2006). The effects of solitary confinement on prison inmates: A brief
history and review of the literature. Crime and Justice, 34(1), 441-528.
Solitary Watch. (2015). Testimony. Retrieved September 19, 2015, from
http://solitarywatch.com/resources/testimony/.
Sorenson, J. R., Cunningham, M. D., Vigen, M. P., & Woods, S. O. (2011). Serious
assaults on prison staff: A descriptive analysis. Journal of Criminal Justice, 39,
143-150.
St. John, P. (2015, September 1). California agrees to move thousands of inmates
out of solitary confinement. Los Angeles Times. Retrieved from http://www.
latimes.com/local/lanow/la-me-ln-california-will-move-thousands-of-inmates­
out-of-solitary-20150901-story.html.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 41

Suedfeld, P., Ramirez, C., Deaton, J., & Baker-Brown, G. (1982). Reactions and
attributes of prisoners in solitary confinement. Criminal Justice and Behavior,
9(3), 303-340.
Suedfeld, P., & Roy, C. (1975). Using social isolation to change the behaviour of
disruptive inmates. International Journal of Offender Therapy and Comparative
Criminology, 19(1), 90-99.
Sundt, J. L., Castellano, T. C., & Briggs, C. S. (2008). The sociopolitical context of
prison violence and its control: A case study of supermax and its effect in Illinois.
The Prison Journal, 88(1), 94-122.
Therolf, G. (2015, May 28). Advocates seek to end solitary confinement options
for young offenders. Los Angeles Times. Retrieved from http://www.latimes.com/
local/crime/la-me-solitary-juvenile-20150528-story.html.
Toch, H. (2001). The future of supermax confinement. The Prison Journal, 81(3),
376-388.
United Nations. (2015a). “Mandela Rules” passed, standards on the treatment of
prisoners enhanced for the 21st century. Press release. New York: United Nations
Office of Drugs and Crime.
United Nations. (2015b). United Nations standard minimum rules for the
treatment of prisoners (the Mandela Rules). New York: United Nations. Retrieved
from http://www.penalreform.org/wp-content/uploads/2015/05/MANDELARULES.pdf.
United States Government Accountability Office. (2013). Improvements needed
in Bureau of Prisons’ monitoring and evaluation of impact of segregated housing.
Washington, DC: United States Government Accountability Office.
Wallace, R. (2013). Out of sight, out of mind: Colorado’s continued warehousing
of mentally ill prisoners in solitary confinement. Denver, CO: American Civil
Liberties Union of Colorado.
Ward, D. A., & Werlich, T. G. (2003). Alcatraz and Marion: Evaluating supermaximum custody. Punishment & Society, 5(1), 53-75.
Weiss, C., Kraner, N. J., & Fisch, J. (2013). 51-jurisdiction survey of solitary
confinement rules in juvenile justice systems. New York: Lowenstein Center for the
Public Interest.
White House Office of the Press Secretary. (2015). Remarks by the President
at the NAACP conference. Washington, DC: The White House, Office of the
Press Secretary.

42 • National Institute of Justice | NIJ.gov

Winerip, M., & Schwirtz, M. (2015, January 14). Rikers setting stricter limits for
isolation. New York Times, p. A1.
Wooldredge, J., & Steiner, B. (2015). A macro-level perspective on prison inmate
deviance. Punishment & Society, 17(2), 230-257.
Wolff, N., Blitz, C. L., Shi, J., Siegel, J., & Bachman, R. (2007). Physical violence inside
prisons: Rates of victimization. Criminal Justice and Behavior, 34(5), 588-599.
Zinger, I., Wichmann, C., & Andrews, D. A. (2001). The psychological effects of
60 days in administrative segregation. Canadian Journal of Criminology 43(1)
(January), 47-83.

Cases Cited
Bruscino v. Carlson, 854 F.2d 162 (7th Cir. 1988).
Dixon v. Goord, 224 F. Supp. 2d 739 (S.D. N.Y. 2002).
Madrid v. Gomez, 889 F. Supp. 1146 (N.D. Cal. 1995).
Ruiz v. Estelle, 503 F. Supp. 1265 (S.D. Tex. 1980).
Ruiz v. Johnson, 37 F. Supp. 2d 855 (S.D. Tex. 1999).
Wolff v. McDonnell, 418 U.S. 539 (1974).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 4 3

APPENDIX TABLE A1: Administrative Segregation in the Federal Bureau of Prisons (BOP)
Special Housing Unit
(SHU)

Special Management Unit
(SMU)

Administrative
Maximum (ADX)

Referral:

Inmates placed in
SHU are in either
administrative
detention (AD)
or disciplinary
segregation (DS).
AD is intended to
be temporary and
nonpunitive. DS is the
possible sanction for
inmates who violate
the rules. Length
of stay is based on
severity of offense.

BOP may consider
designating an inmate to an
SMU who (1) participated
in or had a leadership role
in disruptive geographical
group/gang-related activity;
(2) has a history of serious
disruptive disciplinary
infractions; or (3) committed
any greatest severity-level
prohibited acts after being
classified a member of a
disruptive group, among
other reasons.

Inmates whose
placement in another
facility poses a risk to
the safety of inmates,
staff, or the public
or good order of the
facility and/or inmates
whose status before
or after incarceration
does not allow them
to be safely housed in
another facility.

Population:

• Number of cells:
7,381
• Population: 10,050
• 5.7% of BOP
inmates

• Number of cells: 1,270
• Population: 1,960
• 1.1% of BOP inmates

• Number of cells:
623
• Population: 450
• 0.3% of BOP
inmates

Confinement
conditions:

• Mostly doublebunked.
• 5 hours per
week out-of-cell
exercise.
• May shower/
shave at least 3
times per week.
• Minimum 1
completed call per
month.
• Minimum 4 hours
of visitation per
month.
• Inmates eat all
meals inside cells.

• Conditions are to be made
less restrictive when an
inmate progresses from
level 1 to level 4.
• Mostly double-bunked.
• 5 hours per week out-of­
cell exercise.
• May shower/shave at
least 3 times per week.
• Minimum 2 completed
calls per month.
• Mail and telephone calls
subject to monitoring for
intelligence purposes.
• Minimum 4 hours of
visitation per month.
• Inmates eat all meals
inside cells.

• Nearly all single cells.
• Inmates eat all meals
inside cells.

Compiled from Figures 1 and 2 in the GAO report on segregation in the BOP (United States Government Accountability
Office, 2013, pp. 7, 9).

4 4 • National Institute of Justice | NIJ.gov

APPENDIX TABLE A2: Percentage of Custodial Population (Both Sexes) in Administrative
Segregation (Ad Seg) and Restrictive Housing (Fall 2014) (n = 34) (continues next page)
Total

Ad Seg

% of Total

All Restrictive
Housing

% of Total

Alabama

24,862

729

2.90%

1,253

5.00%

BOP

171,868

1,656

1.00%

11,387

6.60%

Colorado

20,944

207

1.00%

662

3.20%

Connecticut

16,564

74

0.40%

592

3.60%

Delaware

5,977

330

5.50%

847

14.20%

D.C.

2,067

62

3.00%

174

8.40%

Florida

100,869

2,416

2.40%

8,936

8.90%

Georgia

52,959

1,625

3.10%

1,658

3.10%

Indiana

28,318

692

2.40%

1,789

6.30%

Iowa

8,172

142

1.70%

542

6.60%

Kansas

9,529

557

5.90%

664

7.00%

Kentucky

12,103

794

6.60%

794

6.60%

Massachusetts

10,475

313

3.00%

518

4.90%

Michigan

44,925

1,122

2.50%

2,004

4.50%

Missouri

31,945

1,277

4.00%

3,929

12.30%

Montana

2,519

48

1.90%

52

2.10%

Nebraska

5,162

173

3.40%

685

13.30%

New
Hampshire

2,714

17

0.60%

270

9.90%

New Jersey

18,968

1,092

5.80%

168

8.90%

New York

53,613

23

0.00%

4,198

7.80%

North Carolina

37,695

85

0.20%

3,052

8.10%

North Dakota

1,632

23

1.40%

63

3.90%

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 4 5

APPENDIX TABLE A2: Percentage of Custodial Population (Both Sexes) in Administrative
Segregation (Ad Seg) and Restrictive Housing (Fall 2014) (n = 34) (continued)
Total

Ad Seg

% of Total

All Restrictive
Housing

% of
Total

Ohio

50,554

1,553

3.10%

2,064

4.10%

Oklahoma

27,488

1,183

4.30%

1,317

4.80%

Oregon

14,591

239

1.60%

1,025

7.00%

Pennsylvania

49,051

1,060

2.20%

2,339

4.80%

South Carolina

21,575

483

2.20%

1,735

8.00%

South Dakota

3,627

105

2.90%

221

6.10%

Tennessee

21,030

445

2.10%

2,626

12.50%

Texas

150,569

6,301

4.20%

6,301

4.20%

Utah

6,995

95

1.40%

832

11.90%

Washington

16,554

296

1.80%

806

4.90%

Wisconsin

21,996

96

0.40%

1,363

6.20%

Wyoming

2,074

50

2.40%

110

5.30%

Total

1,049,984

25,363

2.57%

64,976

6.91%

Source: Liman Program and Association of State Correctional Administrators (2015, Table 1).

4 6 • National Institute of Justice | NIJ.gov

APPENDIX TABLE A3: Goals and Intended Impacts Associated With Supermax Prisons
Increase prison safety:
• Fewer murders of staff and inmates.
• Fewer assaults on staff and inmates.
• Fewer riots.
• Less concern and fear among inmates and staff about threats to personal safety.
Increase systemwide prison order and control of prisoners:
• Greater compliance with rules by inmates.
• Greater and more consistent fulfillment of daily routines and obligations by inmates.
• Fewer disruptions and outbursts.
• Fewer lockdowns in general-population prisons.
• Fewer use-of-force incidents by staff.
• Fewer warning shots fired by staff.
Improve supermax prisoners’ behavior:
• More successful reintegration of supermax inmates into other prisons and society.
• Greater rule compliance following release from supermax prison.
• Less violence following release from supermax prison.
• Fewer returns to supermax prison.
Reduce the influence of gangs:
• Less gang involvement.
• Less intimidation by gang members of fellow inmates.
• Less drug trafficking.
Punish violent and disruptive prisoners:
• Increase level of punishment for violent and disruptive inmates.
• Increase perceived level of punishment among violent and disruptive inmates.
Increase public safety:
• Fewer escape attempts.
• Fewer successful escapes.
• Lower recidivism rates among supermax and general-population inmates.
• Less crime.
• Less fear of crime among residents.
Improve operational efficiencies:
• Reduce delays for inmates awaiting placement into some type of segregation.
• Reduce costs by operating fewer segregation cells and blocks in different facilities.
• Reduce staff time devoted to transporting inmates between facilities.
Source: Table reproduced from Mears and Watson (2006, p. 242).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 47

CHAPTER 2

The Use of Administrative
Segregation and Its Function
in the Institutional Setting
Ryan M. Labrecque
Portland State University

Introduction

A

dministrative segregation — often referred to as solitary confinement —
involves housing an inmate in conditions characterized by substantial
isolation from other inmates (American Bar Association, 2011).
Administrative segregation is the prison system’s answer for dealing with serious
inmate misbehavior within the institution, in the same way that incarceration
is society’s solution for dealing with dangerous criminals in the community.
From this perspective, administrative segregation represents a form of detention
within the institution — a prison within the prison (Browne, Cambier, & Agha,
2011). The historical origins of this practice trace back to the silent penitentiaries
in the early 19th century, where inmates were subjected to extreme forms of
social isolation and sensory deprivation (Rothman, 1998a). The function and
extent of the use of administrative segregation have since undergone several
changes in the United States, from serving as the main reformation strategy for
entire prison populations to being used as a risk-management tool aimed at
removing select inmates from the general prisoner population (Shalev, 2009).
There is a widely held belief among policymakers and corrections officials that

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 49

using administrative segregation makes prisons and communities safer (see
Mears, 2013). However, those critical of the practice contend that administrative
segregation is an overused correctional policy, which has many damaging effects
on inmates (Cloud, Drucker, Browne, & Parsons, 2015; Haney, 2012a) and staff
(Ferdik, 2015; Haney, 2008).
Beginning in the mid-1980s, there has been an increased amount of scholarly
attention given toward the study of administrative segregation (O’Keefe, 2008).
The majority of the available research is qualitative in nature and includes
interviews with inmates and mental health professionals in administrative
segregation settings (Labrecque & Smith, 2013). As a group, these anecdotal
reports tend to use powerful excerpts from these interviews to suggest that
administrative segregation violates prisoners’ constitutional rights, contributes
to psychological problems, increases criminogenic risk, and is used excessively
in the United States (e.g., Fellner & Mariner, 1997; Haney, 2009; Kupers, 2008;
Lovell, 2008; Toch, 2003). Subsequently, there is a strong consensus in the
literature, as well as a growing public sentiment, that administrative segregation
is responsible for producing these devastating effects (e.g., Bauer, 2013; Gawande,
2009; Goode, 2012; Guenther, 2012; Keim, 2013). These perceived negative
effects have helped make this practice an issue of national prominence for
correctional administrators.
Although administrative segregation is widely used in many jail and prison
systems throughout the United States, it also remains an elusive subject of
scholarly research (Smith, Gendreau, & Labrecque, 2015). From an empirical
standpoint, very little is known about the extent of the use of this policy or its
effects on inmates and staff in the correctional environment. Likewise, there is
a need to better understand if administrative segregation is an effective strategy
for reducing crime and promoting justice. This white paper aids in this endeavor
by examining the use and function of administrative segregation in institutional
settings in the United States. More specifically, this paper synthesizes the
literature on how administrative segregation is used to manage and control
inmates in correctional facilities, discusses the limitations of the current
empirical research, and makes recommendations for future research. The process
for locating relevant studies for this paper included searching several databases
(e.g., Criminal Justice Abstracts, Google Scholar, National Criminal Justice
Reference Service) followed by an ancestry approach, where the reference lists of
each identified study were used to locate additional studies.

Definitional Challenges and the Importance of Terminology
Policy evaluations must begin by defining which specific strategy or intervention
is being tested (Rossi, Lipsey, & Freeman, 2004). However, there is no
universally agreed-upon definition of what constitutes segregated confinement

50 • National Institute of Justice | NIJ.gov

(Butler, Griffin, & Johnson, 2013). A review of state and federal administrative
segregation policies reveals that, in practice, segregation settings are referred to
by a variety of names, including “Security Housing Units,” “Restricted Housing
Units,” and “Intensive Management Units.”1 It is likely that correctional agencies
consciously choose the names of their segregation units to reflect the underlying
notion of what purpose they believe the practice should serve. This idea
about terminology and function has been raised elsewhere — under different
circumstances — for example, when Hans Toch (1978) questioned whether
or not a “correctional officer” by any other name was still a “screw.” It remains
unknown, however, whether changing the name of a segregation setting to
include the terms “management,” “behavioral,” or “modification” — rather than
the terms “security,” “housing,” or “restricted” — represents a fundamental shift
in the underlying ideology of the practice and a move toward redefining how
these units should be used. Do these new labels reflect notions of rehabilitation
or are they mere semantics? Do these settings still represent more of the same old
style of segregation under a new name?

Types of Segregation
Segregation is used in many jails and prisons throughout the United States,
ranging from minimum- to supermaximum-security facilities (Browne et al.,
2011). Correctional institutions use segregation for four distinct purposes:
responding to serious misconduct (disciplinary segregation), ensuring the well­
being and order of the facility (administrative segregation), protecting the inmate
from harm (protective segregation), and meeting other institutional needs
(temporary segregation).

Disciplinary Segregation
Disciplinary segregation — also referred to as punitive segregation — is a form
of punishment for inmates who violate the institution’s rules (Harrington, 2015).
Whenever an institutional violation occurs, a staff member may write up the
perpetrator for the misconduct and a hearing before the rule infraction board
will determine the facts in the case. At the hearing, evidence is presented against
the accused and he or she can either accept blame (i.e., plead guilty) or defend

1

In California, segregation units are called “Security Housing Units” or “SHUs,” and in New York, the same acronym
stands for “Special Housing Units.” Texas segregation units are called “High Security Units”; in Rhode Island, they
are “High Security Centers”; in Louisiana, they are “Closed Cell Restricted” or “CCRs”; and in Pennsylvania they
are called “Restricted Housing Units” or “RHUs.” In the federal prison system, one type of extreme segregation is
the “Communication Management Unit” or “CMU.” In Washington State, the term “Intensive Management Unit” or
“IMU” is used, and in Maine these units are called “Special Management Units” or “SMUs.” This is by no means
a comprehensive list, but even this short review highlights that segregation settings are referred to in many
different ways within the United States.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 51

himself or herself against the charges (e.g., call witnesses). If the inmate is found
guilty, a range of sanctions may be imposed. These punishments can include
the removal of specific privileges, loss of good time, or a sentence for a specific
length of time in disciplinary segregation. The type and severity of the specific
sanction for any one case depend largely on the nature of the misconduct and the
perpetrator’s prior behavioral history in the facility. Departmental regulations
often place limits on the amount of time an inmate may be housed in disciplinary
segregation (e.g., 30 days). However, if the offender is charged with multiple
violations, or if he or she accrues new violations while in segregation, the length
of stay can be extended (Metcalf et al., 2013).

Administrative Segregation
Administrative segregation is used for managerial purposes, including as a
response to an inmate who demonstrates a chronic inability to adjust to the
general population, or when authorities believe an inmate’s presence in the
general population may cause a serious disruption to the orderly operation
of the institution (Shalev, 2008). Administrative segregation is often enforced
for indeterminate periods of time. In some systems, inmates are not told the
reason for their transfer to the administrative segregation unit, and options for
reevaluation or release back to the general prison population are few (Fellner,
2000). For those inmates considered to be a continued threat to the safety and
security of the facility, segregation can be imposed for very long periods (Mears
& Bales, 2010). In more rare cases, some inmates are held in segregation until
they are discharged to the community at the expiration of their sentence (Lovell,
Johnson, & Cain, 2007).

Protective Segregation
Protective segregation — also referred to as protective custody — is used to
separate vulnerable inmates from the general inmate population due to personal
physical safety concerns. These inmates often include sex offenders, police and
prison informants, former police and correctional officers, and those at risk for
self-harm (Wormith, Tellier, & Gendreau, 1988). Although inmates in protective
custody are segregated for their own protection, restrictions on their contact with
others and the programming they receive are often similar to those inmates held
in segregation for disciplinary or administrative purposes (Browne et al., 2011).

Temporary Segregation
Temporary segregation is the placement of an inmate in restrictive housing that
can occur for a wide range of institutional needs. For example, it may be used
as an interim status for inmates pending their transfer to another institution
or awaiting a judicial proceeding, to facilitate a criminal investigation, or when
52 • National Institute of Justice | NIJ.gov

limited bed space in an institution necessitates the use of an otherwise empty
segregation cell (Labrecque, 2015a). Due to its nature, temporary segregation is
usually short in duration, but it can often precede disciplinary or administrative
segregation placements (Harrington, 2015).
Although correctional institutions segregate inmates for many reasons,
the differences in the living arrangements and privileges granted to those
residing in these settings are described as minimal (see Kurki & Morris, 2001).
That is, within a particular segregation unit, inmates held for disciplinary,
administrative, protective, or other purposes are generally exposed to the
same restrictive conditions and treatment by staff. To an outside observer,
the type of segregation being imposed on any particular offender may not be
apparent when walking through the unit. Therefore, the term “segregation”
is used in this paper to refer to the general practice of isolation in restricted
environmental settings. However, where appropriate, the specific type of
segregation being discussed will be acknowledged.

Conditions in Segregation
The conditions in segregation are often characterized by intense isolation and
absolute control (see Shalev, 2008). To assess these conditions, the segregation
policies from state and federal departments of corrections were collected and
reviewed.2 For agencies that did not have their policies available online, a written
copy was requested by email or phone. These policies led to several insights
regarding the living conditions in segregation. First, prisoners are typically
confined to a single cell for 22-23 hours a day and are subjected to increased
cell restrictions and heightened security procedures. Prisoners in segregation
are granted limited access to education, vocation, visitation, recreation, and
other services that are available to the general prison population (see also the
review by Metcalf et al., 2013). Prisoners are often taken out of their cells for
only one to two hours per day, usually for a shower or exercise. Recreation in
many segregation units takes place in a small fenced-in area that is exposed to
the weather. During extreme weather conditions, prisoners must choose between
going into these areas or remaining in their cells, thereby taking no out-of-cell
exercise for the day (Browne et al., 2011). Before leaving their cell for any reason,
inmates are handcuffed, and sometimes even shackled at the waist and placed in
leg irons.
Except when overcrowded conditions require double-bunking, virtually all
forms of social interaction with staff and other inmates are eliminated (Browne
et al., 2011). Inmates eat, sleep, and use the toilet in their cells. Food is delivered
2

Policies were collected from a total of 49 jurisdictions, which include 48 state departments of corrections and the
Federal Bureau of Prisons. A request was made to Delaware and Louisiana, but their policies were not received
at the time of this publication.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 53

through a slot in the door, meetings with counselors and mental health providers
are often conducted through the cell door, and exercise is taken alone. Visits
are often restricted and can be prohibited for a certain period of time when the
inmate first arrives in the segregation unit. When family visits are allowed, the
visitor and the inmate often sit on separate sides of a thick glass window and
must communicate via a telephone. Finally, even mental health and medical
services are extremely limited for prisoners in segregation, which further reduces
their opportunity for human contact (Butler, Johnson, & Griffin, 2014). By
comparison, inmates living in the general prison population have greater access
to various activities (i.e., programming, recreation), which affords them a degree
of meaningful social interaction (Gendreau & Labrecque, 2016).

The Function of Segregation in Institutional Settings
Since the inception of the penitentiary in the early 19th century, segregation
has remained an important component of the American penal system.
Throughout history, the use of segregation has sought to serve many purposes,
including reformation, punishment, protection, behavior modification, and
prisoner management and control (Shalev, 2009). These diverse, and at times
contradictory, objectives make this practice the center of much controversy and
debate (see Haney, 1997; and Scharff-Smith, 2006). Furthermore, each of these
goals is rooted in several different theories about human nature, crime, and
punishment, which make assessing the effectiveness of this correctional policy
difficult (Mears, 2008). To understand the function that segregation serves in
modern correctional institutions, it is helpful to recognize the pretenses under
which the practice was first developed. It is also important to understand how
various social and political events in society have led to changes in the use of
segregation and the role it is expected to serve.

The Penitentiary
Segregation as a penal strategy first emerged in the United States in the early 19th
century (Rothman, 1998a). During this time, penal reformers began to view the
rising national crime rates as evidence that many of the country’s prisons were
not effective at reducing crime (Kann, 2005). However, despite the perceived
inadequacies of these institutions, reformers did not give up on the concept
of prison (Rothman, 1980). For many, imprisonment as a societal response
to crime still represented a vast improvement over the capital and corporal
punishments that were used during the colonial era (Foucault, 1995). With no
ready alternatives to take its place, reformers turned their attention to correcting
what they perceived to be the defects of these early institutions: the destructive
nature of the prison environment (Rothman, 1971). In this social context, the

5 4 • National Institute of Justice | NIJ.gov

penitentiary emerged ready to replace those institutions that were built in the
previous generation (Ignatieff, 1983).
Largely influenced by the ideology of evangelical Quakers, penitentiaries were
built by intent and design to separate inmates from all contact with corruption,
including staff and other inmates (Cullen & Gilbert, 1982). Initially, two competing
organizational schemes emerged: the Pennsylvania “solitary” system and the
Auburn “silent” system. The Pennsylvania model compelled inmates to work alone
in their cells and demanded absolute isolation (Franke, 1992). By contrast, the
Auburn model allowed inmates to congregate during the day for meals, work, and
prayer, but otherwise forced them to remain in their cells alone (McGowen, 1998).
Whereas the Pennsylvania system used physical barriers to separate inmates from
interacting with one another, the Auburn system relied on corporal punishments
(e.g., whipping) to enforce the rule of silence (Rothman, 1980).
Despite this fundamental difference, both systems emphasized the use of
isolation, obedience, and a steady routine of labor as an integral part of the
plan for reformation (Rothman, 1998b). The underlying philosophy of both
models was that isolation would afford prisoners the ability to repent and reform
(Rogers, 1993). Correctional administrators were confident in the power of faith
to reform prisoners and were distinguished in their belief that rehabilitation
was the only real task of the institution (McGowen, 1998). By removing the
person from all temptations and substituting a steady and regular regimen, the
segregation setting would ultimately reform the individual (Lieber, 1838/2010).
Initially, many state facilities followed the Pennsylvania model of total isolation,
but it was the Auburn model that became the blueprint for nearly every prison
built during the mid-19th century (Blomberg & Lucken, 2010). This was,
perhaps, because congregate living was less expensive than unbroken solitary
living; and that the Auburn model promised to hold more inmates, and thus
could bring in more money through convict labor (Rothman, 1998a). The
congregate model also had to rely on corporal punishments to ensure compliance
on the rules of silence. As soon as these institutions became crowded and
corruption became rampant, ensuring silence and isolation simply became
impossible (Rotman, 1990). In addition, there were growing concerns that
isolated confinement caused psychological damage and that despite its hype and
promises, the penitentiary did not eradicate crime (Kann, 2005). Prison officials
were thus forced to rethink how prisons should operate.
Throughout the early decades of the 20th century, there was broad optimism that
prisons could rehabilitate criminal offenders (Allen, 1964). Although much is
credited to the advancements during this time period (e.g., expansion of parole,
probation, juvenile court), a number of historians argue that the actual practices
associated with prisons, despite the rehabilitative rhetoric to the contrary, were
still largely characterized by punishment and control (e.g., Pisciotta, 1994;
Rothman, 1980; Rotman, 1998). During this time, prisoners were not universally

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 55

being placed in solitary cells; however, segregation was still used for those
inmates for whom other methods of discipline (e.g., corporal punishments)
proved ineffective (Miller, 1980).

The Supermax
Maintaining prison order and safety has long been the primary task for
correctional administrators (Reisig, 1998; Useem & Reisig, 1999). There are two
conflicting management strategies for dealing with difficult inmates in the prison
system: the concentration model and the dispersal model. The concentration
model seeks to consolidate the most violent and dangerous inmates from the
entire prison system into one tightly controlled prison (Shalev, 2009). It is a
selective incapacitation strategy that supports the segregation of subgroups of
offenders as a means to achieve safety and order throughout the prison system
(Ward & Werlich, 2003). An example of this approach is the Alcatraz Federal
Penitentiary. When it opened in San Francisco in 1933, officials boasted that it
housed the country’s most notorious criminals (Ward, 2009).
In contrast, the dispersal model argues that the best way to manage difficult
prisoners is to spread them throughout different institutions to dilute their
negative influence in populations of generally conforming inmates (Pizarro &
Stenius, 2004). When Alcatraz was closed in 1963, many of its prisoners were
initially transferred throughout the federal prison system. Shortly thereafter,
another federal prison — United States Penitentiary (USP) Marion (Illinois) —
became known for housing the most violent and disruptive federal and state
prisoners (Richards, 2008). Initially, prisoners in USP Marion were able to
congregate for certain activities (e.g., meals, recreation). However, in 1983,
after several inmates and two officers were killed, the prison declared a state of
emergency and “locked down,” thus becoming the country’s first supermaximum
(or supermax) prison (Ward & Werlich, 2003).
Supermax facilities represent a management style in corrections that focuses
on providing increased control over inmates who are known (or thought
to be) violent, assaultive, major escape risks, or otherwise disruptive in the
general population (Riveland, 1999). Supermax settings seek to hold the most
serious and chronic troublemakers from the general prison population — the
so-called “worst of the worst” (Henningsen, Johnson, & Wells, 1999; Shepperd,
Geiger, & Welborn, 1996). This concentrated approach to managing offenders
represents a return to the tenants of the strict control practices found in the early
penitentiaries (Haney, 1993; Toch, 2001).
In the 1980s, several “get tough” penal policies were enacted in the United
States that helped contribute to an increase in the number of incarcerated
offenders (e.g., mandatory minimum sentencing laws, three-strikes laws, truth
in sentencing laws) (Austin & Irwin, 2012). The coupling of overcrowded living

56 • National Institute of Justice | NIJ.gov

conditions and increased institutional violence led to growing concerns for staff
and inmate safety throughout the country (Steiner & Wooldredge, 2008; Useem
& Kimball, 1991). In response, many states constructed their own supermax
prisons and increasingly relied on segregation to reduce violence throughout the
prison system (Sundt et al., 2008).

Contemporary Use of Segregation
Policymakers often justify the use of segregation — at least in part — on the
premise that the public demands its use (Pizarro et al., 2006). However, there is
little research that has assessed public opinion in this area. One notable exception
is a recent study by Mears and his colleagues (Mears, Mancini, Beaver, & Gertz,
2013). Their 2006 survey of 1,308 Florida residents found that public support
for supermax prisons is strong when residents anticipate a safety benefit (82
percent). Mears et al. (2013) also report that such support diminishes by 21
percent when no such benefit is expected. These results come with the caveat
that they are from only one state, and it is unknown if they will generalize to
other jurisdictions. However, this research is important because it suggests
that the public prefers correctional practices that reduce recidivism rather than
those that do not.
The use of segregation implicitly expresses sentiments of punishment and
retribution; however, its goals often include incapacitation, deterrence, and
rehabilitation (see Mears & Watson, 2006). Two recent studies independently
reviewed the available official segregation policies of state and federal
jurisdictions in the United States. First, Butler et al. (2013) examined the
supermax admission criteria in 42 state jurisdictions. They found that 98 percent
of the state systems they examined had official policies in 2010 that allowed
officials to place inmates in supermax custody because they were believed to
pose a threat to institutional safety. Other reasons found for sending inmates to
supermax include repeat violent behavior (74 percent), escape risk (67 percent),
riotous behavior (45 percent), and belonging to a group that is deemed a security
threat (36 percent). In the second study, Metcalf et al. (2013) reviewed 46 state
and federal segregation policies. They reached a similar conclusion, reporting
that in 2012, many jurisdictions placed inmates in segregation because officials
believed they posed a threat to “the life, property, security, or orderly operation
of the institution” (p. 5).
These findings, coupled with the current review of segregation policies, indicate
that the function of segregation in the modern era is to remove inmates who
pose a threat to the order of the institution from the general prison population,
which can occur for disciplinary, protective, or administrative reasons (Butler
et al., 2013; Browne et al., 2011; Metcalf et al., 2013; Shalev, 2009). Furthermore,
policymakers and corrections officials often justify the use of segregation
because they believe it is an effective strategy for increasing safety and promoting

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 57

order throughout the prison system (Mears, 2013). However, among the many
controversial issues that the practice raises is the contention that segregation
increases (rather than decreases) the likelihood of subsequent criminal behavior
and thus makes prisons and communities less safe over time (see Pizarro et
al., 2006). It has further been widely speculated that long-term durations in
segregation exacerbate the detrimental effects of the setting on inmate outcomes
(i.e., leads to even more criminal behavior; Pizarro, Zgoba, & Haugebrook, 2014).
Before examining the empirical literature on the effects of segregation, this paper
first examines the research on the use of this practice in the United States.

The Use of Segregation
Prevalence of Segregation
In 1997, the National Institute of Corrections conducted a national survey of
departments of corrections that focused on the use of supermax-style housing.
The results of this survey reveal tremendous variation in the supermax facilities
across the state and federal prison systems. Some supermax institutions are
stand-alone buildings, whereas others consist of units within an existing
correctional facility that have been redesigned to meet the strict control needs
of the supermax model. As of 2004, 44 states are known to operate 57 supermax
facilities, collectively housing at least 25,000 inmates nationwide (Naday, Freilich,
& Mellow, 2008).
In 1999, King supplemented the National Institute of Corrections (1997) data
with further information acquired from state and federal departments of
corrections. King (1999) estimates that less than 2 percent of all state and federal
inmates serving one year or more in prison were held in a supermax setting.
His assessment also reveals that the extent of the use of supermax varies widely
among states. For example, some organizations (e.g., Pennsylvania) report
incarcerating less than 1 percent of inmates in supermax settings, while others
(e.g., Mississippi) report incarcerating up to 12 percent.
These studies, while informative, focus specifically on supermax confinement
and ignore the many other forms of segregation, or segregation-like, settings to
which inmates may be exposed during their incarceration (Zinger, 2013). More
broadly, estimates on the prevalence of segregation vary from 25,000 to 100,000
(Metcalf et al., 2013). Some of the best estimates come from the Federal Bureau
of Prisons (BOP), where a recent U.S. Government Accountability Office report
reveals that approximately 6 percent of all federal prisoners are held in some
form of segregated confinement (McGinnis et al., 2014). This report also shows
that this percentage has decreased since at least 2011, when the BOP housed
approximately 7 percent of inmates in segregated confinement.

58 • National Institute of Justice | NIJ.gov

More recently, the Liman Program — in conjunction with the Association of
State Correctional Administrators (ASCA) — conducted a national survey of
departments of corrections to assess how many inmates were held in segregation
during the fall of 2014 (Liman Program & ASCA, 2015). This study found that
approximately 66,000 inmates were under some form of disciplinary, protective,
and administrative segregation, which equated to an average segregated
population of approximately 7 percent. These rates also vary greatly from a low
of 2 percent in Montana to a high of 14 percent in Delaware. However, these
estimates are derived from only 33 of the state and the federal prison systems.
Estimating the use of segregation in the United States is a continued challenge,
particularly because many jails and prisons do not track this information in
a way that is easily accessed by researchers. The lack of a clear definitional
consensus on what practices constitute segregation further make estimating its
use more difficult (Frost & Monteiro, 2016), which is likely a contributing factor
to the differences found in the previous estimates. Further, some prison systems
have been accused of failing to report, or underreporting, their use of segregation
to avoid acknowledging the use of solitary confinement in their department (see
Naday et al., 2008). If the accusations prove to be true, the current estimates may
be low.
Another challenge for determining the extent of the use of segregation is that
inmates are often held in such settings for varying lengths of time (see Mears &
Bales, 2010). A problem with relying on prevalence estimates, therefore, is that
there may be many inmates who occupy a specific segregation cell over a given
length of time. Likewise, the use of only snapshot assessments — at one point in
time — may produce estimates that appear much lower than those that include
the incident counts of all of the inmates held in the setting over a specific time.
There are two recent studies that address this issue by examining how many
inmates are held in segregation over an extended period (see Beck, 2015, and
Labrecque & Smith, 2015).

Incidence of Segregation
The first incidence-based estimate comes from Labrecque and Smith (2015),
who conducted a five-year evaluation of the use of segregation in the state of
Ohio. Labrecque and Smith (2015) report that 36 percent (or 42,632) of the
118,447 admitted to the Ohio Department of Rehabilitation and Correction
(ODCR) between 2007 and 2012 experienced some form of segregation within
the same time frame. It is important to note that this estimate is derived from
longitudinal information and includes all forms of segregation (i.e., disciplinary,
administrative, protective, and temporary), something that is not typical in
estimating the use of segregation. As a comparison, the Liman and ASCA (2015)
report, which uses a prevalence estimate, reports that approximately 4 percent of
the ODRC population was held in segregation during the fall of 2014. Although

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 59

these two investigations are from different years and are not directly comparable,
this illustration highlights the importance of examining both the prevalence and
incidence estimates of the use of segregation.
More recently, the Bureau of Justice Statistics (BJS) released a special report
on the use of restrictive housing that gives a better picture of how the practice
is used in jails and prisons throughout the United States (see Beck, 2015). The
BJS study is important because it includes a national representative sample
of incarcerated persons in both prisons and jails. Information from the latter
group has been much less discussed in the research literature (see Haney,
Weill, Bakhshay, & Lockett, 2015). Beck found that on any given day in 2011­
2012, approximately 4 percent of state and federal inmates and 3 percent of jail
inmates were being held in some form of restrictive housing. Further, this study
revealed that nearly 20 percent of prison inmates and 18 percent of jail inmates
spent time in segregation in the previous year. These estimates include inmates
in segregation for both disciplinary and administrative reasons, but due to the
nature of the survey questions it was not possible to disaggregate how many were
in each type separately.
Continued research is needed in this area. Such work will not only better inform
how segregation is used throughout the United States but will also be an essential
component for effectively reducing its use. That is, to develop any knowledgeable
strategy for reducing the use of segregation, prison officials must first understand
the basics: how many inmates are held in segregation on any given day and how
many inmates experience segregation during their incarceration. This will require
researchers and correctional agencies to continue to work together toward
overcoming the challenges associated with estimating such use.

Duration in Segregation
Many of the criticisms of the use of segregation focus on the perceived
psychological damage that occurs from spending prolonged durations in such
settings. However, there is no universally agreed upon length of time that is
considered an extended period. Those critical of the practice generally argue for
setting standards that would limit stays in segregation to 90 days (e.g., Jackson,
1983; Haney & Lynch, 1997). Likewise, the evaluation literature on this topic
tends to use 90 days as the cut-point for defining long-term segregation (e.g.,
Lovell et al., 2007; Mears & Bales, 2009). Others, however, argue for much
shorter time caps. More than 30 years ago, Gendreau and Bonta (1984) suggested
limiting the use of segregation to 14 days. More recently, the United Nations
(2015) used 15 days to define a prolonged duration. Given these concerns
about duration in segregation, there is a need to better understand how long
inmates spend in these settings, how many times they are placed in them, what
proportion of their total prison sentence is served in segregation, and what
length of time (if any) is optimal for inmates to spend in segregation settings.

60 • National Institute of Justice | NIJ.gov

In a study of the Florida Department of Corrections, Mears and Bales (2010)
found that approximately 2 percent of the inmates released from custody
between 1996 and 2001 experienced at least one stay in supermax confinement
for 30 days or more. Mears and Bales (2010) point out that most accounts
of supermax housing assume that it is a one-time event; however, their data
reveal that the average supermax inmate experiences four separate segregation
placements, and some have more than 10 separate stays. Mears and Bales (2010)
also show considerable range in the duration of time that inmates are held in
supermax, extending from one month to more than three years. Finally, the
authors report that approximately 14 percent of the sample spent half of their
prison term or more in supermax, 39 percent spent a quarter or more of their
prison term in supermax, and 15 percent spent less than 5 percent of their prison
term in supermax.
In their study of segregation in Ohio, Labrecque and Smith (2015) found that
more than half of the inmates who spent time in segregation served fewer than
30 days; however, 9 percent of the sample served 180 days or more. With each
successive placement in segregation, the mean duration of that segregation
increased, from an average of 17 days for the first stay to an average of 28 days
by the fifth stay. Of those experiencing segregation, 45 percent had only one
stay; however, 16 percent had five or more total placements. Finally, more
than half of the sample spent less than 5 percent of their total prison time in
segregation, but 9 percent (or 3,880 inmates) spent more than a quarter of their
total sentence in segregation.
The findings from these two studies indicate that the frequency and duration of
segregation vary widely. These findings are admittedly limited to these two states
and may not necessarily generalize to other jurisdictions; however, the results from
the Liman and ASCA (2015) national-level investigation reveal a similar pattern
in segregation use throughout the United States. Of the 24 jurisdictions reporting
systemwide data on the length of stay in this study, 11 reported that most prisoners
were held in segregation for fewer than 90 days and 18 described holding some
prisoners in segregation for more than three years (Liman & ACSCA, 2015).
This research indicates that although some inmates remain in segregation
settings for very long durations, the vast majority of inmates experience much
shorter stays. What remains a mystery, however, is what is an optimal time for
being held in segregation? The policy recommendations aimed at placing time
caps on segregation to date are based on personal opinion and clinical wisdom
rather than empirical evidence (e.g., Jackson, 1983; Gendreau & Bonta, 1984;
Haney & Lynch, 1997). An effort should be made to assess if there is a tipping
point of diminishing marginal returns for time spent in segregation. Such
research would better inform correctional officials about what limits to place on
inmate stays in segregation.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 61

Inmates in Segregation
Segregation settings are described as targeting the “worst of the worst” inmates,
which includes those who are escape risks, gang members, predators, and high
profile or notorious inmates (Shalev, 2009). Some raise concerns, however, that in
practice these settings actually consist of many “nuisance” inmates (i.e., those who
do not pose a direct threat to the safety and security of the institution), rather than
those who are truly violent or dangerous (Kurki & Morris, 2001; Shames, Wilcox,
& Subramanian, 2015). Given this discrepancy, there is a critical need to better
understand which types of offenders are being held in segregation settings.
Recently, Labrecque (2015b) took stock of the empirical literature on
the predictors of segregation to address the question of “who ends up in
segregation”? More specifically, his meta-analysis examines the differences in
key variables between inmates who are held in segregation settings and those
held in the general prison population. Labrecque (2015b) identified 16 studies
that met his inclusion criteria: (1) the study must have been conducted on
prisoners in a custodial setting (i.e., prison or jail), (2) the study must have
compared the characteristics of inmates in a segregation setting to those in the
general population, and (3) the study had to have contained sufficient data to
calculate an effect size (i.e., Pearson’s r or phi coefficient). The segregation and
non-segregation groups were then compared on a range of available inmate
characteristics, criminal history, institutional behavior, and criminogenic needs
variables.
The findings of this meta-analysis suggest that segregated inmates tend to be
younger and are more likely to be an ethnic minority, have a mental disorder,
be a gang member, and be rated as at high risk to recidivate, when compared
to the inmates from the general prison population. Segregated inmates were
also more likely to have a violent criminal history, have prior juvenile justice
involvement, and be higher risk on their initial institutional classification
rating. Finally, inmates in segregation settings were much more likely to have a
history of engaging in institutional misconduct and to have previously served
time in segregation.
This meta-analysis also examined the differences between the two groups with
respect to their levels of criminogenic need. Across every domain assessed, the
inmates in segregation possessed much greater levels of criminogenic needs
when compared to those in the general prison population. The magnitudes
of these differences also varied by type of criminogenic need. Specifically, the
areas of need with the largest magnitude of difference included motivation for
treatment, education, and antisocial attitudes. The next largest set of differences
was found in the areas of personal/emotional, substance abuse, and antisocial
associates. The areas with the smallest magnitude of differences were found in the
domains of community functioning/leisure, employment, and family/marital.

62 • National Institute of Justice | NIJ.gov

The findings of this meta-analysis are important but must be interpreted with
caution because many of the estimates were derived with small sample sizes and,
in some cases, with only a few effect sizes. The studies included in the metaanalysis were also limited to investigations from a subset of correctional systems
that were willing to share their data on this controversial and potentially litigious
practice. Therefore, these results may not necessarily generalize to all correctional
systems in the United States. Fortunately, the recent BJS report on segregation
also examines the differences between prisoners held in segregation and those
found in the general prison population (see Beck, 2015).
The findings of the BJS study support those from the Labrecque (2015b) metaanalysis. Specifically, the BJS report found that younger inmates, inmates without
a high school diploma, and lesbian, gay, and bisexual inmates were more likely to
have spent time in restrictive housing than older inmates, inmates with at least
a high school diploma, and heterosexual inmates (Beck, 2015). The report also
found that inmates sentenced for violent offenses (not including sex offenses)
and inmates with extensive arrest histories or prior incarcerations were more
likely to spend time in restrictive housing than those held for other offenses and
with no prior arrests or incarcerations. Finally, inmates who were involved in
serious institutional violence (i.e., assaulting other inmates or staff) and those
suffering from serious psychological distress were significantly more likely to
spend time in restricted housing than those who were well behaved and did not
have any mental health issues (Beck, 2015).
This review of the available evidence does not support the contention that
segregation settings are reserved only for the most highly incorrigible and
dangerous offenders. Rather, the available evidence indicates that perhaps a
better way to describe the segregated population is “difficult to manage.” This
research suggests that the segregated population tends to possess those traits
that correlate more highly with antisocial behavior. They are mostly younger,
have more extensive criminal histories, worse institutional behavior, and more
criminogenic needs. On a positive note, this does indicate that prisons are
effectively targeting the inmates who are most at risk for engaging in criminal
behavior for placement in segregation. However, segregation is used under the
assumption that the setting will improve safety and security within the prison
system and beyond (Mears, 2013), but there are many theoretical reasons
that suggest this practice may not be the best strategy for effectively achieving
these goals (see Gendreau & Labrecque, 2016). Next, the empirical literature is
reviewed to assess what effect segregation has on subsequent inmate outcomes.

The Effects of Segregation
There is a critical need to determine if segregation is an effective strategy for
making prisons safer and more humane settings (Labrecque, 2015a). This review

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 6 3

of the evaluation literature begins by first examining its effect on psychological
outcomes. Because improving mental health function is not a goal of segregation,
this literature is only briefly discussed. It is left to other white papers to more
comprehensively examine the psychological impact of segregated confinement
(e.g., Frost & Monteiro, 2016). Instead, the current discussion focuses more
extensively on the effect that segregation has on a variety of criminal behavior
outcomes (e.g., institutional levels of violence, post-release recidivism,
institutional misconduct), because one of the often-cited rationales for the
practice is that it is an effective deterrent of such behaviors (Angelone, 1999;
Gavora, 1996).

Psychological Outcomes
Without a doubt, the most contentious debates in this area involve the
psychological effects of segregation. The belief that segregated confinement
causes psychological damage is not new. After visiting some of the early
United States penitentiaries in the 19th century, several notable European
contemporaries condemned the practice, suggesting it causes inhabitants undue
psychological distress (e.g., de Beaumont & de Tocqueville, 1833; Dickens,
1842/1985). The majority of the research conducted to date, however, has largely
been limited to qualitative investigations (e.g., Grassian, 1983; Haney, 2003;
Kupers, 2008; Lovell, 2008). Some scholars insist these anecdotal reports are
the unequivocal evidence that segregation causes serious mental health issues
amongst its inhabitants (Haney, 2012b).
More recently, however, other scholars point out the methodological shortcomings
in much of the literature that contributes to this conclusion (e.g., selection bias,
response bias, inadequate or no control groups), which in their estimation limits
the credibility of their results (Gendreau & Labrecque, 2016; Suedfeld, Ramirez,
Deaton, & Baker-Brown, 1982; Zinger, Wichmann, & Andrews, 2001). Two recent
independent meta-analytic reviews on this topic (Labrecque, Smith, & Gendreau,
2013; Morgan et al., 2014) conclude that not only has segregation been an elusive
subject of empirical research, but also the effects on inmate physical and mental
health functioning found from the available studies tend to be in the “small” to
“moderate” range, rather than “large” as has been predicted by those critical of the
practice (see also Smith, Gendreau, & Labrecque, 2015). These findings suggest
that segregation may not produce any more of an iatrogenic effect than do other
housing options in prison (i.e., general population).
Although more research is clearly needed in this area before any definitive
conclusions should be drawn, these findings serve as a caution to reviewers about
making judgments regarding the effects of segregation too hastily, especially
when they are based on qualitative rather than quantitative evidence. More
empirical research is needed on the psychological effects of segregation. It
should address the number of research design issues that have been identified

6 4 • National Institute of Justice | NIJ.gov

in this literature base. It should also assess the moderating effects of the quality
of staff-inmate interactions, conditions of confinement, increased length of time
in segregation, and other offender-level characteristics (e.g., age, gender, mental
health status, risk level; see Gendreau & Labrecque, 2016). This type of research
is important to corrections administrators because it can help them identify
which inmates to exclude from placement in segregation. It could also serve as
a guide for improving the conditions in such settings in order to achieve more
desirable outcomes (e.g., less psychological deterioration, improved behavior).

Behavioral Outcomes
Despite the many calls for more empirical evaluations of the effects of segregation
(e.g., Briggs, Sundt, & Castellano, 2003, p. 1342; Kurki & Morris, 2001, p. 393;
and Ward & Werlich, 2003, p. 54), a very limited number of behavioral outcome
studies have been conducted to date. Within this limited research base, three
types of outcomes are examined: institutional violence, post-release recidivism,
and institutional misconduct. One of the key reasons that the field lacks sound
empirical knowledge on the effects of segregation is due, in part, to the type of
research methodology employed in these investigations. Therefore, the current
literature base is reviewed and is further separated by the type of research
methodology used.

Institutional violence
One of the often-stated purposes of the use of segregation includes improving
the systemwide order in prison systems (see Mears & Watson, 2006). Very little
empirical research has assessed whether segregation is effective in reducing
aggregate levels of violence. The limited research in this area typically examines
trends in measures of institutional violence across correctional systems over time. It
looks specifically for differences before and after changes in the use of segregation
(e.g., construction of supermax facility) and has produced mixed findings.
In their discussion on the effects of policy changes in the Texas prison system
in the late 1980s, Marquart and colleagues attribute the decline in institutional
violence and inmate murders to the massive lockdown of the state’s gang
members into segregation settings (Crouch & Marquart, 1989, 1990; Ralph
& Marquart, 1991). Ralph and Marquart acknowledge that this policy change
drastically increased the number of inmates held in administrative segregation
but remain convinced that the concentration strategy is effective in reducing
levels of violence throughout the prison system. More recently, Useem and
Piehl (2006) used a similar analytical approach with national-level prison data
over a longer period. However, they concluded that the decrease in the number
of prison riots, disturbances, arsons, escapes, assaults, and murders between
the 1980s and early 2000s did not correlate with the changes in segregation
practices, which they point out actually declined between 1982 and 2001. The

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 65

use of segregation, therefore, was not directly responsible for the improvements
in prison order. It is worth pointing out that this type of research design is
speculative because it fails to consider the other historical threats to validity
(i.e., other changes within the system that occurred during the same time).
More recently, researchers have employed more advanced statistical techniques
to assess the influence of segregation on institutional violence, which has
also produced mixed findings. For example, using a multilevel model design
with a nationally representative sample of 4,168 male inmates from 185 state
correctional facilities, Huebner (2003) assesses the effect of different types of
administrative control on inmate assaults. She found segregation use — defined
as “the percent of the total inmate population that received solitary confinement
as a disciplinary response to the most recent rule infraction” (p. 110) — was
ultimately unrelated to levels of inmate-on-inmate or inmate-on-staff assaults.
Another study conducted by Briggs et al. (2003) uses a multiple interrupted
time series design to examine whether the emergence of supermax housing
in three states (Arizona, Illinois, and Minnesota) produced a reduction in
systemwide levels of violence, when compared to a comparison state that did not
construct a supermax prison in the same period (Utah). They found supermax
prisons did not reduce levels of inmate-on-inmate violence but did find mixed
support for their ability to increase staff safety. Specifically, the implementation
of a supermax prison had no effect on levels of inmate-on-staff violence in
Minnesota, temporarily increased staff injuries in Arizona, and reduced assaults
against staff in Illinois. It is worth noting that only four of the 24 states sampled
provided the researchers with sufficient data to conduct the time series analysis
(i.e., monthly violence estimates over the five years before the construction of the
supermax facility in their state), which “raises concern about the generalizability
of the sample” (p. 1352).
The evidence does not support the contention that supermax prisons are
responsible for reducing systemwide levels of violence. This finding calls into
question the justification of the practice on the basis that it improves safety and
order throughout the prison system. However, more research is needed in this
area, particularly for investigations that can overcome some of the shortcomings
found in the prior research.

Recidivism
Approximately half of the respondents in a national survey of prison wardens
identify rehabilitation as a goal of segregation (see Mears & Castro, 2006).
Likewise, several empirical investigations assess the effect of segregation on
measures of post-release recidivism. These studies, however, vary widely in their
methodological quality and results.

6 6 • National Institute of Justice | NIJ.gov

In a study from the federal prison system, Ward and Werlich (2003) use a
nonequivalent comparison group design to examine the differences in the returnto-prison rates between a group of inmates released from Alcatraz (i.e., segregation
group; n = 1,550) and a random subsample of inmates released from Leavenworth
(i.e., non-segregation comparison group; n = 257) between 1934 and 1963 (see
also Ward, 2009). Ward and Werlich (2003) found that inmates released from
Alcatraz were more likely to be returned to federal custody during follow-up
(50 percent) compared to those released from Leavenworth (37 percent). However,
the offenders sent to Alcatraz had more extensive and serious criminal histories
(Ward, 2009). Likewise, it is reasonable to suspect that differences between these
two groups may have also had an influence on the results.
In 2011, Seale and colleagues also conducted a nonequivalent comparison
group recidivism study in the California Department of Corrections and
Rehabilitation. They found that, of the inmates who were released from custody
during fiscal year 2006-2007, those who had served time in the Security
Housing Unit (SHU) had a higher return-to-prison rate during a three-year
follow-up (70 percent) compared to those who did not spend any time in SHU
during their commitment (65 percent). However, prior group differences may
have affected the results. For example, prior research suggests that inmates in
segregation are more likely than inmates in the general population to possess
many characteristics (e.g., younger age, greater criminal histories, higher risk
for recidivism, gang affiliation) that also place them at a greater likelihood for
recidivating (Lovell, Cloyes, Allen, & Rhodes, 2000). The results from these
studies should be interpreted cautiously.
In Washington state, Lovell and colleagues (2007) employed another type
of research methodology to assess the effects of segregation on recidivism:
the matched comparison group design. They identified 200 inmates who
were released from the Washington Department of Corrections in 1997 and
1998 and who also served at least 12 weeks in a supermax setting during
their commitment. These supermax inmates were matched one-to-one to a
comparison group of non-supermax inmates from the larger pool of 6,453
offenders released during the same time. These two groups were matched based
on nine demographic and criminal history variables. Lovell and colleagues
(2007) found that the inmates who experienced supermax were more likely to be
found guilty of a new felony within three years of release (53 percent) compared
to the inmates who did not experience supermax (46 percent); however, this
difference was not statistically significant. They also reported that inmates who
were released into the community directly from the supermax setting were more
likely to be found guilty of a new felony (69 percent), compared to those who
experienced segregation, but spent three or more months in general population
before being released into the community (46 percent).
It is worth noting that offenders from the former group were also younger
and had more extensive criminal histories, compared to those from the latter

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 67

group. When the direct-release and later-release inmates were matched for age
and criminal history — two well-known predictors of criminal behavior — the
significance of the group difference disappeared. Lovell and colleagues (2007)
argue that this finding may be an artifact of the small sample size in their
study. However, it is also possible that segregation may not have any effect on
recidivism once other relevant factors are considered.
More recently, researchers have started employing more analytically advanced
matching techniques — most notably propensity score matching — in an attempt
to reduce the selection group biases in segregation research. In a study of the
Florida Department of Corrections, Mears and Bales (2009) examined three-year
recidivism outcomes between a group of supermax inmates who spent more than
90 days in a supermax setting (n = 1,267) with a comparison group of inmates
who were propensity score matched from the larger pool of inmates in the Florida
system during the sampling time frame (n = 58,752). Although the differences
found between the two groups for any recidivism was not significant (59 percent
for supermax compared to 58 percent for non-supermax), inmates in the supermax
group were significantly more likely to violently recidivate during follow-up than
those in the control group (24 percent for supermax compared to 21 percent for
non-supermax). Further, Mears and Bales (2009) found no evidence that the
duration spent in segregation, or the timing of release from segregation (direct or
later release), had any significant effect on the outcomes examined.
Another outcome evaluation using propensity score analysis matched 57 inmates
who had served time in the Ohio supermax prison during select periods in
2003 and 2005 to a control group of inmates from the general prison population
who did not serve any time in the supermax setting (n = 1,512) (Butler, Steiner,
Makarios, and Travis, 2013). Inmates were matched on the characteristics of
age, race, risk level, sentence type and severity, gang member status, sex offender
status, education, and time served. The results showed no statistically significant
difference between the two groups in terms of rearrest or return to prison during
a seven-year follow-up period.
It is important to note that the recidivism studies with the weaker methodological
designs produce much larger effect sizes than those with more scientific integrity.
The findings from the more methodologically rigorous studies reveal a null effect
of segregation on recidivism. These findings challenge the use of segregation to
reduce post-release recidivism. However, much more work is needed in this area.
In particular, studies are needed that can overcome the challenges of identifying
appropriate comparison group cases.

Institutional misconduct
Increasing prison safety is an often-cited goal of segregation (see Mears & Castro,
2006). Likewise, researchers have begun to assess what effect segregation has on
individual levels of institutional misconduct. Recently, Morris (2016) evaluated

68 • National Institute of Justice | NIJ.gov

the effect of short-term exposure to segregation on subsequent misbehavior
in a sample of male inmates who were admitted into the Texas Department
of Corrections between 2004 and 2006. He used a multilevel counterfactual
research design (i.e., propensity score matching) to assess group differences
between inmates who were sent to segregation as a punishment for an initial act
of violence (the treatment group) and those who were not sent to segregation as
a punishment for an initial act of violence (the control group) during the first
two years of their commitment. He found that the use of short-term disciplinary
segregation (i.e., 15 days or less) had no statistically significant effect on the
occurrence or timing of subsequent violent infractions.
Another recent study conducted by Labrecque (2015a) examined the impact of
segregation on subsequent misconduct among inmates who were admitted into
the Ohio Department of Rehabilitation and Correction between 2007 and 2010
and who experienced at least one stay in disciplinary segregation during their
commitment (n = 14,311). He used a pooled time series design to assess whether
the experience of segregation in a preceding time wave (and the number of
days spent in segregation) influenced the probability of engaging in misconduct
during the next time wave.3 This research design is particularly useful because it
takes advantage of the within-individual variation in the exposure to segregation
to assess whether the experience had an influence on the probability of being
found guilty of an institutional misconduct charge in the subsequent time wave.
Labrecque also found that neither the experience of disciplinary segregation,
nor the number of days spent in such settings, had any significant effect on the
prevalence or incidence of the finding of guilt for subsequent violent (e.g., assault),
nonviolent (e.g., damage to property, theft), or drug infractions (e.g., possession of
drugs and alcohol).
These two studies — although certainly not without their limitations — represent
methodologically rigorous tests using sophisticated analytical procedures to
assess the influence of segregation on subsequent measures of institutional
behavior. The findings of both studies indicate that the experience of disciplinary
segregation does not decrease, or increase, institutional misbehavior. Instead,
they support the contention that segregation has no significant effect on
criminal behavior. This research is naturally in need of replication, and further
investigations of institutional behavior should include segregation stays of a
longer duration and for different reasons (i.e., administrative).

Discussion
This white paper explored how segregation is used to manage and control
inmates within correctional institutions. It shows that definitional and reporting
3

Time waves were constructed into three-month intervals beginning with the inmate’s initial admission date.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 69

challenges make it difficult to determine how this practice is used throughout
the United States. Despite these obstacles, this investigation revealed several
important findings on the use of segregation that are important for researchers
and practitioners alike. Further, these findings help move the National Institute
of Justice forward in its attempt to advance knowledge on the use and effect
of segregation in the United States and to help translate research findings so
criminal justice professionals can make informed decisions regarding the future
use of this practice (Rodriguez, 2015).
First, the differences found in the prevalence estimates between studies are likely
due to the parameters that researchers place on the definition of segregation.
Estimates derived from only examining the number of inmates in supermax
settings are much lower than those that also include other forms of segregation
(e.g., administrative, disciplinary, protective, temporary). Second, estimates
of the incidence of segregation suggest that many more inmates experience
such settings during their commitment, when compared to the estimates of the
prevalence of segregation. This difference is due to the fact that longitudinal
examinations are able to capture the many inmates who cycle through these
units over time rather than relying on a one-time snapshot assessment of the
number of inmates held there on any given day. Third, the use of segregation
varies considerably not only across jurisdictions but also between inmates. Some
inmates experience segregation as a one-time event, while others experience it
many times during their commitment. Fourth, the length of time inmates serve
in segregation also varies considerably, from days to multiple years, with some
inmates who spend the large portion of their entire incarceration sentence in
such settings.
This white paper also examined the characteristics of the inmates who are sent
to segregation. Inmates housed in segregation differ significantly from those
in the general prison population in many easily identifiable factors that are
routinely collected by many departments of corrections (Beck, 2015; Labrecque,
2015b). Inmates in segregation settings tend to be younger and are more likely
to be an ethnic minority, have a mental disorder, be a member of a gang, have
a more extensive criminal history, and have a record of prior misbehavior in
the institution. They also are likely to be rated as at high risk to recidivate when
compared to the inmates from the general prison population. This is important
information because corrections officials could use it to proactively identify and
treat inmates with greater propensities toward being placed in segregation settings
in an effort to reduce the need for segregation. It is worth noting that Helmus
(2015) recently developed a risk assessment scale — the Risk of Administrative
Segregation Tool (RAST) — to predict inmate placements in administrative
segregation in the Canadian federal prison system (see also Helmus, Johnson, &
Harris, 2014).4
This work represents a crucial first step in assisting correctional agencies to
better identify which inmates are at high risk for being sent to segregation. Such

70 • National Institute of Justice | NIJ.gov

knowledge could certainly be beneficial to efforts aimed at diverting offenders
from such placements. A certain amount of caution, however, should be exercised
before correctional agencies adopt risk instruments like the RAST. As Labrecque
(2015b) notes, there is an inherent risk in using a segregation risk assessment that
includes only static (i.e., unchangeable) factors. This information could potentially
be used to justify isolating inmates based on their risk score. And because of the
static nature of the items in the assessment, there is nothing the offender could do
to reduce his or her risk. In short, the use of a static assessment has the potential of
increasing the need for segregation rather than for reducing it.
The Labrecque (2015b) meta-analysis reveals that inmates in segregation differ
from those in the general prison population not only on demographic and criminal
history variables but also in their criminogenic needs. This finding has important
implications for treatment because correctional administrators could use this
information to help identify which criminogenic needs to target with intervention.
The use of only a static risk assessment would provide no such information.
Therefore, future efforts should be made to develop segregation assessment tools
that include both risk and need items.5 The development of such a tool has the
potential of helping prison officials to improve institutional safety and reduce
their need for segregation in both the short and long term.
Finally, the primary function in the contemporary use of segregation is to
increase systemwide order, safety, and control (see Mears & Castro, 2006).
However, upon review of the limited outcome evaluation research, it is
questionable that these settings are capable of effectively achieving these goals.
The empirical research on the effects of segregation on systemwide levels of order
reveals mixed findings. Some of the early, largely speculative, research suggests
that the increased use of administrative segregation in Texas was responsible for
reducing systemwide levels of institutional violence (Crouch & Marquart, 1989,
1990; Ralph & Marquart, 1991). However, the studies conducted in this area that
employ more advanced research designs tend to find much less support for this
contention (Briggs et al., 2003; Huebner, 2003). Most notably, Briggs et al. (2003)
found evidence that the emergence of supermax prisons did not reduce levels
of inmate-on-inmate violence but appeared to reduce inmate-on-staff violence
in one state (Illinois). In sum, the research does not support the contention that
segregation is an effective strategy for reducing systemwide levels of disorder.
Another way researchers have attempted to evaluate the effectiveness of
segregation is to use subsequent individual-level behavioral outcomes. Although
there is some evidence that the experience of segregation may increase post­
4

The RAST includes six static items (age, prior convictions, prior segregation placements, sentence length, criminal
versatility, and prior violence) and has been found to have a high predictive validity.

5

It should be noted that Helmus (2015) attempted to include criminogenic needs in her risk assessment. However,
she was unable to improve the predictive validity of the tool beyond using the six static items alone, so she chose to
eliminate the dynamic needs factors from her final RAST model.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 71

release recidivism, this finding is generally limited to those research studies
using the weakest type of research methodology: the nonequivalent comparison
group design. Therefore, these findings should be interpreted cautiously. The
results from the more methodologically rigorous studies (e.g., propensity score
matching) reveal no statistically significant differences in recidivism outcomes
between the inmates who were housed in segregation and a matched sample
of those who were housed in the general prison population. Further, the
institutional misconduct literature — while much less extensive — similarly
suggests that disciplinary segregation has a null effect on subsequent behavior.
However, before any definitive conclusions are made about the effects of
segregation on behavioral outcomes, it must be acknowledged that the
evaluations in this area are few and have limitations that must be addressed by
future research.

Recommendations for Research and Practice
This white paper provided a summary of the existing literature on the use and
function of segregation, but more importantly, it seeks to serve as a springboard
for future research. This paper also intends to help inform practitioners who
work in segregation environments on the current state of research about the
practice. The following six recommendations are made to help improve the
state of the segregation research and to assist correctional authorities in making
informed decisions regarding the use of segregation.

Obtain better estimates on the use of segregation
With increasing pressure to reduce the use of segregation throughout the United
States, it is important to have a solid understanding of how this practice is used
across the country. Prior estimates vary widely. Moving forward, researchers and
correctional agencies should work together to obtain more reliable estimates of
segregation use. Future research should also focus on estimating the prevalence
and incidence of segregation, as both forms have important policy and practical
implications. Attempts should also be made to further unpack how the different
types of segregation are used in correctional institutions. It may be important
to know, for example, what proportion of inmates is held for administrative
versus disciplinary purposes. Finally, more research should be conducted on
the duration, frequency, and proportion of total incarceration time spent in
segregation. This information will be invaluable to correctional agencies seeking
to develop alternative strategies for the use of segregation.

Develop segregation risk/needs assessments
Differences exist in the characteristics of the inmates who are sent to segregation
units and those living in the general prisoner population. Researchers should

72 • National Institute of Justice | NIJ.gov

use this information to develop risk/needs assessment tools that can predict the
probability of an inmate being sent to segregation and identify factors that can
be targeted with intervention to decrease such risk. Priority should be given
to research that includes additional information about criminogenic needs, as
most of the prior group comparisons have been limited to data that correctional
agencies collect for other purposes (e.g., education, substance abuse). Such a
research strategy would provide useful knowledge about what types of services
might have the best effect in helping agencies reduce their use of segregation.

Conduct case studies of segregation units
One of the main ways the function of segregation has been estimated is by
analyzing the available state and federal segregation policies. However, there are
some limitations to this approach that should be acknowledged. For example,
some jurisdictions fail to provide their policies to researchers (see Butler et al.,
2013), and some have policies that do not include enough information to be
analyzed (see Metcalf et al., 2013). There is also the potential that differences may
exist between what is written in policy and what is done in practice. Another
approach to assessing the purpose of segregation is to survey correctional
officials (see Mears & Castro, 2006). This method, while informative, also does
not take into account differences that may exist between intent and practice.
Researchers need to go into the prison environment and see firsthand how these
units operate. These case studies in segregation units will not only help better
determine what role segregation plays in modern institutions but also may be
useful for determining which practices are more and less likely to help achieve
positive outcomes (e.g., improved behavior, decreases in post-release recidivism).

Increase the number and quality of empirical evaluations of segregation
More methodologically rigorous empirical evaluations are needed on the effects
of segregation. Such research should strive to investigate aggregate levels of
disorder, as well as individual-level behavioral outcomes (e.g., institutional
misconduct, post-release recidivism). This research should not only include
violent outcomes but also other less serious and nonviolent measures. It is
imperative that research in this area addresses concerns related to selection bias,
as it is well known that inmates who are sent to segregation tend to possess a
greater pre-existing disposition toward criminal behavior (Lovell et al., 2000).
Likewise, the failure to appropriately match cases on these characteristics will
likely lead to biased results. As the review of the prior empirical research shows,
studies employing weak methodological research designs tend to reveal a large
negative effect of segregation, whereas those studies with more scientifically
rigorous designs generally find no statistically significant differences between the
segregation and non-segregation groups.
It must be acknowledged that a randomized control trial — the gold standard
of research designs — is not a reasonable possibility in segregation research

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 73

because it would simply be unethical to isolate inmates purely for research
purposes, given the concerns that such placement may have a negative impact
on their mental health. It is up to researchers to continue to come up with better
methodologies and statistical techniques to obtain comparable matches for
evaluation purposes. One technique that has been particularly helpful in this
area is the use of propensity score matching. Propensity score matching affords
researchers the ability to ascertain a comparison group that is as close as possible
to the segregated group on the number of observable covariates (Rosenbaum &
Rubin, 1983; Rubin, 2006). However, it must also be understood that propensity
score matching is not a magic bullet and carries its own set of limitations.
Propensity score analysis, for example, can only make comparisons between
groups with the information that is available. Thus, this technique cannot
account for any unobserved factors that may influence placement in segregation,
which is a problem if those variables also have an impact on the outcome of
interest. Given what is known about segregation, it is likely that there are some
predictors of placement that are not captured on official records (e.g., situational
information). Therefore, obtaining good matches in the future will require not
only the use of more advanced statistical techniques but also the acquisition of
more offender and situational information. Funding efforts should be made to
encourage researchers to enter the institutions and collect this kind of data.

Prioritize research that investigates moderators
It has long been observed that the context in which segregation is delivered is
crucial to its effect on outcomes (Bonta & Gendreau, 1990). However, there is
very little research available that assesses for any differential effects of segregation
on behavioral outcomes. Research should therefore be prioritized to include
the investigation of the influence of moderators, especially for those offender
characteristics that have been the subject of recent policy changes (i.e., mentally
ill, juveniles, women). This research should also strive to include other situational
variables (e.g., physical conditions, officer-prisoner relationships, how inmates
are treated, institutional climate, reasons for being sent to segregation, health
care and treatment services, in-cell provisions, access to outside contacts), which
may also be responsible for mediating the effect of segregation on criminal
behavior (see Gendreau & Labrecque, 2016). This type of research will provide a
number of benefits to correctional agencies, including helping them to identify
which offenders may be more likely to suffer the iatrogenic effects of segregation
(e.g., young, mentally ill, and female inmates) and what types of modifications
to these units may help alleviate such negative effects (e.g., ensuring a positive
culture, increasing out-of-cell time, providing programming).

Continue to explore and evaluate changes to segregation units
As correctional systems continue to alter their segregation practices by
modifying conditions and incorporating treatment options, it is imperative

74 • National Institute of Justice | NIJ.gov

that these strategies are well documented and evaluated. Such information
will be imperative for establishing “what works” and “what does not work” in
segregation. This research will be essential for helping correctional agencies
choose which practices to adopt and which to avoid.

Conclusion
This review of the evidence finds very little support for the contention that
the use of segregated confinement (otherwise known as restrictive housing) is
responsible for reducing individual or aggregate levels of criminal behavior. The
finding of a null effect should not be misinterpreted as support for the continued
use of segregation, however, especially at its current rate in the United States.
This result, rather, calls into question the logic of relying on an expensive and
ineffective crime control strategy, when there are other potentially more viable
options available that may achieve better outcomes (e.g., principles of effective
intervention; Andrews & Bonta, 2010).
The best course of action for correctional administrators today is to use
segregation judiciously and sparingly, while striving to create a system with
little to no need for the practice in the first place. However, any effort aimed
at reforming the use of segregation must acknowledge that this practice is a
management tool that officials rely on for the effective management of jails and
prisons (Mears, 2006). In fact, many continue to insist that its use is needed to
ensure the safety and security of these institutions (e.g., Angelone, 1999; Gavora,
1996). This raises the important question, “If not for segregation, what do
corrections officials do with difficult prisoners?”
It is unlikely that any significant progress will be made in reducing the segregated
inmate population in the United States until correctional officials have alternative
options available for offenders and until they are confident that their use will not
affect institutional safety and security in a negative way. As the nation rethinks
the use and function of segregation in institutional settings, the availability of
empirical research is crucial for the development of evidence-based policies and
practices. The recommendations and conclusion reached here are a starting point
for research in this endeavor.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 75

References
Allen, F. (1964). The borderland of criminal justice: Essays in law and criminology.
Chicago, IL: University of Chicago Press.
American Bar Association. (2011). ABA standards for criminal justice: Treatment
of prisoners (3rd ed.). Washington, DC: American Bar Association.
Andrews, D. A., & Bonta, J. (2010). The psychology of criminal conduct (5th ed.).
Cincinnati, OH: Anderson Publishing.
Angelone, R. (1999). Why “super-max” prisons work. Richmond, VA: Virginia
Correctional Security Report.
Austin, J., & Irwin, J. (2012). It’s about time: America’s imprisonment binge (4th ed.).
Belmont, CA: Wadsworth.
Bauer, S. (2013, January 13). Solitary confinement: The cruelest punishment.
The Week. Retrieved from http://www.theweek.com.
Beck, A. J. (2015). Use of restrictive housing in U.S. prisons and jails, 2011-12.
Washington, DC: U.S. Department of Justice.
Blomberg, T. G., & Lucken, K. (2010). American penology: A history of control
(2nd ed.). New Brunswick, NJ: Transaction Publishers.
Bonta, J., & Gendreau, P. (1990). Reexamining the cruel and unusual punishment
of prison life. Law and Human Behavior, 14(4), 347-366.
Briggs, C. S., Sundt, J. L., & Castellano, T. C. (2003). The effect of supermaximum
security prisons on aggregate levels of institutional violence. Criminology, 41(4),
1341-1376.
Browne, A., Cambier, A., & Agha, S. (2011). Prisons within prisons: The use of
segregation in the United States. Federal Sentencing Reporter, 24(1), 46-49.
Butler, H. D., Griffin III, O. H., & Johnson, W. W. (2013). What makes you the
“worst of the worst”? An examination of state policies defining supermaximum
confinement. Criminal Justice Policy Review, 24(6), 676-694.
Butler, H. D., Johnson, W. W., & Griffin III, O. H. (2014). The treatment of the
mentally ill in supermax facilities: An evaluation of state supermax policies.
Criminal Justice and Behavior, 41(1), 1338-1353.
Butler, H. D., Steiner, B., Makarios, M., & Travis, L. (2013). Assessing the effects of
exposure to different prison environments on offender recidivism. Presented at the
annual meeting of the Academy of Criminal Justice Sciences in Dallas, TX.

76 • National Institute of Justice | NIJ.gov

Cloud, D. H., Drucker, E., Browne, A., & Parsons, J. (2015). Public health and
solitary confinement in the United States. American Journal of Public Health,
105(1), 18-26.
Crouch, B. M., & Marquart, J. W. (1989). An appeal to justice: Litigated reform of
Texas prisons. Austin, TX: University of Texas Press.
Crouch, B. M., & Marquart, J. W. (1990). Resolving the paradox of reform: Litigation,
prisoner violence, and perceptions of risk. Justice Quarterly, 7(1), 103-123.
Cullen, F. T., & Gilbert, K. E. (1982). Reaffirming rehabilitation. Cincinnati, OH:
Anderson Publishing Company.
de Beaumont, G., & de Tocqueville, A. (1833). On the penitentiary system in the
United States and its application in France. Retrieved from http://www.archive.
org.
Dickens, C. (1985) American notes. New York, NY: Fromm International
Publishing Corporation. (Original work published in 1842).
Fellner, J. (2000). Out of sight: Supermaximum security confinement in the United
States. New York, NY: Human Rights Watch.
Fellner, J., & Mariner, J. (1997). Cold storage: Super-maximum security
confinement in Indiana. New York, NY: Human Rights Watch.
Ferdik, F. (2015). The relationship between correctional officer safety and
wellness and administrative segregation. Presented at the National Institute of
Justice meeting on the use of administrative segregation in the United States in
Arlington, VA.
Foucault, M. (1995). Discipline and punish: The birth of the prison. New York, NY:
Random House.
Franke, H. (1992). The rise and decline of solitary confinement: Socio-historical
explanations of long-term penal changes. The British Journal of Criminology,
32(2), 125-143.
Frost, N. A., & Monteiro, C. E. (2016). Administrative segregation in prison.
Washington, DC: U.S. Department of Justice.
Gavora, J. (1996, August 25). Violent offenders should be placed in supermax
prisons. The Washington Post. Retrieved from http://www.washingtonpost.com.
Gawande, A. (2009, March 30). Hellhole. The New Yorker. Retrieved from http://
www.newyorker.com.
Gendreau, P., & Bonta, J. (1984). Solitary confinement is not cruel and unusual
punishment: Sometimes people are! Canadian Journal of Criminology, 26(4), 467-478.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 77

Gendreau, P., & Labrecque, R. M. (2016). The effects of administrative
segregation: A lesson in knowledge cumulation. In J. Wooldredge & P. Smith
(Eds.), Oxford handbook on prisons and imprisonment. Oxford, UK: Oxford
University Press.
Goode, E. (2012, March 30). Fighting a drawn-out battle against solitary
confinement. The New York Times. Retrieved from http://www.nytimes.com.
Grassian, S. (1983). Psychopathological effects of solitary confinement. American
Journal of Psychiatry, 140(11), 1450-1454.
Guenther, L. (2012, August 26). The living death of solitary confinement. The
New York Times. Retrieved from http://www.nytimes.com.
Haney, C. (1993). “Infamous punishment”: The psychological consequences of
isolation. National Prison Project Journal, 8(2), 3-7, 21.
Haney, C. (1997). Psychology and the limits of prison pain: Confronting the
coming crisis of the Eighth Amendment law. Psychology, Public Policy, and Law,
3(4), 499-588.
Haney, C. (2003). Mental health issues in long-term solitary and “supermax”
confinement. Crime & Delinquency, 49(1), 124-156.
Haney, C. (2008). A culture of harm: Taming the dynamics of cruelty in
supermax prisons. Criminal Justice and Behavior, 35(8), 956-984.
Haney, C. (2009). The social psychology of isolation: Why solitary confinement is
psychologically harmful. Prison Service Journal, 181, 12-20.
Haney, C. (2012a). Testimony of Professor Craig Haney to the Senate Judiciary
Subcommittee on the Constitution, civil rights, and human rights hearing on
solitary confinement. Retrieved from http://www.judiciary.senate.gov.
Haney, C. (2012b). Prison effects in the era of mass incarceration. The Prison
Journal. Advance online publication. doi: 10.1177/0032885512448604.
Haney, C., & Lynch, M. (1997). Regulating prisons of the future: The
psychological consequences of supermax and solitary confinement. New York
University Review of Law and Social Change, 23, 477-570.
Haney, C., Weill, J., Bakhshay, S., & Lockett, T. (2016). Examining jail isolation: What
we don’t know can be profoundly harmful. The Prison Journal, 96(1), 126-152.
Harrington, M. P. (2015). Methodological challenges to the study and
understanding of solitary Confinement. Federal Probation, 79(3), 45-47.

78 • National Institute of Justice | NIJ.gov

Helmus, L. (2015). Developing and validating a risk assessment scale to predict
inmate placements in administrative segregation in the Correctional Service of
Canada. Unpublished Ph.D. dissertation, Carleton University, Ottawa, ON.
Helmus, L., Johnson, S., & Harris, A. J. R. (2014). Developing the Risk of
Administrative Segregation Tool (RAST) to predict admissions to segregation.
Ottawa, ON: Solicitor General Canada.
Henningsen, R. J., Johnson, W. W., & Wells, T. (1999). Supermax prisons:
Panacea or desperation? Corrections Management Quarterly, 3(2), 53-59.
Huebner, B. M. (2003). Administrative determinants of inmate violence: A multilevel
analysis. Journal of Criminal Justice, 31, 107-117.
Ignatieff, M. (1983). State, civil society and total institutions: A critique of
recent social histories of punishment. In S. Cohen & A. Scull (Eds.), Social
control and the state: Historical and comparative essays. Oxford, UK: Martin
Robertson & Company.
Jackson, M. (1983). Prisoners of isolation: Solitary confinement in Canada.
Toronto, ON: University of Toronto Press.
Kann, M. E. (2005). Punishment, prisons, and patriarchy: Liberty and power in the
early American republic. New York, NY: New York University Press.
Keim, B. (2013, July 10). The horrible psychology of solitary confinement.
Wired. Retrieved from http://www.wired.com.
King, R. D. (1999). The rise and rise of supermax: An American solution in
search of a problem? Punishment and Society, 1(2), 163-186.
Kupers, T. A. (2008). What to do with the survivors? Coping with the long-term
effects of isolated confinement. Criminal Justice and Behavior, 35(8), 1005-1016.
Kurki, L., & Morris, N. (2001). The purposes, practices, and problems of
supermax prisons. Crime and Justice, 28, 385-424.
Labrecque, R. M. (2015a). The effect of solitary confinement on institutional
misconduct: A longitudinal evaluation. Washington, D.C.: U.S. Department of Justice.
Labrecque, R. M. (2015b). Who ends up in administrative segregation? Predictors
and other characteristics. Paper presented at the National Institute of Justice
meeting on the use of administrative segregation in the United States in
Arlington, VA.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 79

Labrecque, R. M., & Smith, P. (2013). Advancing the study of solitary confinement.
In J. Fuhrman & S. Baier (eds.), Prisons and prison systems: Practices, types, and
challenges (pp. 57-70). Hauppauge, NY: Nova Science Publishers.
Labrecque, R. M., & Smith, P. (2015). Unpacking the ‘black box’ of solitary
confinement: A five-year longitudinal study of inmates in Ohio. Paper presented at
the Academy of Criminal Justice Sciences in Orlando, FL.
Labrecque, R. M., Smith, P., & Gendreau, P. (2013). The effects of solitary
confinement on prisoners in custodial settings: A meta-analysis. Paper presented
at the Academy of Criminal Justice Sciences in Dallas, TX.
Lieber, F. (2010). A popular essay on subjects of penal law and on uninterrupted
solitary confinement at labor, as contradistinguished to solitary confinement at
night and joint labor by day, in a letter to John Bacon. Port Chester, NY: Adegi
Graphics. (Original work published in 1838).
Liman Program & Association of State Correctional Administrators. (2015).
Time-in-cell: ASCA-Liman 2014 national survey of administrative segregation in
prison. New Haven, CT: Yale Law School.
Lovell, D. (2008). Patterns of disturbed behavior in a supermax prison. Criminal
Justice and Behavior, 35(8), 985-1004.
Lovell, D., Cloyes, K., Allen, D., & Rhodes, L. (2000). Who lives in supermaximum custody? A Washington state study. Federal Probation, 64(2), 33-38.
Lovell, D., Johnson, L. C., & Cain, K. C. (2007). Recidivism of supermax
prisoners in Washington State. Crime and Delinquency, 53(4), 633-656.
McGinnis, K., Austin, J., Becker, K., Fields, L., Lane, M., Maloney, M., …, Felix,
T. (2014). Federal Bureau of Prisons: Special housing unit review and assessment.
Arlington, VA: CAN.
McGowen, R. (1998). The well-ordered prison: England, 1780-1865. In N. Morris
& D. J. Rothman (Eds.), The Oxford history of the prison: The practice of punishment
in western society (pp. 71-99). New York, NY: Oxford University Press.
Mears, D. P. (2006). Evaluating the effectiveness of supermax prisons. Washington,
DC: Urban Institute, Justice Policy Center.
Mears, D. P. (2008). An assessment of supermax prisons using an evaluation
research framework. The Prison Journal, 88(1), 43-68.
Mears, D. P. (2013). Supermax prisons: The policy and the evidence. Criminology
and Public Policy, 12(4), 681-719.
Mears, D. P., & Bales, W. D. (2009). Supermax incarceration and recidivism.
Criminology, 47(4), 1131-1166.

80 • National Institute of Justice | NIJ.gov

Mears, D. P., & Bales, W. D. (2010). Supermax housing: Placement, duration, and
time to re-entry. Journal of Criminal Justice, 38(4), 545-554.
Mears, D. P., & Castro, J. L. (2006). Wardens’ views on the wisdom of supermax
prisons. Crime and Delinquency, 52(3), 398-431.
Mears, D. P., Mancini, C., Beaver, K. M., & Gertz, M. (2013). Housing the
“worst of the worst” inmates: Public support for supermax prisons. Crime and
Delinquency, 59(4), 587-615.
Mears, D. P., & Watson, J. (2006). Towards a fair and balanced assessment of
supermax prisons. Justice Quarterly, 23(2), 232-269.
Metcalf, H., Morgan, J., Oliker-Friedland, S., Resnik, J., Spiegel, J., Tae, H., Work,
A. R., & Holbrook, B. (2013). Administrative segregation, degrees of isolation, and
incarceration: A national overview of state and federal correctional policies. New
Haven, CT: Liman Public Interest Program.
Miller, M. B. (1980). At hard labor: Rediscovering the 19th-century prison. In T.
Platt & P. Takagi (Eds.), Punishment and penal discipline (pp. 79-88). Berkeley,
CA: Crime and Social Justice Associates.
Morgan, R. D., Van Horn, S. A., MacLean, N., Bolanos, A., Gray, A. L., Batastini,
A., & Mills, J. F. (2014). Administrative segregation: Is it a harmful correctional
practice? Presented at the annual meeting of the American Psychology-Law
Society in New Orleans, LA.
Morris, R. G. (2016). Exploring the effect of exposure to short-term solitary
confinement among violent prison inmates. Journal of Quantitative Criminology,
32(1), 1-22.
Naday, A., Freilich, J. D., & Mellow, J. (2008). The elusive data on supermax
confinement. The Prison Journal, 88(1), 69-93.
National Institute of Corrections. (1997). Supermax housing: A survey of current
practice. Longmont, CO: U.S. Department of Justice.
O’Keefe, M. L. (2008). Administrative segregation from within: A corrections
perspective. The Prison Journal, 88(1), 123-143.
Pisciotta, A. W. (1994). Benevolent repression: Social control and the American
reformatory prison movement. New York, NY: New York University Press.
Pizarro, J. M., & Stenius, V. M. K. (2004). Supermax prisons: Their rise, current
practices, and effect on inmates. The Prison Journal, 84(2), 248-264.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 81

Pizarro, J. M., Stenius, V. M. K., & Pratt, T. C. (2006). Supermax prisons: Myths,
realities, and the politics of punishment in American society. Criminal Justice
Policy Review, 17(1), 6-21.
Pizarro, J. M., Zgoba, K. M., & Haugebrook, S. (2014). Supermax and recidivism:
An examination of the recidivism covariates among a sample of supermax exinmates. The Prison Journal, 94(2), 180-197.
Ralph, P. H., & Marquart, J. W. (1991). Gang violence in Texas prisons. The Prison
Journal, 71(2), 38-49.
Reisig, M. D. (1998). Rates of disorder in higher-custody state prisons: A comparative
analysis of managerial practices. Crime and Delinquency, 44(2), 229-244.
Richards, S. C. (2008). USP Marion: The first federal supermax. The Prison
Journal, 88(1), 6-22.
Riveland, C. (1999). Supermax prisons: Overview and general considerations.
Washington, DC: National Institute of Corrections.
Rodriguez, N. (2015). Director’s corner: Setting the agenda for administrative
segregation research. Washington, DC: U.S. Department of Justice.
Rogers, R. (1993). Solitary confinement. International Journal of Offender
Therapy and Comparative Criminology, 37(4), 339-349.
Rosenbaum, P. R., & Rubin, D. B. (1983). The central role of the propensity score
in observational studies for causal effects. Biometrika, 70(1), 41-55.
Rossi, P. H., Lipsey, M. W., & Freeman, H. E. (2004). Evaluation: A systemic
approach (7th ed.). Thousand Oaks, CA: Sage.
Rothman, D. J. (1971). The discovery of the asylum: Social order and disorder in
the new republic. Boston, MA: Little, Brown.
Rothman, D. J. (1980). Conscience and convenience: The asylum and its
alternatives in progressive America. Boston, MA: Little, Brown.
Rothman, D. J. (1998a). The invention of the penitentiary. In T. J. Flanagan, J. W.
Marquart, & K. G. Adams (eds.), Incarcerating criminals: Prisons and jails in social
and organizational context (pp. 15-23). London, UK: Oxford University Press.
Rothman, D. J. (1998b). Perfecting the prison: United States, 1789-1865. In
N. Morris & D. J. Rothman (eds.), The Oxford history of the prison: The
practice of punishment in western society (pp. 100-116). New York, NY: Oxford
University Press.

82 • National Institute of Justice | NIJ.gov

Rotman, E. (1990). Beyond punishment: A new view of rehabilitation of criminal
offenders. New York, NY: Greenwood Press.
Rotman, E. (1998). The failure of reform: United States, 1865-1965. In N. Morris &
D. J. Rothman (eds.), The Oxford history of the prison: The practice of punishment in
western society (pp. 151-177). New York, NY: Oxford University Press.
Rubin, D. B. (2006). Matched sampling for causal effects. New York, NY:
Cambridge University Press.
Scharff-Smith, P. (2006). The effects of solitary confinement on prison inmates: A
brief history and review of the literature. Crime and Justice, 34(1), 441-528.
Seale, L., Atkinson, J., Grealish, B., Fitzgerald, T., Grassel, K., & Viscuso, B.
(2011). California Department of Corrections and Rehabilitation: 2011 adult
institutions outcome evaluation report. Sacramento, CA: Office of Research,
Research and Evaluation Branch.
Shalev, S. (2008). A sourcebook on solitary confinement. London, UK: Mannheim
Centre for Criminology.
Shalev, S. (2009). Supermax: Controlling risk through solitary confinement.
Portland, OR: Willan Publishing.
Shames, A. Wilcox, J., & Subramanian, R. (2015). Solitary confinement:
Common misconceptions and emerging safe alternatives. New York, NY: VERA
Institute of Justice.
Shepperd, R. A., Geiger, J. R., & Welborn, G. (1996). Closed maximum security:
The Illinois supermax. Corrections Today, 58(4), 84-87.
Smith, P., Gendreau, P., & Labrecque, R. M. (2015). The impact of solitary
confinement on inmate behavior: A meta-analytic review. Paper presented at the
North American Correctional and Criminal Justice Psychology (N3) Conference,
Ottawa, Canada.
Steiner, B., & Wooldredge, J. (2008). Inmate versus facility effects on prison rule
violations. Criminal Justice and Behavior, 35(4), 438-456.
Suedfeld, P., Ramirez, C., Deaton, J., & Baker-Brown, G. (1982). Reactions and
attributes of prisoners in solitary confinement. Criminal Justice and Behavior,
9(3), 303-340.
Sundt, J. L., Castellano, T. C., & Briggs, C. S. (2008). The sociopolitical context
of prison violence and its control: A case study of supermax and its effect in
Illinois. The Prison Journal, 88(1), 94-122.
Toch, H. (1978). Is a “correctional officer,” by any other name, a “screw”?
Criminal Justice Review, 3(2), 19-35.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 83

Toch, H. (2001). The future of supermax confinement. The Prison Journal, 81(3),
376-388.
Toch, H. (2003). The contemporary relevance of early experiments with
supermax reform. The Prison Journal, 83(2), 221-228.
United Nations. (2015) United Nations standard minimum rules for the treatment
of prisoners (the Mandela rules). New York, NY: United Nations.
Useem, B., & Kimball, P. (1991). States of siege: U.S. prison riots 1971-1986.
New York, NY: Oxford University Press.
Useem, B., & Piehl, A. M. (2006). Prison buildup and disorder. Punishment & Society,
8(1), 87-115.
Useem, B., & Reisig, M. D. (1999). Collective action in prisons: Protest,
disturbances, and riots. Criminology, 37(4), 735-760.
Ward, D. A. (2009). Alcatraz: The gangster years. Berkeley, CA: University of
California Press.
Ward, D. A., & Werlich, T. G. (2003). Alcatraz and Marion. Punishment and
Society, 5(1), 53-75.
Wormith, J. S., Tellier, M. C., & Gendreau P. (1988). Characteristics of protective
custody offenders in a provincial correctional center. Canadian Journal of
Criminology, 30, 39-58.
Zinger, I. (2013). Segregation in Canadian federal corrections: A prison
ombudsman’s perspective. Presented at the International Conference on Human
Rights and Solitary Confinement in Winnipeg, Manitoba, Canada.
Zinger, I., Wichmann, C., & Andrews, D. A. (2001). The psychological effects of 60
days in administrative segregation. Canadian Journal of Criminology, 43(1), 47-83.

8 4 • National Institute of Justice | NIJ.gov

CHAPTER 3

The Conditions of Confinement
in Restrictive Housing
Holly Foster, Ph.D.
Texas A&M University

Restrictive Housing: Purpose, Terms,
and Report Objectives

R

estrictive housing practices are used in American correctional institutions
to manage and contain perceived threats to the safety of inmates and staff,
and to keep order in the facility. The process involves segregating some
inmates from the general prisoner population under specific circumstances,
including violence and disruption (U.S. Department of Justice, 2016). Other
than general overviews, however, little systematic and comparative information
is available on the conditions of confinement involved in restrictive housing.
This report will cover the following conditions of confinement: (1) physical
and temporal dimensions, (2) incentives and disciplinary sanctions used, (3)
social circumstances (e.g., family contacts), (4) psychological conditions, and
(5) service provision (e.g., medical and educational services for prisoners). The
absence of information about confinement conditions means that practitioners
and researchers lack solid data on which aspects are especially influential on
prisoners in both the short term and over time. More study of this topic would
provide practitioners and researchers with systematic knowledge of how the
conditions work to affect prisoner adaptation. This knowledge may assist
practitioners in reducing the harmful effects that prisoners may incur from

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 85

living in these conditions. This report provides a synthesis of extant research
on the conditions of confinement in restrictive housing to identify the current
limitations of the practice and point to ways to address them.
Court decisions (e.g., Madrid v. Gomez; Simon, 2014) indicate that restrictive
housing conditions are unconstitutional for mentally ill inmates. Systematic
reviews are needed of the conditions that inmates face under different forms of
restrictive housing over time. Some descriptive information on current conditions
in restrictive housing is found in the U.S. Department of Justice’s recent (2016)
review, Report and Recommendations Concerning the Use of Restrictive Housing,
and other key reports (Liman Program/Association of State Correctional
Administrators [ASCA], 2015; New York Civil Liberties Union [NYCLU], 2015).
Yet, missing from the literature is systematic research on how prisoners experience
the conditions described. Such information would better inform the knowledge
base regarding the levels and types of stress that prisoners face.
More detailed information is needed on the specific conditions that inmates face
when placed in restrictive housing and the degree to which these conditions
are similar or vary across its various forms. There are three major types of
restrictive housing practices that reflect different correctional intentions (Shalev,
2009, p. 2) and that may involve both similarities and differences. Punitive
segregation is used to discipline prisoner misconduct, usually for a period
following a disciplinary hearing. Protective segregation (and custody) is used to
house and separate vulnerable prisoners from the general population for their
own protection. Administrative segregation is imposed for managerial reasons.
It separates inmates who are deemed to be high risk from the general prison
population — often for long periods (e.g., perceived gang members, members
of security threat groups) (Liman Program/ASCA, 2015, p. 1). Additionally,
supermaximum or “supermax” facilities are those designed to hold prisoners in
long-term restrictive housing (Shalev, 2009, p. 9). When implemented for these
three categories, restrictive housing reduces interaction with other inmates,
limits programming opportunities, and restricts an inmate’s privileges (Beck,
2015). Myriad terms are used to describe these conditions; this report will use
“restrictive housing” to refer to all of them. However, the terms in the study’s
source documents will be used when referring to specific research studies
when necessary.
To further synthesize and critique the extant research on conditions of
confinement in restrictive housing, this white paper adopts Simon’s (2014)
explanation for the rising rates of incarceration in the United States from the
1970s to approximately 2009 (Garland, 2001; Clear & Frost, 2013). It also
evaluates the practice through qualitative research, involving analyses of
subjective data such as personal stories, and quantitative research, in which
numerical data are analyzed. As Simon indicates, the qualitative story of mass
incarceration is revealed through an analysis of major court decisions, with the
quantitative story revealed through statistical trends and patterns. The qualitative

86 • National Institute of Justice | NIJ.gov

story may also be investigated through interviews with and observations of
those in restrictive housing. The quantitative story, however, requires testing
hypotheses and relationships among variables gathered through data collection.
Criminological and sociological theories should guide the understanding of both
the qualitative and quantitative information about restrictive housing. For the
qualitative story, there is a need to list and describe the commonalities and the
different experiences faced by various groups in restrictive housing, including
those of overrepresented racial and ethnic minority groups, and often overlooked
female and transgender prisoners. For the quantitative story, comparative designs
are needed to more fully understand the frustrations and deprivations faced
by those living in prison and the types and levels of “pains of imprisonment”
that arise over time Sykes (1958/2007). Moving beyond description, a more
empirically based, quantitative approach guided by theory would show how the
pains of imprisonment affect prisoner outcomes, and adaptations for those living
in restrictive housing (Agnew, 2001, 2006; Pearlin, Menaghan, Leiberman, &
Mullan, 1981; Sykes, 1958/2007).
With very limited information available on restrictive housing conditions,
this paper will draw on findings from several key reports on the topic (Liman
Program/ASCA, 2015; NYCLU, 2012). These sources illuminate the conditions in
restrictive housing; they also highlight the need for more information about these
conditions and the repercussions they may have for inmates. For example, the
most systematic resource on conditions of administrative segregation is the data
gathered from correctional administrators by Yale University’s Liman Program
and the Association of State Correctional Administrators (ASCA). Although the
Liman Program/ASCA study is an important source of information, it does not
include prisoners’ experiences of the conditions of confinement in restrictive
housing that may be gleaned through the qualitative and quantitative approaches
proposed here.
The report further identifies a need for research to explore gender differences
and similarities pertaining to inmates’ experiences of restrictive housing.
Arguments are emerging about the need for gender-responsive programming
in the criminal justice system (Covington & Bloom, 2006), although not yet
in the realm of restrictive housing. Literature indicates that restrictive housing
conditions were established to house “dangerous men” (Simon, 2014). However,
within the limited information about restrictive housing, even less is known
about how women experience these conditions. A lack of systematic attention
to gender perpetuates the tendency to overlook the needs of female offenders
in programming decisions, as criminal justice services are usually designed
with the needs of men in mind (Covington & Bloom, 2006). This report uses
data regarding female inmates in administrative segregation made available for
analysis by the Liman Program, but not published elsewhere, alongside published
information about the majority of those individuals living in these conditions — men
(Liman Program/ASCA, 2015). A descriptive look at the conditions of restrictive
housing for female prisoners suggests further research is needed, in particular,

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 87

to develop the gendered qualitative stories of restrictive housing. This focus on
gender also leads to further consideration of other elements of social location
(e.g., race and ethnicity) that may shape experience.

Trends and Patterns in Restrictive Housing
Part of the quantitative story surrounding restrictive housing is revealed by a
description of trends and patterns regarding who tends to experience them,
which links further back to trends in mass incarceration. As the National
Research Council report (2014) documents, increases in incarceration rates
in the United States are the result of policy changes, such as the War on
Drugs, and changes from indeterminate to determinate sentencing practices
constituting more punitive approaches to criminal justice. With these policy
changes came a concomitant rise in the use of restrictive housing beginning
in the 1980s. Shalev (2009) reports that all forms of prisoner segregation
increased in the 1990s. Between 1995 and 2000, in particular, the number of
prisoners who were isolated from the general prison population rose by 40
percent (p. 2). The most recent estimates indicate that approximately 80,000
to 100,000 people are confined in isolated conditions in U.S. prisons (Liman
Program/ASCA, 2015, p. 3). Estimates across American jurisdictions report
that the percentage of the custodial population confined in restrictive housing
ranges from 2.1 percent to 14.2 percent (Liman Program/ASCA, 2015, p. 15).
Furthermore, using data gathered from inmates as part of the National Inmate
Survey (NIS), roughly 10 percent of all prison inmates and 5 percent of all jail
inmates have spent 30 or more days in restrictive housing (Beck, 2015).
The survey also found race and ethnic disparities among the inmates who
experience restrictive housing conditions. Black prison inmates were more
likely than white and Latino prison inmates to have spent time in restrictive
housing. Furthermore, inmates with lower education levels (i.e., without a high
school diploma), young inmates, and lesbian, gay, and bisexual inmates were
more likely to have spent time in restrictive housing than were more educated,
older, and heterosexual inmates (Beck, 2015, p. 4). However, the NIS data show
no difference in the percentage of each gender having spent time in restrictive
housing (Beck, 2015, p. 4).
Further descriptive statistical information on inmates placed in administrative
segregation is available by prisoner gender, race, and ethnicity and reflects
broader demographic trends in mass imprisonment. For example, the data show
that administrative segregation is more commonly used with male prisoners
(7.5 percent) than female prisoners (0.1 percent) (Liman Program/ASCA,
2015, p. 17). The median estimates for female prisoners held in administrative
segregation is less than 1 percent, although in one jurisdiction, a high of 6.4
percent of the female custodial population is held in these conditions (Liman
Program/ASCA, 2015, p. 20).

88 • National Institute of Justice | NIJ.gov

The use of administrative segregation also varies by race and ethnicity. Research
shows higher percentages of black and Latino men in administrative segregation
than in the general population of prisons across responding jurisdictions (Liman
Program/ASCA, 2015, p. 30). Furthermore, black women are overrepresented in
administrative segregation compared to the general female custodial population
(Liman Program/ASCA, 2015, p. 36).
The NIS data also reveal criminal justice status and history among inmates held
in restrictive housing compared to other inmates. Those in restrictive housing
are more likely to have been sent to prison for a violent offense than for other
offenses (Beck, 2015, p. 5). Inmates in restrictive housing are also more likely
to have been arrested more than once and to have had prior incarcerations as
juveniles (Beck, 2015, p. 5).
The use of restrictive housing not only increased in prevalence alongside
mass incarceration, but it also became increasingly severe with increased
technological surveillance and more extreme facilities (Shalev, 2009), including
the development and growth of supermax security prisons involving extended
periods of isolation. In the 1980s, there were few supermax facilities other
than those in the two founding locations — Marion, Illinois, and Florence,
Arizona. However, as of 2005, as many as 44 states had these facilities (Shalev,
2009; Simon, 2014; Mears, 2005), with roughly 25,000 people held in supermax
security conditions (Mears, 2013). Mears (2013) and Western (2007) link the
policy context of mass incarceration with the increase of these facilities.
During this period, states also moved away from rehabilitative ideals in prison
programming and toward “total incapacitation,” which emphasizes sending
people to prison to prevent crime rather than using other approaches (Haney,
2003; Simon, 2014). Simon argues that the rise of the supermax prison helped
to legitimize mass incarceration, with an emphasis on locking away “dangerous
men” (2014, p. 52).
Despite the considerable variation in the types of restrictive housing used
in federal and state correctional systems across America, they share some
common features (Haney, 2009; Liman Program/ASCA, 2015, p. 1). This report
next provides a synthesis of what is currently known about the conditions
of confinement in restrictive housing facilities, as well as an assessment
of the limitations of extant knowledge. Academic books, journal articles,
and comprehensive reports by research institutes constitute the basis of the
information for this review. Furthermore, because information continues to
emerge on proposed changes in restrictive housing practices in some states (e.g.,
California, New York), newspaper coverage is also used to update information on
the conditions of confinement.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 89

How Can Theory Be Used To Understand
Restrictive Housing Conditions?
Theory is a powerful tool for guiding research, but it could also be used by
practitioners as an integrative framework in working with restrictive housing
practices. Theory may assist in considering the substantive meaning of the
conditions of confinement in restrictive housing. Furthermore, since theoretical
propositions can be tested through data analysis, they can, therefore, be found
by evidence to be false, which occurs when empirical results do not support the
theorized relationships (Bernard, Snipes, & Gerould, 2010). When theories are
not supported empirically, they can be refined and subjected to further testing.
However, if certain theoretical propositions are consistently supported by data,
practitioners can be confident in the robustness of those results — providing a
compelling rationale for decision-making with scarce resources. Furthermore,
theory may provide a framework for how things work in the short and long
term, with practical implications for when and why to adopt or curtail specific
programs and practices.
One such theoretical tool for further synthesizing information on the conditions
of confinement in restrictive housing is provided by the focus on stressors faced
in prison, as highlighted by Robert Agnew’s general strain theory (GST) (2001;
2006) for understanding how strains (or stresses) are associated with criminal
behavior. Recent directions in GST apply this framework to understanding
criminal justice settings (Blevins, Listwan, Cullen, & Johnson, 2010; Delisi, 2011;
Agnew & Delisi, 2012), but not (so far) to the conditions of restrictive housing.
Theorizing the connections between restrictive housing conditions and GST is
one objective of this report.
Further theoretical insights are provided by classic research conducted more
generally in prisons by Sykes (1958/2007) in his conceptualization of the “pains
of imprisonment.” GST attends to the implications of stressors and strains or
events and conditions that are disliked (Agnew, 2006, p. 4). Among myriad
strains that individuals generally face in prison, two are the most prominent:
being treated in an aversive manner and losing something of value (Agnew,
2006). Prisoners in restrictive housing experience loss of privileges and extreme
isolation resulting in a lack of meaningful human contact (Smith, 2006) — both
of which qualify as strains or pains of imprisonment. In fact, a recent book
develops the use of isolation as a contemporary “pain of mass imprisonment”
(Fleury-Steiner & Longazel, 2014, emphasis added). However, further research
is needed to understand the specific types of strains experienced by those in
restrictive housing as well as their totality.
Sykes’ (1958/2007) classic research in a men’s maximum-security prison
identified five primary pains of imprisonment: (1) the deprivation of liberty,
which includes restricted freedom of movement and isolation from friends,
family, and relatives (p. 65); (2) the deprivation of goods and services; (3) the

9 0 • National Institute of Justice | NIJ.gov

frustration of sexual desire (originally described in terms of lack of access to
heterosexual relationships); (4) the deprivation of autonomy or the restricted
ability to make choices (p. 73); and (5) the deprivation of security, where
the prisoner may need to fight for safety and possessions. Insight into the
contemporary pains of imprisonment for prisoners in the United States will be
enhanced by considering the more extreme circumstances of restrictive housing.
Moreover, consideration of the pains of imprisonment in restrictive housing
would best include comparisons to the general prison population. Although the
conditions of confinement in restrictive housing differ from the circumstances in
which general population prisoners are held, both entail pains of imprisonment.
The recent report by the National Research Council (2014) includes an
overview of conditions of confinement in prisons as environments tasked with
maintaining order and safety, as well as meeting punishment and reformation
goals. Conditions of confinement include a range of stressors, such as “material
deprivations; restricted movement and liberty; a lack of meaningful activity;
a nearly total absence of personal privacy; and high levels of interpersonal
uncertainty, danger, and fear” (p. 174).
Stressors encountered in prison, and those experienced prior to prison, can lead
to mental health problems, according to criminal justice models of importation
(taking into account pre-prison influences and experiences) and deprivation
(taking into account influences and experiences encountered in prison)
(Goodstein and Wright, 1989). Both importation strains (stressors experienced
prior to prison) and deprivation strains (stressors experienced in prison)
influence an inmate’s adjustment to prison, according to criminal justice models
and the framework of GST (Agnew, 2006; Foster, 2012). The importance of
further research into deprivation strains is described below.
General strain theory also includes a key role for coping resources that may
offset the influences of strains on antisocial behavior (Agnew, 2006; Pearlin
et al., 1981). Accordingly, although prisoners in the general population are
subject to stressors and strains, they also have access to some mitigating factors
that may include sharing a cell; having some contact with other prisoners in
particular areas at designated times; and being offered some degree of vocational,
educational, and therapeutic programs (Shalev, 2009). Furthermore, classic
sociological research on prisoners reveals the importance of social interactions
with other inmates in mitigating the pains of imprisonment, particularly when
prisoners take on specific social roles and develop interpersonal relationships
to contribute to the social order in men’s and women’s prisons (Clemmer,
1940/1965; Giallombardo, 1966; Heffernan, 1972; Sykes, 1958/2007). These early
findings suggest that efforts to increase prisoner coping resources in restrictive
housing may lead to more orderly settings as well, by offsetting the pains of
imprisonment. Little is known about inmate and institutional coping resources
available to prisoners held in restrictive housing conditions. Contemporary studies
of imprisoned men describe how prison gangs have become a means of achieving

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 91

order in prisons (Skarbek, 2014). Yet, exploring other means of establishing order
by fostering coping resources, even in extreme circumstances, may be more
conducive to inmate and staff well-being according to GST premises.
Furthermore, confinement conditions within restrictive housing vary, from less
to more restrictive and harsh circumstances regarding the severity of isolation,
amount of deprivation, number of restrictions, and degree of degradation
(Haney, 2009; National Research Council, 2014). During the development and
growth of supermax prisons in the mid-1980s, a continuum from lesser to
harsher and more restrictive conditions developed, with the harshest conditions
incorporating intense social isolation and control (Browne, Cambier, and Agha,
2011, p. 47). A synthesis of what is known about the conditions of confinement
in restrictive housing will begin to describe some of the pains of imprisonment in
these circumstances and provide a basis for future comparative research.
In addition to using GST and the pains of imprisonment framework, another
theoretical tool for understanding restrictive housing is provided by examining
critical sociological perspectives focused on the body (Foucault, 1977; Wacquant,
2004). These perspectives draw attention to “embodied experience,” which
acknowledges that extreme restrictions enacted over the body, such as those
used in restrictive housing, also have a direct effect on prisoners’ cognitive and
emotional health. A connection with more critical sociological perspectives
provides a basis for interpreting how the pains of imprisonment may vary by
social location (e.g., across demographic groups). This perspective underpins
a call for more information on prisoners’ experiences in restrictive housing to
bring about more just and humane policies and practices. It will also provide
insight into potentially diverse experiences within restrictive housing. This
direction of research will begin to fill some of the missing aspects of the
qualitative story surrounding restrictive housing.

Synthesis and Critique of Evidence on
Restrictive Housing Conditions
Data on the conditions of confinement in restrictive housing are generally
lacking and are mostly embedded within broader reports and articles, rather
than being the focus of empirical inquiry. For example, extant resources include
some detailed observational information on conditions in supermax facilities
(e.g., Shalev, 2009); however, these findings are often descriptive in nature. The
most nationally informative and recent data on restricted housing conditions
are available for conditions in administrative segregation, specifically, where
comparative, survey-based information has been gathered across jurisdictions
(Liman Program/ASCA, 2015). These data were gathered through surveys
administered to correctional officials in 46 responding jurisdictions across the
United States. That report, Time-In-Cell: The ASCA-Liman 2014 National Survey
of Administrative Segregation in Prison, follows a systematic review of policies

92 • National Institute of Justice | NIJ.gov

covering the conditions of confinement in administrative segregation, which
was also conducted by the Liman Program (2013). Key findings of the Liman
Program/ASCA data are synthesized here to provide a rare glimpse into extant
conditions. New information also will be gleaned from data made available by
the Liman Program/ASCA on the 10 jurisdictions that reported information
on administrative segregation conditions for women: Delaware, Georgia, Iowa,
Kansas, Montana, Nevada, South Carolina, Texas, Wisconsin, and Wyoming.
The Liman Program/ASCA data pertain to institutionally reported information,
but they also contain some sparse information from prisoners themselves.
The prisoner-reported data tend to be based on convenience samples, which
are samples composed of participants who are accessible but not randomly or
representatively selected. One key report is based on communication with 100
prisoners who served time in punitive segregation in New York (NYCLU, 2015).
However, it is limited to men’s experiences, as 99 percent of the prisoners in
punitive segregation, or “extreme isolation,” in New York state are male.
Some sparse qualitative research is also available, the most prominent being
Rhodes’ (2004) work with incarcerated men. The only available qualitative
research on the restrictive housing of incarcerated women is embedded in
a larger (but older) study of a women’s prison in California (Owen, 1998).
Information about women in restrictive housing was included as an aside and
was not the purpose of the study. Qualitative research may involve smaller
numbers of cases, but it yields more nuanced information to add to the
knowledge base.
One question that has not been addressed — but should be — is to what
degree are there similarities and differences across genders? In the Liman
Program/ASCA data, 26 jurisdictions across the United States answered a
general question about whether facilities have different staffing, programming,
or privileges in administrative segregation for women and men, which
elicited qualitative, text-based responses.1 Most correctional administrators
indicated that there was no difference in the conditions for men and women in
administrative segregation. Among the responses that support this point, the
following were selected as examples:
• Alaska: “Administrative Segregation policy and procedure applies to both
male and female prisoners, without variance.”
• Connecticut: “There are not any notable differences between how the males
and females are confined or managed.”
1

The questionnaire item reads, “If your system also houses female inmates in administrative segregation, please
describe any differences in the facilities, staffing, programming, privileges, or other aspects of confinement that
differ from what you have described above” (Liman Program, 2015, Appendix C (Section III): http://www.law.yale.
edu/system/files/area/center/liman/document/liman-asca_adseg_appendix_cappendix_c.pdf.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 93

• Florida: “No differences — same as males.”
• Nebraska: “We have the same policies and physical layout for restrictive
housing at the men’s and women’s facilities.”
Yet, other qualitative responses suggest that there are some differences in
conditions for men and women in administrative segregation:
• Utah: “Restrictive housing is similar to the males. One major difference is the
step-down process, which is not as extensive or structured. Programming is
about the same.”
• Nevada: “There is only [one] institution that houses female inmates for
administrative segregation. Limited housing and there is no other institution
that the inmate can be transferred to.”
Therefore, the qualitative data suggest that similarities exist across genders, but
they also indicate some differences that seem to pertain to women’s smaller
numbers in administrative segregation conditions.

Physical and Temporal Dimensions
Supermax conditions also vary, but there are some commonalities of space
that adhere to legal standards requiring that cells measure 70-80 square feet
(Shalev, 2009). In California, Pelican Bay State Prison’s Secure Housing Unit
(built in 1989) has cells that measure 7.6 feet by 11.6 feet, with some cells
facing concrete walls (Goode, 2015). Arrigo and Bullock (2008) further explain
that secure housing units tend to be small cells (6 feet x 8 feet), with solid steel
doors (p. 624). Pizarro and Narag (2008) synthesize the literature on supermax
conditions, indicating that inmates tend to be held in 7- by 12-foot cells, often
without windows.
Double-celling can also occur in restrictive housing conditions (NYCLU, 2015,
p. 34). Some inmates have a “bunkie” or cellmate in these small spaces, which
can lead to tension and has the potential for violence (p. 34, endnote 115). The
NYCLU report describes cellmates sharing “roughly 100 square feet — about
the size of a parking space — that includes a toilet, open shower stall, writing
platform, and bunk beds” (p. 35). A key issue, according to prisoners in these
confined conditions, is the lack of privacy. This finding points to how the physical
dimensions of restrictive housing cells may well constitute deprivation strains.
Regarding time-in-cell, a common feature of restrictive housing includes the
“physical isolation of individuals in which they are confined in their cells
for around twenty-three hours each day (typically twenty-two to twentyfour hours)” (Smith, 2006, p. 448). Jurisdictions report that most men in
administrative segregation spend 23 hours per weekday in their cells. In roughly

9 4 • National Institute of Justice | NIJ.gov

half of the jurisdictions, prisoners spend 23 hours per weekend day in their cells,
as well (Liman Program/ASCA, 2015, p. 37).
Reporting jurisdictions state that prisoners can stay in administrative segregation
for less than 90 days but that stays can range up to three or more years (Liman
Program/ASCA, 2015, p. 28). In supermax facilities, this means long-term
isolation for indeterminate terms (Haney, 2003); however, legal settlements in
California and New York have called for ending the use of indeterminate stays in
restrictive housing (Canon, 2015; Schwirtz & Winerip, 2015).
Some sources indicate that the cells in restrictive housing units lack windows
and are often illuminated by artificial light 24 hours per day, where “prisoners
have no means of controlling the brightness or dimness of their units” (Arrigio &
Bruce, 2008, p. 625). Conditions can include isolation cells that have no windows,
as at Pelican Bay (Lovett, 2015). Some supermax units have skylights in the
hallways (Shalev, 2009). Most jurisdictions indicate that both men and women in
administrative segregation have access to natural light (Liman Program/ASCA,
2015, p. 39), and two-thirds of jurisdictions for men report that prisoners can
control their in-cell lights (Liman Program/ASCA, 2015, p. 39). Sixty percent of
jurisdictions report that female inmates can control in-cell lighting. However,
these results do not address the quality or duration of inmates’ exposure to light,
which is part of inmates’ sensory experiences — their embodied experience of
restrictive housing.
Other physical conditions of restrictive housing include air conditioning and
noise levels. Among both men and women in administrative segregation, most
jurisdictions report having air conditioning (Liman Program/ASCA, 2015, p. 39),
although whether it is functioning is an open question. Some literature about
noise suggests an “eerie silence” in supermax conditions (Shalev, 2009), while
other reports indicate uncontrollable noise levels, with prisoners banging their
fists against cell doors and yelling. Descriptive reports of prisoner frustration and
yelling are also indicated by Owen (1998) for women in secure housing (p. 163).
Additionally, the NYCLU (2015) report describes the noise conditions in the
recreation pen as catcalls, “guys screaming like crazy people” (2015, p. 39), and
pens filled with men “yelling and screaming about nothing” (p. 39).
Another physical concern is access to showers in restrictive housing. In most
supermax conditions, both male and female inmates leave their cells for showers
approximately three times per week (Pizarro & Narag, 2008), although some
reported less-frequent showering (Liman Program/ASCA, 2015, p. 43).
Physical conditions may also encompass recreation and exercise provisions.
Under supermax conditions, prisoners exercise alone, with no recreational
equipment, in a cage or concrete exercise yard outdoors for one hour per day
(Shalev, 2009). Sometimes, the exercise yard is indoors with an open, barred top.
Prisoners describe the exercise “pen” as being surrounded by concrete walls or
heavy metal grating that obstructs their view of the sky (NYCLU, 2015, p. 39).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 95

The vast majority of jurisdictions report that men’s administrative segregation
facilities have outdoor exercise areas and that 60 percent have indoor exercise
areas (Liman Program/ASCA, 2015, p. 41). For women, 90 percent of the 10
reporting jurisdictions indicate that outdoor exercise facilities are available,
while 50 percent report indoor exercise facilities.
Some jurisdictions report similar access to group exercise activities among men
(37 percent) (Liman Program/ASCA, 2015, p. 42) and women (40 percent).
Roughly 20 percent of jurisdictions reporting on conditions for men in
administrative segregation indicate that inmates can receive more hours of
exercise as they progress through a step-down classification system (Liman
Program/ASCA, 2015, p. 41). For men, most jurisdictions report a median of
five hours of exercise per week. Among reporting jurisdictions for women,
hours allotted for exercise range from three to 10 hours per week, with the
median similar to that estimated for men — about five hours per week. The
Liman Program/ASCA report indicates that group exercise tends to be offered
as part of the final stages of a step-down classification system in administrative
segregation conditions (2015, p. 42). Inmates in administrative segregation
in general can skip their exercise periods. They can also be denied exercise
privileges due to inclement weather, as punishment for rule violations, and if
they are considered a threat to security.
Meals also constitute part of the physical conditions of confinement in supermax
custody. Inmates eat meals in their cells (Pizarro & Narag, 2008), their food often
passed through slots in the door (Browne et al., 2011).
Furthermore, inmates in restrictive housing are generally permitted no physical
contact. The touch of another human being tends to happen only when
correctional officers place handcuffs and other restraints on inmates, such as
shackling them at the waist or placing them in leg irons (Browne et al., 2011).
This finding speaks to the distinctive feature that human contact is generally
lacking for prisoners held in restrictive housing (Smith, 2006). For example,
in supermax conditions, inmates may experience cell extractions (the forceful
removal of a prisoner from a cell) (Pizarro & Narag, 2008). Extractions are
employed when inmates cover the glass windows of their cell doors or refuse to
come out of their cells. Other physical conditions of confinement include fourpoint restraint, which can include strapping an inmate to a bed. When inmates
become violent or refuse to follow orders, guards can place them in special cells
that have no amenities — including no beds or toilets. Food delivery and other
services can also be denied (Pizarro & Narag, 2008, p. 26).
Finally, technology and surveillance are part of the physical conditions that
inmates in restrictive housing face (Shalev, 2009). Supermax units include
surveillance by video cameras pointed at cells and intercom systems in the
recreational yard, for example, to reduce contact between prisoners and staff.

9 6 • National Institute of Justice | NIJ.gov

Descriptive information indicates that at Pelican Bay, not only do doors open and
close electronically but corrections officers speak to inmates through intercoms
(Goode, 2015).

Limitations and Recommendations
The data show that the physical conditions in restrictive housing are consistent
with guiding standards. However, there is little information about what these
conditions mean for prisoners, how conditions are experienced, and how
these conditions affect prisoners’ adaptation to prison life, including well­
being and antisocial behavior. Systematic research is needed on prisoners’
embodied experiences in restrictive housing, which must be generated with
in-depth approaches filling in the qualitative story of what it is like to live in
restrictive housing. For example, the present literature notes that the cells are
small. However, it is not known how a small cell with a view of a concrete
wall, or no windows, or a steel door combine to affect those who live in them.
Prisons help ameliorate the constraints of small, general-population cells
with out-of-cell time, recreation, and other programming (National Research
Council, 2014). These mitigating factors, or coping resources, are generally not
available in restrictive housing, where, for example, out-of-cell time is greatly
limited. A key recommendation for future research is that prisoners who have
experienced restrictive housing should be asked directly and systematically
about their experiences of these physical conditions. Although the descriptive
data on the physical conditions in administrative segregation show similarities
among men and women, little is known about their respective experiences of
these conditions.

Totality Versus Separate Strains
Each of these conditions alone may induce deprivation strain, but considering
the totality of physical deprivations that inmates encounter in restrictive housing
may be especially important in making an overall assessment. The National
Research Council report describes some of the conditions in restrictive housing
under the title, “extreme conditions of confinement” (p. 178). However, how the
totality of these conditions is experienced by inmates and how they may together
constitute extreme conditions is lost when research focuses on each specific
dimension separately. New surveys of inmates are needed to systematically
measure these deprivations to provide an overall index of the deprivation strains
that inmates encounter. This effort would be similar to the research conducted on
“cumulative burden,” the total stress load faced in general community population
samples (Turner, Wheaton, & Lloyd, 1994). Such a research direction will yield
comparative, empirical assessments of how stressful various conditions of
restrictive housing are for prisoners. Comparisons could then be made across
prisoner groups, such as comparing men’s and women’s experiences.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 97

Control of the Body and Coping Resources
Sociological theory is informative in underlining that the body is, ultimately,
the site at which the physical conditions of restrictive housing are experienced
(Foucault, 1977; Wacquant, 2004). As Haney observes, “in most of these units in
the United States prisoners cannot come out of their cells without being cinched
up in elaborate ways — handcuffs, leg irons, restraint chains and the like” (2009,
p. 19). This perspective draws attention to the aspects of restrictive housing
that control prisoners’ bodies through the cell conditions imposed and use of
restraints. Coates’ (2015) recent work on being a black male in America, Between
the World and Me, sensitizes audiences to pay attention to race and the body. His
work, along with reports on demographic trends in restrictive housing (Beck,
2015), suggests that it is especially important to acknowledge that the restrictive
physical conditions of confinement are disproportionately experienced by black
bodies. Inequities are seen not only in the demographic patterns of restrictive
housing but also through inquiry into prisoners’ embodied experience. The lack
of qualitative research on this topic obscures this key point, as well as what it
means for the well-being of racial and ethnic minority groups.
According to GST, coping resources can offset strains and stressors — which
have been described as the pains of imprisonment in restrictive housing (Agnew,
2001; 2006; Pearlin et al., 1981; Sykes, 1958/2007). While touring a men’s prison,
the author of this paper observed a prisoner in restrictive housing shadow boxing
in his cell through a small window in its steel door. He was creatively passing
his time in restrictive housing with extremely limited resources at his disposal.
This boxing example and the throwing of bodily substances (Rhodes, 2004)
demonstrate that prisoners may also use their bodies to act out some degree of
resistance. More information is needed on institutional efforts that try to help
prisoners cope with restrictive housing conditions and how prisoners themselves
cope with these conditions — if at all.

Routines, Sanctions, and Incentives
Regarding the sanctions used in restrictive housing, NYCLU describes a “culture
of deprivation,” where deprivation orders may be imposed for seven days, but
there is no cap on the amount on the total time such orders can span (2015, p. 37).
However, the Liman Program/ASCA’s (2015) systematic overview of the use
of both sanctions and incentives in administrative segregation provides more
specific information and also includes some data about women (Figure 1). The
most common sanctions include limiting inmates’ social telephone calls, time
spent listening to the radio, commissary privileges, social visits, possession of
personal property, and exercise time. Sanctions are also imposed that limit
reading material, group programming, individual out-of-cell programming, incell programming, social correspondence, verbal exchanges between prisoners,
and showers. Taken together, the data most often show that the sanctions used

98 • National Institute of Justice | NIJ.gov

Figure 1. Percentage of Jurisdictions Reporting Activities as Disciplinary
Sanctions Among Men and Women.
Social Phone Calls
Radio
Commissary
Social Visits
Personal Property
Television
Exercise
Reading Material
Group Programming
Individual Out-of-Cell
Programming
In-Cell Programming
Social Correspondence
Verbal Exchanges
Between Prisoners
Showers

■ Men

0

20

40

60

■ Women

80

100

Sources: Data on men (n = 43) from Liman Program/ASCA (2015, p. 50); data on women (n = 10)
made available by Liman Program/ASCA.

are similar across gender. However, data also indicate that some sanctions (e.g.,
regarding social visits) are experienced more often by men than women in
administrative segregation.
More details about how sanctions are implemented are embedded in the
literature, including those that affect recreation and food. In New York state,
when prisoners first arrive in secure housing units, they must wear handcuffs
secured to a waist chain during recreation (NYCLU, 2015, p. 38). Prisoners find
that these restraints severely limit exercise options within the already confined
conditions of a recreational cage. After 30 days, if they do not have a disciplinary
infraction, prisoners may go to the recreation cage without the restraints.
Of course, this privilege can be revoked if the inmate incurs a disciplinary
infraction. Food restrictions also constitute sanctions for disciplinary
infractions in restrictive housing (NYCLU, 2015). In New York, there is a recent
recommendation to eliminate “nutraloaf,” a dry, flavorless, 1,100 calorie product
sometimes given to inmates as a sanction in place of the standard meal (McKinley,
2015; The Economist, 2015). Prisoners also report mealtime “drive-bys,” during
which a correctional officer passes their cells without delivering food. Other

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 99

Figure 2. Percentage of Jurisdictions Reporting Activities as Incentives Among
Men and Women.
Commissary
Personal Property
Social Phone Calls
Social Visits
Television
Exercise
Radio
Reading Material
Group Programming
Individual Out-of-Cell
Programming
In-Cell Programming
Showers
Verbal Exchanges
Between Prisoners
Social Correspondence

■ Men

0

10

20

30

■ Women

40

50

Sources: Data on men (n = 41) from Liman Program/ASCA (2015, p. 50); data on women (n = 10)
made available by Liman Program.

descriptive information from prisoners reveals that a tray may be delivered
with a cover, as is shown on security cameras, but with no food under the cover
(NYCLU, 2015, p. 39).
The Liman Program/ASCA data (2015) also provide an overview of the
incentives that are available in administrative segregation across jurisdictions
(Figure 2). The most common incentives are access to the commissary, personal
property allowed in cell, and social telephone calls and visits. A comparative
assessment by gender on the use of incentives in administrative segregation
shows that fewer programs offer incentives to female inmates than do programs
for men.

Limitations and Recommendations
Some of the descriptive information about sanctions is derived from convenience
samples; more systematic data are available on administrative segregation.
However, more systematic data are still needed pertaining to sanctions and

10 0 • National Institute of Justice | NIJ.gov

incentives used across the various living conditions in restrictive housing and
gender similarities and differences in their use. The quantitative, descriptive
results tend to show similarities in the provision of sanctions and incentives to
women and men. This finding is in keeping with the qualitative data reported by
the responding administrators in jurisdictions involved in the Liman Program/
ASCA report. However, the quantitative data also support other findings that
there are some differences. Specifically, this report finds that more sanctions
are used with men than with women in administrative segregation, and fewer
incentives to illicit good behavior are provided to women. Further comparative
inquiry would increase the knowledge base regarding how gender, as one example
of social location, shapes inmates’ experiences with the conditions of confinement
in restrictive housing. More information on this topic is needed to better
understand and address prisoners’ embodied experiences of restrictive housing.

Social Conditions of Confinement
Social relationships in prison include those with family members, staff, and
other inmates, but restrictive housing precludes most of these social contacts.
For example, Browne and colleagues (2011) reported that family visits are
reduced, or may be completely prohibited, for a year or more. They further
note that when family visits are allowed, they are conducted by speaker or
telephone through a thick glass window, with no opportunity for human touch
(Browne et al., 2011, p. 47). In some facilities, visits even occur over closedcircuit television (Arrigo & Bullock, 2008, p. 625).
The Liman Program/ASCA report and data (2015) indicate that all reporting
jurisdictions permit social visits for both men and women. The jurisdictions also
reported that 20 percent of women’s visits are contact visits, but no jurisdictions
reported contact visits for men. The research indicates that, for prisoners in
restrictive housing units, approximately 5 percent of visits to male prisoners
occur only via video. No facilities reported restricting female inmates to videoonly visits; however, a combination of contact, non-contact, and video are used
for women (10 percent) and men (7 percent). Whether these patterns have
differential meanings to inmates by gender is unknown.
Both male and female inmates in administrative segregation are allowed social
telephone calls (in addition to calls for legal or religious purposes) (Liman
Program/ASCA, 2015). The vast majority of these telephone calls are monitored,
as are calls by general population prisoners. However, it may be important
to consider the influence of such monitoring as part of the totality of other
circumstances and restraints that inmates face in restrictive housing.

Limitations and Recommendations
Research is needed on the social effects of solitary confinement on families in
the community. Prisoners’ perceptions that their families are also affected by
Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 101

their incarceration may add to the pains of imprisonment they experience. The
issue is relevant to practitioners, in that it may be a heretofore unacknowledged
form of strain for prisoners in restrictive housing. Investigating this area would
contribute to research on the collateral consequences of incarceration or
“spillover effects” (Comfort, 2007). Literature pertaining to incarcerated men
indicates that female spouses and partners are deeply affected by their loved ones’
circumstances (Comfort, 2007; Wildeman, Schnittker, & Turney, 2012). There
is a notable lack of information about how a family fares while a family member
is in restrictive housing (Smith, 2006, p. 497). The recent NYCLU report (2015,
p. 28) indicates that families also suffer when prisoners are sent to restrictive
confinement. More systematic research designs would yield further insight.
Drawing on GST (Agnew, 2006), given the severity of conditions in restrictive
housing, it would seem that the effects of having a loved one in these conditions
may be even more acute for families on the outside; however, only comparative
research would shed light on that issue.
Social conditions in restrictive housing also include contacts with other
prisoners and staff. Those in restrictive housing tend to be excluded from
normal prison programming, routines, and collective activities, greatly
reducing social interaction (Haney, 2003). Although prisoners can yell to those
in the next cell or pass notes, and must interact with guards, Haney argues that
these personal exchanges do not constitute normal social interaction. Smith
(2006) goes on to pinpoint harm in solitary confinement: “The central harmful
feature is that it reduces meaningful social contact to an absolute minimum: a
level of social and psychological stimulus that many individuals will experience
as insufficient to remain reasonably healthy and well-functioning” (p. 503).
More information about how prisoners experience the social conditions of
restrictive housing — or the lack of them — would illuminate whether the
conditions of confinement used in supermax facilities today are serving
society’s best interests over the long term.
Furthermore, little information on staff-inmate interactions has been gathered in
a systematic fashion. Given the constraints of restrictive housing, some inmateto-staff communication in restrictive housing occurs through their bodies, in
the form of cutting, suicide attempts, and throwing bodily substances (Rhodes,
2004). Inmate contact with prison staff also occurs in the course of routine
activities in restrictive housing, such as being escorted to the exercise yard or
the toilet, or through brief encounters when meals are delivered to the cell door
(Smith, 2006, p. 448). Ethnographies that expand on staff-inmate interactions are
extremely rare in the literature. However, they are needed given that staff-inmate
interactions constitute part of the social conditions of confinement.

Gender Similarities or Differences?
The Liman Program/ASCA data provide further survey-based information on
gender and the degree of communication among prisoners. All jurisdictions

102 • National Institute of Justice | NIJ.gov

report that male and female prisoners can talk with one another, but the quality
of those contacts is not revealed by the findings. The data show that more men
than women are allowed to talk via group programming, but it is not clear
whether that is due to more restrictions on women or less group programming
for women. The data also indicate that more women than men talk during
recreation-yard activities; therefore, qualitative differences may exist by gender in
how social contacts occur in prison.
Some insight about women in restrictive conditions, and their social contacts, is
embedded in a broader view of Owen’s (1998) work in California. Her research
includes conditions in the prison’s general population and shows that women
form “play families” as part of their social relations. A family member’s stay
in restrictive housing conditions affects her prison family. Owen’s descriptive
research suggests that women find it painful to be cut off from play-family
relationships in the prison, wondering, for example, if partners on the inside
will wait for them while they are in restrictive housing (Owen, 1998, pp.134­
137). This research then suggests that perceived impact on within-prison ties is a
pain of imprisonment for women in restrictive housing. These play- or pseudofamily ties can also lead to time in secure housing (e.g., as punishment for
violence against a partner). This insight raises new questions about how women
experience restrictive housing when it separates them from their in-prison
partner or other in-prison family members. However, more updated research is
needed on both men’s and women’s social relationships in prison.
Although Owen’s (1998) qualitative research on women includes only a few
cases of those in restrictive housing, it also raises some questions about staffprisoner relationships. As part of her overall study, Owen interviewed some
staff and prisoners in restrictive housing. One striking finding comes from the
comments of a prison guard about the degree of perceived physical threat from
the prisoners. The guard reports that “[t]here is no danger for staff in here, not
even in the Seg unit. The danger in here is getting manipulated” (Owen, 1998,
p. 165). This perception contrasts with the extremes of supermax conditions
that focus on constraining “dangerous men” (Simon, 2014). However, in a visit
to a restrictive housing component in a women’s prison, the author found that
staff were highly concerned about perceived threats to their safety and that of
visitors. Visitors were given protective vests to wear and instructed to not walk
too close to the cells (which had steel bars) to prevent being spit upon. The
message conveyed throughout the visit was that these were “dangerous women.”
This message conflicts with the prison guard’s perception in Owen’s research.
Therefore, more research is needed to clarify how both women and men are
perceived as threats to prison functioning and the reasons they are placed in
restrictive housing.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 103

Psychological Conditions
Even with variations in the conditions of confinement, Smith states that “the
overall conclusion must be that solitary confinement — regardless of specific
conditions and regardless of time and place — causes serious health problems
for a number of inmates” (2006, p. 503). It is estimated that about 45 percent
of supermax prisoners suffer from psychosocial impairments (Lovell, 2008).
Another study shows that two-thirds or more of those in supermax conditions
suffer from psychological and emotional trauma, and the psychopathological
effects of isolation (Haney, 2003). Furthermore, a recent review of extant
literature by the National Research Council (2014) states that “the overwhelming
majority of studies document the painful and potentially damaging nature of
long-term prison isolation. Occasional studies have found little or no harm …
[h]owever numerous methodological concerns have been expressed that limit
any straightforward interpretation of these counterintuitive results” (pp.186-187).
Haney and Lynch’s review from 1997 discusses some studies that found no
adverse psychological effects of restricted housing and suggest that this is a
problematic conclusion of some early literature. Smith (2006) reviews this
research as well, mentioning, in particular, some studies on sensory deprivation
that found no negative effects. His review notes that this research was based
on a voluntary sample of college students in an experimental setting for 10
days or less (Kurki and Morris, 2001, p. 431), which is not comparable to the
involuntary conditions of confinement. However, a more recent longitudinal
study in Colorado examined 270 male inmates with and without mental illness
in administrative segregation and in the general prison population. They found
no adverse mental health effects from exposure to administrative segregation
during a one-year period (O’Keefe, Klebe, Metzner, Dvoskin, Fellner, & Stucker,
2013). As the authors themselves note, more research is needed on these findings
because they are from one state and conditions of confinement vary across
jurisdictions. In addition, mental health effects may emerge after the one-year
interval in which they were studied.

Limitations and Recommendations
On the whole, studies tend to show adverse psychological conditions for inmates
living in restrictive housing (Mears, 2008, p. 691), although evidence is mixed.
More research is needed to clarify the psychological conditions that may arise
in restrictive housing and to evaluate how variations in these conditions of
confinement (e.g., level and types of deprivation strains) are associated with
mental health outcomes. This research would be further enhanced by considering
the potentially offsetting role of coping resources, when available. The GST
framework provides a rationale for future research on processes that may explain
how strains and coping resources in these circumstances work to affect the
mental health of inmates in restrictive housing. This type of empirical research
may suggest policy-relevant points of prevention and intervention.

10 4 • National Institute of Justice | NIJ.gov

Service Provision in Restrictive Housing
Some information on the services provided to prisoners in restrictive housing is
embedded in the recent Liman Program/ASCA report (2015) and in studies of
supermax prisons (Shalev, 2009). Concerns about service provision in restrictive
housing have emerged from the court case in California regarding the quality
of health care and therapeutic programming in Madrid v. Gomez at the Pelican
Bay Secure Housing Unit (Arrigo & Bullock, 2008, p. 625). The consensus in
the limited extant literature is that mental health and medical services are very
limited for those in restrictive housing (Browne et al., 2011).
Supermax facilities such as Pelican Bay often can provide more in-depth
information about service provision in restrictive housing. Programming for
inmates in supermax conditions is provided within cells due to safety concerns
(Pizarro & Narag, 2008). Educational services may be provided, as they are at
Pelican Bay, by placing inmates in a row of educational cells that face computer
monitors, through which the instructors teach the lesson (Shalev, 2009). Other
information on supermax conditions suggests that when education is provided,
it is by teachers talking to inmates through openings in the cell doors (Pizarro
and Narag, 2008). In some cases, prisoners may sign up for in-cell study packets
(e.g., for GED, substance use, aggression management) (NYCLU, 2015, p. 32).
However, extant literature generally points to an absence of programming.
The literature pertaining to services also indicates that mental health and
medical services are extremely limited for prisoners in restrictive housing
conditions (Browne et al., 2011), where visits with mental health counselors
and staff are conducted through the cell door. However, cell-front therapy
allows other inmates to hear the prisoner-therapist discussion. If out-of-cell
treatment is provided, the inmate is led to treatment in shackles and remains
shackled throughout the session. Shalev (2009) provides some information on
other medical services in supermax conditions. For example, the Pelican Bay
Secure Housing Unit has on-site medical and dental clinics, but more systematic
information on medical services across jurisdictions is needed. Telemedicine is
also used in supermax facilities, with medical examinations performed via video
conferencing links between the prison and medical centers. One 10-minute
telemedicine conference requires roughly two hours of preparation, which
involves a strip search of the prisoner, an escort to and from the telemedicine
clinic, and two guards who remain with the prisoner during the examination
(Shalev, 2009).
Prisoners in secure housing units in New York state explain that medical
staff come to a cell when the inmate submits a sick-call slip; the inmate talks
with staff through a locked cell door and the food slots, affording no privacy
(NYCLU, 2015, p. 40). Visits with a psychiatrist are sometimes conducted
via teleconferencing (p. 41). Over all, prisoners report difficulty in receiving
attention from medical personnel and social workers; when they do receive
medical services, issues of privacy remain a concern.
Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 105

Finally, the literature shows an overall dearth of programming for prisoners in
supermax conditions, drawing attention to the idleness often observed among
these inmates (NYCLU, 2015; Haney, 2009) — who live a deeply monotonous
existence with pronounced deprivations (Haney, 2009). Restrictions on
programming and services also likely amplify the overall pains of imprisonment
experienced by prisoners.

Limitations and Recommendations
Information in extant literature on the availability of legal, medical, educational,
and mental health services in restrictive housing is sparse. However, service
provision is potentially very pertinent in terms of the GST framework, as services
represent a programmatically modifiable form of coping resources for prisoners.
More information is needed, as well, about the effectiveness of the services
provided, including, for example, different modes for providing medical services
in these circumstances.

Gender Similarities or Differences?
Data on women made available by the Liman Program and the Liman Program/
ASCA report (2015) were synthesized to provide insight into gender similarities
and differences in service provision for those in administrative segregation. This
information indicates that in-cell programming is the most common for both
men and women, whereas out-of-cell programming is less common. In-cell
programming includes mental health care, GED, and education services. Outof-cell programming includes mental health groups, education, and visits with
a counselor. Although important in understanding the parameters of service
provision in restrictive housing, this information is limited — the data do not
speak to the quality of the programming or how it is experienced by prisoners.
Information pertaining to legal visits is available for both men and women in
administrative segregation. All jurisdictions permit legal visits for both, but
these meetings are a mix of contact and non-contact visits that are monitored.
Therefore, the information about legal visits by gender shows similarities, but
more information is needed about the quality of these visits and how they are
experienced by inmates. In addition, the monitoring of legal visits should be
viewed within the totality of constraints that these prisoners face. Is it yet another
form of deprivation strain that prisoners must endure?

Systematic Research on Step-Down Programs
Step-down programs allow inmates to alter the extreme conditions of
confinement in restrictive housing and regain certain privileges or coping
resources by reaching specific milestones. By establishing step-down programs,
states “tie an inmate’s departure from segregation to the completion of certain
goals, such as behavioral plans and classes” (Liman Program, 2013, p. 18). Some

10 6 • National Institute of Justice | NIJ.gov

states — including Connecticut, Massachusetts, Mississippi, New Jersey, New
York, and Virginia — have structured programs that target behavior in some
way. Of the states listed above, Massachusetts, Mississippi, and Virginia — like
Washington and Colorado — are working to find ways to hold prisoners in
restrictive housing while allowing for more opportunities for group activities
and therapy in those circumstances (Liman Program, 2013, p. 18). Without
such measures, inmates serving the remainder of their sentences in restrictive
housing cannot access the transitional programs for community re-entry that are
available to the general prison population (e.g., assistance with acquiring letters
of recommendation and in developing post-release plans, a resume, and cover
letter) (NYCLU, 2015, p. 33). In fact, one survey found that 4,400 prisoners in
2013 were released directly from secure housing to the community (p. 29). In
some of the jurisdictions listed above, transition release programs are part of the
step-down programs offered in restrictive housing (Liman Program/ASCA, 2015,
p. 30). Information on the effects and workings of step-down programs is not
widely available; however, these programs should be the basis of research inquiry
because they may reduce the deprivation strains that prisoners encounter and
help prepare prisoners for re-entry.
Finally, changes in the confinement conditions in some restrictive housing
circumstances are either in place or are being recommended. In California, for
example, prisoners will no longer be held in isolation indefinitely (Ford, 2015;
Lovett, 2015; Schwirtz & Winerip, 2015). The implementation of these changes
should be monitored and studied for the benefit of practitioners and prisoners
throughout the nation.

Considering Conditions for Subgroups
Much of what is known about the conditions of confinement in restrictive
housing practices pertains to adult male prisoners. To further consider
diversity in prisoner experiences of the conditions of confinement, this
report features women as a group of inmates that tend to be overlooked
when considering restrictive housing. To move toward an even deeper
understanding of embodied experience, other overlooked groups must also be
included in research and policy considerations. In fact, information is needed
on a range of subgroups, including racial and ethnic minorities, transgender
inmates (Sylvia Rivera Law Project, 2007), juveniles (ACLU, 2012; Birckhead,
2015), and mentally ill prisoners (American Civil Liberties Union Colorado
(ACLUC), 2013; Rodriguez, 2013).
Research shows that transgender women in men’s prisons are 13 times more
likely than non-transgender inmates to be sexually assaulted by other inmates
(Jenness, Maxon, Sumner, & Matsuda, 2010; Jenness & Fenstermaker, 2014).
Given high rates of sexual victimization, New York’s Rikers Island Jail has created
a special housing unit for transgender inmates (Mathias, 2014). Other responses

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 107

to the elevated threat of sexual victimization among transgender inmates involve
placing those inmates in protective custody (Sylvia Rivera Law Project, 2007).
Therefore, it is important to consider the degree to which transgender inmates
are differentially exposed to these conditions compared to non-transgender
(cisgender) inmates. The experiences of transgender women in restrictive housing
also need further research (Sylvia Rivera Law Project, 2007). For example, as
reported by a transgender female prisoner in a larger study in California, the
likelihood of being moved to restrictive housing serves as a deterrent to reporting
sexual assault (Jenness, personal communication, January 12, 2016).

Discussion and Conclusions
This report brings together information on the conditions of confinement in
restrictive housing. Although some details pertaining to physical conditions,
sanctions and incentives, social conditions, psychological conditions, and
service provision are synthesized in this report, there is an overwhelming lack of
systematic information on the topic. This lack of information stems from a gap in
the research — prisoners’ personal experiences of the conditions of confinement.
The Liman Program/ASCA data provide some of the most informative, nationally
representative data on these conditions across jurisdictions in the United States
and are used extensively in this report. However, many issues raised by the current
synthesis and critique point to areas that require further empirical inquiry.
The majority of the information available describes the basic physical conditions
in some forms of restrictive housing, but systematic information about restrictive
housing is lacking. The most comprehensive information available tends to
come from descriptions of supermax facilities, but very little is known about the
social conditions and services provided in restrictive housing. Even when these
elements are present in the literature, their coverage is sometimes inconsistent
(e.g., total isolation or whether family contact is permitted), which may reflect the
myriad conditions encompassed by the term “restrictive housing.” As is shown in
this report, there are also some differences in conditions labeled “administrative
segregation” for men and women (e.g., regarding sanctions and incentives),
although there are also similarities. Furthermore, details are also lacking about
the effects of restrictive housing conditions over an extended period of time.
Nationally representative, descriptive information is available on the demographic
and criminal justice history patterns of groups of inmates who experience
restrictive housing (Beck, 2015). Most important, however, is that very little is
known about their experiences in these conditions of confinement. Research
also must move beyond trace mentions of the conditions of restrictive housing
embedded in other studies and further address the nuances and workings of these
conditions. Such efforts may be of considerable benefit to practitioners; they may
help them to better monitor prisoners’ behavior and provide effective conditions
when and if restrictive housing is absolutely necessary.

108 • National Institute of Justice | NIJ.gov

Quantitative Story of Restrictive Housing
The quantitative story of restrictive housing addresses the trends and patterns
surrounding who experiences it and descriptive information on the prevalence
of particular conditions of confinement. Given recently available information
on the criminal justice histories of those in restrictive housing (Beck, 2015), it is
clear that prisoners often bring some importation strains (pre-prison stressors)
with them to restrictive housing. However, a review of evidence covered in this
report suggests that inmates endure considerable deprivation strains (stressors
experienced while living in these conditions). More research on these deprivation
strains is needed to better flesh out the quantitative story of restrictive housing.
General strain theory (Agnew, 2001; 2006) and the pains of imprisonment
concept (Sykes, 1958/2007) offer theoretical tools to better systematically
detail the deprivation strains involved in restrictive housing. This framework
points not only to detailing the types and levels of strains that prisoners face
in restrictive housing but also leads to a consideration of the totality of strains
that they endure in these conditions (Turner, Wheaton, & Lloyd, 1994).
Developing the quantitative story of restrictive housing would entail creating
questionnaires requiring inmate feedback, which would systematically measure
strains in prisons and different conditions within them. Furthermore, using
a comprehensive theoretical framework will help to move research from
description to empirical analyses of how these conditions affect prisoner
outcomes, including health. Inquiry along these lines may also offer theoretically
guided and empirically based recommendations for ways to reduce the
deprivation strains that prisoners endure in restrictive housing. General strain
theory may be further used to consider instituting programming to modify the
coping resources available to incarcerated individuals living in these conditions.
It predicts that mitigating those strains should reduce antisocial behavior in
keeping with rehabilitative goals, and in considering the future re-entry of these
inmates to the general population of prisoners and, eventually, to society. The
development of coping resources, according to GST, should also foster order
within restrictive housing conditions in keeping with the managerial goals of
correctional institutions. Investment in coping resources, therefore, may be
particularly justified when restrictive housing conditions are imposed. More
information is particularly needed on programs that are effective in reducing
deprivation strains and on the potential of step-down programs.
Furthermore, a number of questions remain about the conditions of confinement
considered to be deprivation strains. For example, if deprivation strain is
experienced in restrictive housing, as suggested by the conditions described, how
does it spill over to inmates’ families? How does that affect prisoners? What are
the long-term effects of exposure to restrictive housing conditions for inmates
and their families? For practitioners, further inquiry along these lines may lead
to the development of effective policies and programs when and if restricted
housing is needed.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 109

Qualitative Story on Restrictive Housing
The qualitative story of restrictive housing points to the body as a site of
deprivation. In line with critical sociological theories (Foucault, 1979; Wacquant,
2004), a focus on embodied experience may illuminate why different social
locations may matter in restrictive housing conditions. As an example of moving
toward embodied experience, this report has touched on gender similarities
and differences in restrictive housing. Most of the research (although sparse)
has been conducted with men; even less information is available about female
inmates living under these circumstances. Although the data show that men
and women face fairly similar conditions in solitary confinement (e.g., types
of physical conditions and a lack of services), there are differences that must
be understood. For example, fewer incentives are provided to women in
administrative segregation, and more sanctions are provided to men in these
circumstances. Such differences may have implications for inmates’ adjustment
to living in restrictive housing. This limited inquiry into gender and restrictive
housing suggests a need for more systematic research on embodied experience.
Further development of this concept will yield insight into how different
subgroups experience strains in restrictive housing and may inform better policy
and practice.

Conclusions
This report begins to synthesize what is known about the conditions of
confinement in restrictive housing. It offers some initial insights based on extant
research. Due to a lack of information, both the quantitative and qualitative stories
of restrictive housing presented here are limited — and are in dire need of more
research. Taken as a whole, the conditions of confinement present for inmates
living in restrictive housing are extremely challenging across all the dimensions
considered (physical, sanctions and incentives, social, psychological, and service
provision). These conditions suggest that the use of restrictive housing needs to be
avoided as much as possible. Fleshing out these qualitative and quantitative stories
will lead to ways to make the conditions of confinement as humane as possible, and
to use restrictive housing only when absolutely unavoidable.
Further empirical information is needed to systematically understand how
the conditions and processes of restrictive housing affect the inmates who live
there. One pressing area of research involves the systematic investigation of the
conditions that constitute the pains of imprisonment; their types and levels;
and how they work together, separately, and as a totality in affecting prisoner
experiences. The GST and pains of imprisonment frameworks may be used to
illuminate the types of strains that inmates endure in restrictive housing and
the role of coping resources in these conditions. Research is urgently needed

110 • National Institute of Justice | NIJ.gov

on services, programs, and opportunities for social interaction (e.g., visits from
inmates’ family members). Such areas may be clear points at which programming
can be modified to be more responsive to prisoner needs. Another area that is
amenable to change through policy may informed by more empirical research
on the promise of step-down programs for reintegrating those held in restrictive
housing into both the general population of prisoners and society at large.
Research with implications for programming will be useful to practitioners in
shaping safer and more humane practices as they conduct their vital work.
Finally, future research may consider the promise of mixed research methods,
which involve using both qualitative and quantitative approaches to understand
a social problem (Creswell & Plano-Clark, 2011). A systematic program of
research may begin with qualitative studies that may fill in missing information
on the meaning of myriad conditions that inmates face in restrictive housing and
yield new insight into inmates’ embodied experience in these circumstances. As
a next step, a program of research could build on the qualitative information by
generating systematic questionnaires to quantify the strains, coping resources,
and psychological outcomes among a representative sample of inmates living
in restrictive housing and in the general prison population. A mixed-methods
program of research on restrictive housing would then yield comparative
assessments of the conditions of confinement. By investigating the processes
involved, this program of research would also inform practitioners about
promising areas of intervention and prevention — grounded in experience and
theory — for problems that arise among inmates when restrictive housing cannot
be avoided.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 111

References
Agnew, R. (2001). Building on the foundation of general strain theory: Specifying
the types of strain most likely to lead to crime and delinquency. Journal of
Research in Crime and Delinquency, 38, 319-361.
Agnew, R. (2006). Pressured into crime: An overview of General Strain Theory.
Los Angeles, CA: Roxbury.
Agnew, R., & Delisi, M. (2012). General Strain Theory, the criminal justice system
and beyond: Introduction to the special issue. Crime and Justice, 40, 174-175.
American Civil Liberties Union (2012). Growing up locked down: Youth in solitary
confinement in jails and prisons across the United States. Human Rights Watch.
http://www.aclu.org.
American Civil Liberties Union Colorado (2013). Out of sight, out of mind:
Colorado’s continued warehousing of mentally ill prisoners in solitary confinement.
http://aclu-co.org/wp-content/uploads/files/imce/ACLU-CO%20Report%20
on%20Solitary%20Confinement_2.pdf.
Arrigo, B. A., & Bullock, J. L. (2008). The psychological effects of solitary
confinement on prisoners in supermax units. International Journal of Offender
Therapy and Comparative Criminology, 52, 622-640.
Beck, A. J. (2015). Use of restrictive housing in U.S. prisons and jails, 2011-12.
NCJ 249209. Washington, DC: Bureau of Justice Statistics.
Bernard, T. J., Snipes, J. B., & Gerould, A. G. (2009). Vold’s theoretical criminology:
6th edition. New York, NY: Oxford University Press.
Birckhead, T. R. (2015). Children in isolation: The solitary confinement of youth.
Wake Forest Law Review, 50(1). Forthcoming: UNC Legal Studies Research Paper
No. 2512867. http://ssrn.com/abstract=2512867.
Blevins, K. R., Listwan, S. J., Cullen, F. T., & Johnson, C. L. (2010). A General
Strain Theory of prison violence and misconduct: An integrated model of inmate
behavior. Journal of Contemporary Criminal Justice, 26, 148-166.
Browne, A., Cambier, A., & Agha, S. (2011). Prisons within prisons: The use of
segregation in the United States. Federal Sentencing Reporter, 24, 46-49.
Canon, G. (2015). Inside the landmark court case that will end indefinite solitary
confinement in California. http://www.motherjones.com/politics/2015/09/
california-limits-solitary-confinement.
Clear, T. A., & Frost, N. L. (2013). The punishment imperative: The rise and failure
of mass incarceration in America. New York, NY: New York University Press.

112 • National Institute of Justice | NIJ.gov

Clemmer, D. (1940/1965). The prison community. New York, NY: Holt, Rinehart
and Winston.
Coates, T. N. (2015). Between the world and me. New York, NY: Spiegel & Grau.
Comfort, M. (2007). Punishment beyond the legal offender. Annual Review of
Law and Social Science, 3, 271-296.
Covington, S. S., & Bloom, B. E. (2006). Gender-responsive treatment and
services in correctional settings. Women and Therapy, 29, 9-33.
Creswell, J. W., & Plano-Clark, V. L. (2011). Designing and conducting mixed
methods research. 2nd edition. Thousand Oaks, CA: Sage.
Delisi, M. (2011). How general is general strain theory? Journal of Criminal
Justice 39, 1-2.
Fleury-Steiner, B. & Longazel, J. (2014). The pains of mass imprisonment.
New York: Sage.
Ford, M. (2015). The beginning of the end for solitary confinement? The Atlantic.
http://www.theatlantic.com/politics/archive/2015/09/scaling-back-solitary/403441/.
Foster, H. (2012). The strains of maternal imprisonment: Importation and
deprivation stressors for women and children. Journal of Criminal Justice, 40,
221-229.
Foucault, M. (1977). Discipline & punish. New York, NY: Vintage Books.
Garland, D. (2001). Introduction: The meaning of mass imprisonment. In D.
Garland (ed.), Mass imprisonment: Social causes and consequences (pp. 1-3).
Thousand Oaks, CA: Sage.
Goode, E. (2015, August 4). Solitary confinement: Punished for life. New York
Times. http://www.nytimes.com/2015/08/04/health/solitary-confinement­
mental-illness.html?_r=2.
Goodstein, L., & Wright, K. N. (1989). Inmate adjustment to prison. In L.
Goodstein and D. Layton MacKenzie (eds.), The american prison: Issues in research
and policy (pp. 229-251). New York, NY: Plenum Press.
Giallombardo, R. (1966). Society of women: A study of a women’s prison.
New York, NY: John Wiley & Sons.
Haney, C., & Lynch, M. (1997). Regulating prisons of the future: A psychological
analysis of supermax and solitary confinement. New York University School of
Law and Social Change, 23, 477-570.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 113

Haney, C. (2003). Mental health issues in long-term solitary and “supermax”
confinement. Crime & Delinquency, 49, 124-156.
Haney, C. (2009). The social psychology of isolation: Why solitary confinement is
psychologically harmful. Prison Service Journal, 181, 12-20.
Heffernan, E. (1972). Making it in prison: The square, the cool, and the life. New York,
NY: Wiley-Interscience.
Jenness, V., & Fenstermaker, S. (2014). Agnes goes to prison: Gender authenticity,
transgender inmates in prisons for men, and pursuit of “The Real Deal.” Gender
& Society, 28, 5-31.
Jenness, V., Maxon, C. L., Sumner, J. M., & Matsuda, K. N. (2010). Accomplishing
the difficult but not the impossible: Collecting self-report data on inmate-on­
inmate sexual assault in prison. Criminal Justice Policy Review, 21, 3-30.
Kurki, L. & Morris, N. (2001). The purposes, practices, and problems of supermax
prisons. Crime and Justice, 28. 385-424. Chicago: University of Chicago Press.
Liman Program, Yale Law School & the Association of State Correctional
Administrators. (2015). Time-in-cell: The ASCA-Liman 2014 National Survey of
Administrative Segregation in Prison. New Haven, CT: Yale University. https://
www.asca.net/system/assets/attachments/9195/*asca-liman_administrative_
segregation_report_sep_2_2015.pdf?1450213204 or http://www.asca.net.
Liman Program, Yale Law School. (2013). Administrative segregation, degrees of
isolation, and incarceration: A national overview of state and federal correctional
policies. New Haven, CT: Yale University. https://www.law.yale.edu/system/files/
area/center/liman/document/Liman_overview_segregation_June_25_2013_TO_
POST_FINAL%281%29.pdf.
Lovell, D. (2008). Patterns of disturbed behavior in a supermax population.
Criminal Justice and Behavior, 35, 985-1004.
Lovett, I. (2015). California agrees to overhaul use of solitary confinement.
The New York Times. http://www.nytimes.com/2015/09/02/us/solitary­
confinement-california-prisons.html.
McKinley, J. (2015). New York prisons take unsavory punishment off the table.
The New York Times. http://www.nytimes.com/2015/12/18/nyregion/new-york­
prisons-take-an-unsavory-punishment-off-the-table.html.
Mathias, C. (2014, November 18). New York’s largest jail to open housing
unit for transgender women. Huffington Post. http://www.huffingtonpost.
com/2014/11/18/rikers-transgender-women_n_6181552.html.
Mears, D. P. (2005). Evaluating the effectiveness of supermax prisons. Washington, DC:
The Urban Institute Justice Policy Center.
114 • National Institute of Justice | NIJ.gov

Mears, D. P. (2008). An assessment of supermax prisons using an evaluation
research framework. The Prison Journal, 88, 43-68.
Mears, D. P. (2013). Supermax prisons: The policy and the evidence.
Criminology & Public Policy, 12, 681-719.
National Research Council. (2014). The growth of incarceration in the United
States: Exploring the causes and consequences. Committee on Causes and
Consequences of High Rates of Incarceration. J. Travis, B. Western, and S.
Redburn (eds.). Committee on Law and Justice, Division of Behavioral and
Social Sciences and Education. Washington, DC: The National Academies Press.
New York Civil Liberties Union. (2015). Boxed in: The true cost of extreme
isolation in New York’s prisons. New York, NY: New York Civil Liberties Union.
http://www.nyclu.org.
O’Keefe, M. L., Klebe, K. J., Metzner, J., Dvoskin, J., Fellner, J., & Stucker, A.
(2013). A longitudinal study of administrative segregation. Journal of the
American Academy of Psychiatry and Law, 41, 49-60.
Owen, B. (1998). In the mix: Struggle and survival in a women’s prison.
Albany, NY: State University of New York Press.
Pearlin, L., Menaghan, E., Leiberman, M. A., & Mullan, J. T. (1981). The stress process.
Journal of Health and Social Behavior, 22, 337-356.
Pizarro, J. M., & Narag, R. E. (2008). Supermax prisons: What we know, what we
do not know, and where we are going. The Prison Journal, 88, 23-42.
Rhodes, L. A. (2004). Total confinement: Madness and reason in the maximum
security prison. Berkeley, CA: University of California Press.
Rodriguez, S. (2013, April 13). Lawsuit filed against solitary confinement of 800
“seriously mentally ill” prisoners in Pennsylvania. Prison Legal News. https://
www.prisonlegalnews.org/news/2013/apr/15/lawsuit-filed-against-solitary­
confinement-of-800-seriously-mentally-ill-prisoners-in-pennyslvania/.
Shalev, S. (2009). Supermax: Controlling risk through solitary confinement.
Portland, OR: Willan.
Schwirtz, M., & Winerip, M. (2015, December 16). New York State agrees to
overhaul solitary confinement in prisons. The New York Times. http://www.
nytimes.com/2015/12/17/nyregion/new-york-state-agrees-to-overhaul-solitary­
confinement-in-prisons.html?_r=0.
Skarbek, D. (2014). The social order of the underworld: How prison gangs govern
the American penal system. New York, NY: Oxford University Press.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 115

Simon, J. (2014). Mass incarceration on trial: A remarkable court decision and the
future of prisons in America. New York, NY: The New Press.
Smith, P. S. (2006). The effects of solitary confinement on prison inmates: A brief
history and review of the literature. Crime and Justice, 34, 441-528.
Sykes, G. (1958/2007). The society of captives: A study of a maximum security
prison. Princeton, NJ: Princeton University Press.
Sylvia Rivera Law Project. (2007). “It’s a war in here”: A report on the treatment
of transgender and intersex people in New York state men’s prisons. New York, NY:
Sylvia Rivera Law Project. http://www.srlp.org.
The Economist. (2015, December 21). Solitary confinement is cruel — but
soon it will at least be more unusual. http://www.economist.com/blogs/
democracyinamerica/2015/12/american-prisons.
Turner, R. J., Wheaton, B., and Lloyd, D. A. (1995). The epidemiology of social
stress. American Sociological Review, 60, 104-125.
U.S. Department of Justice (2016). Report and recommendations concerning the
use of restrictive housing. Final Report. Washington, DC: U.S. Department of
Justice. http://www.justice.gov/restrictivehousing.
Wacquant, L. (2004). Body & soul: Notebooks of an apprentice boxer. New York,
NY: Oxford University Press.
Western, B. (2007). Introduction to the Princeton classic edition. The society
of captives: A study of a maximum security prison. Princeton, NJ: Princeton
University Press.
Wildeman, C., Schnitkker, J., and Turney, K. (2012). Despair by association?
The mental health of mothers with children by recently incarcerated fathers.
American Sociological Review, 77, 216-243.

116 • National Institute of Justice | NIJ.gov

CHAPTER 4

Gang Affiliation and Restrictive
Housing in U.S. Prisons
David C. Pyrooz
University of Colorado Boulder

Introduction

P

rison gangs came to the forefront of issues faced by correctional
authorities toward the end of the 20th century. Gang activity has been
documented in U.S. prisons as early as the 1940s but was not linked in
a major way to prison violence (Camp & Camp, 1985; Crouch & Marquart,
1989; DiIulio, 1990; Irwin, 1980). Emerging in the 1970s was a gang dynamic
responsible for producing an unprecedented amount of disorder and violence
in U.S. prisons. Between 1975 and 1979, there were 124 gang-related homicides
in California prisons (Crouch & Marquart, 1989, p. 204; Porter, 1982). Texas
prisons witnessed 52 homicides during the “war years” of 1984 and 1985, with 90
percent being gang-related (Ralph & Marquart, 1991). Gangs are also implicated
in orchestrating prison riots (Goldstone & Useem, 1999; Useem & Reisig,
1999), such as the 1993 Lucasville riot in Ohio that left nine inmates and one
correctional officer dead (Huff & Meyer, 1997). Although homicide and violence
in contemporary prisons are at historically low levels (Mumola, 2005; Useem
& Piehl, 2008), gangs remain disproportionate contributors to violence and
misconduct in prisons (e.g., Griffin & Hepburn, 2006; Huebner, 2003; Shelden,
1991). They are one of the more challenging issues correctional officials face in
managing prisons.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 117

One response to combat the influence of gangs in U.S. prisons involves moving
affiliates of gangs out of general population housing and into restrictive housing
facilities or units. National surveys of correctional officials indicate that this
response is not only practiced at a high rate (Knox, 2012) but also viewed as the
most effective solution for countering the role of gangs in facilitating misconduct
and prison disorder (Winterdyk & Ruddell, 2010). Indeed, the use of restrictive
housing has been described as a “silver bullet” for addressing gang activity in
prisons (Vigil, 2006, p. 33). From this standpoint, segregating the affiliates of
gangs from the general population could reduce prison violence and disorder
systemwide as well as reducing misconduct among individual gang members
(Fischer, 2002; Ralph & Marquart, 1991).
Critics challenge the use of restrictive housing on gang affiliates on legal
(Tachiki, 1995), humanitarian (Toch, 2007), and empirical (Mears & Reisig,
2006) grounds. One of the most vexing issues involves the wholesale placement
of the members or associates of entire gangs into restrictive housing for
indeterminate periods. Gang affiliates can be subjected to restrictive housing
conditions based not on their behavior, but on their status, which runs contrary
to traditional notions of restrictive housing (King, 1999; Kurki & Morris, 2001).
It is not uncommon to learn of gang affiliates who spend more than a decade
in isolated conditions that restrict them from the basic privileges provided to
the general prison population (Reiter, 2012). The longer gang affiliates stay in
restrictive housing, the more likely they are to misbehave when they return to
general prison population (Labrecque, 2015a). Moreover, a recent review of
the literature regarding restrictive housing concluded that it may have some
potentially devastating psychological consequences (Frost & Monteiro, 2016).
This raises serious questions about its use on gang populations, who already
maintain elevated levels of depression and suicidal tendencies (Watkins & Melde,
2016). The conditions associated with this practice have led to unrest in states
like California, where an uprising led to a hunger strike involving 30,000 inmates
(Rodriguez, 2013) along with a class action lawsuit — Ashker v. Governor of
California. The lawsuit was settled recently, leading to sweeping changes in how
restrictive housing is used on those who affiliate with gangs (St. John, 2015). In
light of these events, any short-term benefits to prison management may not be
worth the long-term consequences (Griffin, 2007).
This white paper examines key issues related to gang affiliation and restrictive
housing in U.S. prisons. It was commissioned by the National Institute of Justice
of the U.S Department of Justice’s Office of Justice Programs to synthesize
what is known about these topics and to identify knowledge gaps that should
be addressed in future research. It begins by addressing the wide range of
terminology and definitions applied to restrictive housing, gangs, and gang
members. Here, the focus is on how these topics are measured and how they are
used in correctional practice. Next, a brief overview of the extent, magnitude,
and nature of gangs, gang affiliation, and restrictive housing is provided to

118 • National Institute of Justice | NIJ.gov

document their emergence and trends as central issues in correctional policy and
practice. The core of this white paper examines how gang affiliation is related to
restrictive housing by focusing on five areas:
1. The logic behind segregating gangs and gang affiliates.
2. Gang affiliation as a correlate of restrictive housing.
3. Pathways into restrictive housing among gang affiliates.
4. Pathways out of restrictive housing among gang affiliates.
5. Gang affiliation, restrictive housing, and the reduction of misconduct and disorder.
This white paper makes several conclusions based on a synthesis of the literature
in these focus areas. It also identifies aspects of these areas that need further
attention from the practitioner, policymaking, and research communities. It
offers recommendations for future research to advance the understanding of the
relationship between gang affiliation and restrictive housing. It should be noted
that much of what is known about the focus areas synthesized in this white paper
is derived from a small number of states, particularly California and Texas, which
might not represent common policies or practices in corrections generally.

Terminology and Definitions: Restrictive Housing,
Gangs, and Gang Affiliation
Restrictive Housing
One of the challenges in surveying the current state of the evidence on the
relationship between gang affiliation and restrictive housing is terminology. Fred
Patrick, Director of the Vera Institute’s Center on Sentencing and Corrections,
noted that the practice of placing inmates in restrictive housing goes by many
names, including “isolation, restricted housing, administrative segregation,
protective custody, special housing, disciplinary segregation, etc.” (Shames,
Wilcox, & Subramanian, 2015, p. 2). A 2010 review of 42 state correctional
policies revealed more than 20 variations in the terminology used for long­
term segregated housing alone (Butler, Griffin, & Johnson, 2013); however,
a consensus on terminology is unfolding. The executive committee of the
American Correctional Association (ACA) (2014) recently passed a resolution to
adopt “restrictive housing” as the terminology used to apply to the broad practice
of separating inmates from the general prison population.
Additional factors are essential to characterizing restrictive housing, as outlined
in the joint report of Yale University’s Liman Program and the Association of

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 119

State Correctional Administrators (Baumgartel, Guilmette, Kalb, Li, Nuni, Porter,
& Resnik, 2015, pp. 1-2), including —
1. Discretion: the latitude prison staff maintain for placing an inmate in
restrictive housing.
2. Duration: short- versus long-term placement of inmates in restrictive housing.
3. Isolation: the amount of human contact and interaction, particularly with
fellow inmates.
4. Time-in-cell: the daily length of time an inmate is confined to a cell.
Overall, it is generally agreed that restrictive housing constitutes the
discretionary practice of housing inmates in cells separate from the general
population with extended physical and social isolation over fixed or
indeterminate sentences (Baumgartel et al., 2015; Frost & Monteiro, 2016;
Shames et al., 2015; Smith, 2006).
This white paper follows the ACA standards and uses “restrictive housing” to
refer broadly to these practices.1 There are several “pathways” into restrictive
housing, reflecting its various purposes, including “safety, punishment, or
protection” (American Corrections Association, 2014). When the purpose
is protecting inmates, restrictive housing is commonly known as “protective
segregation.” Examples of inmates found in this type of housing include
celebrities, former law enforcement officers, gang dropouts, and inmates with
other types of sensitive needs. Restrictive housing for the purpose of punishment
is commonly termed “disciplinary segregation,” and for many agencies,
placement is based on custody levels that elevate with inmate misconduct.
Finally, when restrictive housing is used to manage threats and safety, it is
commonly known as “administrative segregation.” In this context, safety
generally refers to the safety of the institution — broadly construed — due to
the threat or risk posed by the inmate. Administrative segregation is especially
relevant when it comes to discussing the relationship between restrictive housing
and gang affiliation. As described in a later section, administrative segregation
is commonly used on gang affiliates. This use attracts the greatest concern from
advocacy groups, such as Solitary Watch and the American Civil Liberties Union,
because it often involves both independent and cumulative increases in staff
discretion, isolation, time-in-cell, and indeterminate placement. Distinguishing
these pathways into restrictive housing is critical to understanding the
relationship between gang affiliation and restrictive housing.

1

In reviewing the literature, terminology consistent with its original usage is reflected in this white paper if it more
appropriately represents the work of the authors, particularly for research on long-term placement in administrative
segregation or supermax housing.

120 • National Institute of Justice | NIJ.gov

Gangs
Unlike restrictive housing, the terminology applied to gangs in prison settings
is much more universal across agencies. “Security threat group” (STG) is a term
commonly applied to gangs, although it is not uncommon for agencies to use
“street gangs,” “prison gangs,” “disruptive groups,” “cliques,” and “unauthorized
organizations,” among other terms. It is beyond the scope of this white paper
to delve into the complexities of gang definitions. Drawing from the California
Department of Corrections and Rehabilitation’s (2014, p. 2) definition,2 the
essential features of STGs are —
1. Group: a formal or informal organization or association of three or more inmates.
2. Collective identity: a common name or identifying signs or symbols.
3. Durability: ongoing or durable across time.
4. Criminal activity: multiple acts of organization, threats, finance, soliciting,
or misconduct are conducted by its affiliates, individually or collectively, on
behalf of the group.
These factors distinguish STGs from fleeting associations as well other groups
(e.g., religious) found in prison settings. It is worth noting that agencies have
developed terminology and definitions for STGs that are not exclusive to gangs
in theory but are in practice. Any cursory review of the names of STGs that
agencies have recognized reveals that they are overwhelmingly street or prison
gangs. Accordingly, this white paper will use “gangs” as the terminology applied
to groups such as those described above unless referring to a specific policy
where “STG” is used.
Many agencies identify two tiers of gangs (e.g., STG-I and STG-II). This is a
critical distinction because it has implications for the potential placement of
inmates into restrictive housing. Although this issue will be addressed in more
detail in later sections, in some agencies the gang tier may dictate a specific set of
restrictive housing policies and procedures, among other responses. Generally,
three important, although interrelated, factors are used to distinguish between
tiers of gangs, including —

2

Because of recent reforms to gang management policies, finalized in 2014, the definition provided by the California
Department of Corrections and Rehabilitation includes changes to the definition of STGs. The author is unaware of
any comprehensive investigation into either academic or administrative gang definitions in prison settings. The works
of Camp and Camp (1985), Lyman (1989), and the National Institute of Corrections (1991) are often referenced for
providing the essential features of prison gangs, and Hill (2009) provided the only comparison in documentation
practices across agencies.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 121

1. Whether the origins and activities of the group are primarily in street or
prison settings.
2. If a group maintains an organizational structure that involves established
procedures, hierarchy, bylaws, and collective behavior.
3. Perhaps most importantly, the level of threat posed by the group to inmates,
correctional officers and staff, and facilities, especially the propensity for
violence and the disruption of institutional security.
Some states (e.g., the California Department of Corrections and Rehabilitation
[CDCR]) refer to both tiers as STGs,3 while other states (e.g., the Texas
Department of Criminal Justice [TDCJ])4 distinguish STGs from other
collectives that are termed street gangs or disruptive groups. And, some agencies
(e.g., the Federal Bureau of Prisons) use alternative language such as “disruptive
groups” rather than STGs or gangs for the upper tier while recognizing that
there are additional gangs and gang members present in their facilities. Agencies
typically conduct threat assessments to determine the status of groups (appendix
A identifies the criteria the CDCR uses to certify associations or groupings of
inmates as STGs).

Gang affiliation
While corrections officials use threat assessments to certify when a grouping
or association of inmates is a gang, the determination of gang affiliation for
individual inmates is based on what is commonly termed “validation” (Camp
& Camp, 1985, p. 132; Tachiki, 1995). Validation refers to a fact-finding process
where knowledgeable prison employees — usually in a classification or STG unit
3

For example, the California Department of Corrections and Rehabilitation (2014, pp. 31-32) identifies STG-I as follows:
“groups, gangs, and/or historically based prison gangs that the CDCR has determined to be the most severe threat to
the security of the institutions and communities based on a history and propensity for violence and/or influence over
other groups.” Alternatively, STG-II is defined as: “other groups or gangs such as street gangs or disruptive groups
comprised of members and associates who may be determined to be in a subservient role to the more dominant
STG-I groups.”

4

The Texas Department of Criminal Justice recognizes 12 STGs and monitors numerous disruptive groups. The
case of the Tangos may provide readers with a useful way to distinguish between gang tiers. The Tangos are a
Latino “hometown” gang (e.g., Tangos of Houston, Tangos of El Paso) that has maintained a presence in Texas
prisons for more than two decades, yet departs from the traditional Latino prison gang structure of groups like
the Texas Syndicate and the Texas Mexican Mafia (Tapia, 2014). The Tangos also have the largest gang presence
in Texas as well as the TDCJ prisons and “represent the greatest statewide gang threat” (Texas Department of
Public Safety, 2015, p. 17). However, the TDCJ classifies the Tangos as a disruptive group and not an STG, given
that the group — regardless of its size in numbers — lacks the strong organizational structure that is found in
STGs such as the Texas Mexican Mafia or Aryan Brotherhood. For example, unlike top-tier STGs, gang member
“rank” (e.g., status, shot-calling) among Tangos in TDCJ is localized to units or pods and does not transfer
when a Tango gang member moves to a different facility. Overall, the TDCJ threat assessment deems that the
threat associated with the Tangos is lesser than the more established STGs, which in turn, warrants a lower
classification as a disruptive group.

122 • National Institute of Justice | NIJ.gov

— review evidence regarding an inmate’s history and association with recognized
gangs to make an official determination of gang affiliation.5 Gang affiliation
is determined based on criteria commonly used by law enforcement agencies
to document and record street-gang members (Huff & Barrows, 2015). These
criteria focus on gang signs and symbols in written documents or photographs,
socializing with known gang members, activity on behalf of the gang, and other
forms of gang intelligence. How inmates score on an explicit list of criteria, often
termed “source items,” are entered into what some agencies call a “validation
packet” to determine gang affiliation (e.g., Arizona Department of Corrections;
see Fischer, 2002). Some agencies apply equal weighting to all source items,
while others apply smaller or larger weights depending on the source item
(see appendix B for CDCR’s 14 source items and their associated weights).
There is a great deal of overlap across prison systems in how gang members are
documented and recorded, but some variation remains (Hill, 2009, documents
the similarities and differences in 43 prison systems).
Similar to how agencies recognize tiers for different gangs, they also recognize
different levels in the status of individual gang members. These determinations
are equally important in understanding the relationship between gang affiliation
and restrictive housing because the level of gang affiliation may trigger specific
restrictive housing policies and practices. Inmates with non-zero levels of
involvement in gangs are considered “gang affiliates,” which is the operational
definition used in this white paper. A study of the patterns and correlates of
violence and misconduct in the Federal Bureau of Prisons (Gaes, Wallace,
Gilman, Klein-Saffran, & Suppa, 2002, pp. 362-363) compared a three-tier
gang member classification system with the street gang literature on gang
“embeddedness” (see Klein, 1971; Pyrooz, Sweeten, & Piquero, 2013) and
reported the following tiers:
1. A member is a “full-fledged, core gang member.”
2. A suspect is “thought to be a gang member whose credentials have not been
fully established.”
3. An associate refers to an inmate whose “actions indicate he is conducting
business or looks out for the interests of a gang” but has not or cannot join for a
variety of reasons (e.g., race or ethnicity, residence, or cultural background).
There is the fourth, or non-embedded, level that includes inmates who are
unaffiliated or unassociated with gangs altogether (i.e., zero level). This bottom

5

Although an inmate can be validated as a gang member at any point while incarcerated, the most active phase of
intelligence gathering occurs at intake or reception (Goodman, 2008; Hatcher, 2006). This concerted focus on gang
affiliation and status at intake is due to the need for finding appropriate housing for inmates. Wrongly housing gang
dropouts or rival gang affiliates could have deadly consequences for inmates or staff.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 123

level represents the largest group of inmates, as it is well known that most
inmates are not embedded in gangs (Hill, 2009; Winterdyk & Ruddell, 2010).
The notion of gang embeddedness has been likened to a bulls-eye (Esbensen,
Winfree, He, & Taylor, 2001), where members would be at the center of the
target, suspects at the inner ring, associates at the outer ring, and the nonembedded would not be located on the target at all. These distinctions can be
found in the validation process used to determine gang affiliation, where the
number of points an inmate accumulates dictates whether he or she is considered
a member, suspect, associate, or nonmember.
This section has provided an overview of the terminology and definitions of
restrictive housing, gangs, and gang members. There are several types of restrictive
housing, gangs, and gang affiliations, each of which is important to distinguish, and
the associated terminology is used throughout this white paper. Understanding
the distinctions across each of these concepts is critical to understanding their
interrelationship. Some gangs and some gang affiliates are subject to some forms
of restrictive housing. However, gangs are not monolithic to restrictive housing,
nor is restrictive housing monolithic to gangs. It is important to understand the
relationship between restrictive housing and gang affiliation, and to determine why,
when, where, and how restrictive housing is used.

A Brief Overview of Restrictive Housing, Gangs, and Gang
Affiliation in U.S. Prisons
The use of restrictive forms of housing has a long tradition in U.S. prisons
(Haney & Lynch, 1997; Hinds & Butler, 2015; King, 1999). Recent events such as
the Pelican Bay hunger strikes in 2011 and 2013 have propelled the practice of
restrictive housing into a national spotlight (Lovett, 2015). Moreover, national
leaders have expressed both interest and concern about the practice — recent
high-publicity examples include President Obama’s speech to the National
Association for the Advancement of Colored People (Baker, 2015) and U.S.
Supreme Court Justice Kennedy’s concurring opinion in Davis v. Ayala (Ford,
2015). Several U.S. legislative bills have been proposed and hearings held on
restrictive housing in the past couple of years. Although corrections issues
occasionally spill outside of prisons and into the public discourse, rarely has
there been such a concerted interest in restrictive forms of housing.
The practice of restrictive housing has grown considerably over the past several
decades (Butler et al., 2013; King, 1999; Naday, Freilich, & Mellow, 2008). In
1984, only one state (Illinois) maintained a supermax prison facility (Kurki &
Morris, 2001). In 1996, that figure increased to 32 states (National Institute of
Corrections, 1997). By 2004, 44 states had supermax prison facilities (Mears,
2005). No single source of data has systematically tracked the number of inmates,

124 • National Institute of Justice | NIJ.gov

beds, or facilities using restrictive housing over an extended period, but several
studies have produced estimates of the scope of restrictive housing.
Naday and colleagues (2008) reported the number of U.S. prison inmates housed
in administrative segregation annually between 1997 and 2002. They examined
data from The Corrections Yearbook, a report produced by the Criminal Justice
Institute based on surveys of correctional agencies and 51 agencies with at least
one valid data point during the six-year period (excluding the Bureau of Prisons).
An average of 26,177 inmates were housed in administrative segregation each
year. When paired with state-level figures of incarceration over the same sixyear period (Harrison & Beck, 2003), this translates to a 2.2 percent rate of
administrative segregation. These findings indicate that, on any given day, the
vast majority of inmates are not segregated from the general population.
Baumgartel and colleagues (2015) and Beck (2015) report the most recent data
on the scope of restrictive housing in U.S. prisons. The Baumgartel team’s findings
are based on the national survey conducted jointly by the Liman Program and
the Association of State Correctional Administrators, in which 40 state and
federal agencies reported on their custodial population in fall 2011 and fall 2014.
They found that 6.4 percent of the custodial population in 34 jurisdictions was
in restrictive housing in 2014. In terms of administrative segregation, where
40 agencies had valid data, Baumgartel and colleagues’ findings reveal that
29,848 inmates were housed in administrative segregation, or 2.6 percent of the
custodial population. There was a 19 percent reduction in the number of inmates
in administrative segregation between 2011 and 2014. The majority of states
(62 percent) reported a decline, with the Colorado Department of Corrections
observing an 86 percent reduction. When comparing the findings with Naday and
colleagues (2008), it is clear that there remains a continued — and an apparently
increased — reliance on administrative segregation in U.S. prisons.
Beck (2015) examined restrictive housing using data from the 2011-12 National
Inmate Survey, which consists of a sample of 38,251 inmates in 233 state
and federal prisons. The survey had inmates self-report the time they spent
in restrictive housing in the 12 months prior to the interview. This is a key
distinction from the work of Naday and colleagues (2008) and the Baumgartel
team (2015), both of whom relied on administrative data. Approximately 4.4
percent of inmates reported spending “last night” in restrictive housing, which
would be equivalent to the “on hand” or daily snapshot of the population housed
in restrictive housing. This number is twice that of the findings reported by
Naday and colleagues (2008), but it is important to recognize that this figure
includes restrictive housing broadly defined.6 Beck also found that nearly 20
percent of inmates reported spending at least one day in restrictive housing in
the past year, a figure that reveals a much greater exposure to restrictive housing
than daily snapshots do.
6

Baumgartel and colleagues (2015) did not report the 2011 estimates of restrictive housing, so a comparison to
Beck (2015) is not possible.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 125

The rise in the use of restrictive housing has coincided with the growth of prison
gangs. This growth may be the reason for the increased use of restrictive housing
in states like California and Texas (DiIulio, 1990; Mears, 2005, p. 33). There are
reports of prison gangs as early as the 1940s (e.g., DiIulio, 1990, p. 132), although
what is contemporarily viewed as prison gangs — large, racially and ethnically
homogenous, organized and structured, and violent groups (see Pyrooz, Decker,
& Fleisher, 2011) — did not emerge until the latter half of the 20th century.
The U.S. prison population underwent large-scale demographic changes in the
1970s, including the incarceration of more youthful, violent, and racial or ethnic
minority offenders, which has been identified as a source of the growth of prison
gangs (Skarbek, 2014). Street gangs proliferated during this era as well (Klein &
Maxson, 2006).
The state of the evidence on the landscape of prison gang activity is based
on a collection of one-off studies. Camp and Camp (1985) documented the
emergence of prison gangs in a survey of administrators from 49 prison systems
in 1984. They found that two-thirds of prison administrators reported prison
gangs in their institutions. Among the 29 agencies that reported information on
gangs, there were a total of 114 gangs in the United States. The Camp and Camp
study also provided the first national estimate of prison gang members. Although
based only on 23 agencies, they found that the number of gang members totaled
12,634, with California (2,050), Illinois (5,300), and Pennsylvania (2,400)
recording the largest numbers of gang members. Overall, states reported that an
average of 3 percent of the prison population were gang members. The findings
from Camp and Camp are important because they provided much-needed
baseline information about the nature and extent of gang activity in prisons.
Several studies have been conducted to update the estimates found in Camp and
Camp (1985), including Knox (2012), the National Gang Intelligence Center
(2011), Wells, Minor, Angel, and Carter (2002), Hill (2009), and Winterdyk
and Ruddell (2010). Since 1991, Knox has conducted several surveys of chief
administrators of jails and prisons. Between 133 and 323 institutions and
anywhere from 39 to 50 states were included in Knox’s surveys. It is unclear if
the sample of institutions in the studies were representative of the United States,
and whether the institutions included in the sample were jails or prisons cannot
be distinguished. The surveys covered a wide range of general issues, but also
included several measures relevant to gangs. Knox found much higher rates of
gang membership than Camp and Camp, and that the rate of gang membership
rose steadily throughout the 1990s and the turn of the century. However, Knox
found this dramatic increase only among the male prison population.7 The
remaining contemporary estimates of the rate of gang membership range from
12 to 15 percent, which stands in sharp contrast to the findings reported by
7

For example, Knox’s surveys revealed that the rates of gang membership for males increased from 9.3 percent
in 1995 to 29.5 percent in 2012. Rates of gang membership for females hovered around 4 to 8 percent.

126 • National Institute of Justice | NIJ.gov

Knox. In particular, Winterdyk and Ruddell extrapolated their findings to the
federal and state prison population and reported 189,000 validated and 208,000
affiliated gang members. Drawing together these estimates, it is safe to conclude
that roughly one in seven U.S. prisoners are members of gangs in prison, and it is
likely that an even larger number maintain an affiliation with gangs.
The aforementioned studies verify that gangs and gang affiliates have an
established presence in U.S. prisons, although they constitute a minority of the
population. The problems associated with gangs are greatly disproportional
to their composition of the prison population, however. This is not a new
observation. In Illinois, Jacobs (1974, p. 399) held that “the ‘gang thing’ is the
most significant reality behind the walls.” Gangs present serious challenges
in managing prisons, including inmate and staff safety, programming, and
institutional quality of life (Fleisher & Decker, 2001). And, in prison systems
like California, DiIulio (1990, p. 130) noted that “the question of how to manage
prisons has resolved itself into the question of how to manage prison gangs.”
While many correctional responses to gangs have been introduced,8 the next
section of this white paper focuses on one particular response: segregating
gang members from the general population to counter the disruptive problems
associated with gangs.

The Relationship Between Gang Affiliation and
Restrictive Housing
This section reviews the research on the relationship between gang affiliation and
restrictive housing. It examines the logic underlying restrictive housing and its
application to gangs and gang affiliates, whether there are disparities in the use
of restrictive housing among gang members, the unique policies and practices
leading gang members into and out of restrictive housing, and the empirical
evidence supporting or countering the use of this practice on gangs.

The Logic Behind Segregating Gangs and Gang Affiliates
Mears and colleagues (Mears & Reisig, 2006; Mears & Watson, 2006) have outlined
the logic of placing inmates in restrictive housing. There are three different theories
explaining why restrictive housing should result in greater systemwide safety, order,
and control in prisons: deterrence, incapacitation, and normalization theories.
Mears and colleagues called these theories “pathways” because they establish
8

Strategies include (1) out-of-state transfers to dilute the gang population, (2) enhanced prosecution of
criminal activities, (3) inmate informants, (4) interrupting communications, (5) facility lockdowns, (6) restricting
privileges such as contact visits and programming, (7) gang renunciation and debriefing, and (8) gathering
gang intelligence (Carlson, 2001; Fleisher & Decker, 2001; National Institute of Corrections, 1991; Winterdyk &
Ruddell, 2010).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 127

the ways in which placing an inmate into restrictive housing should lead to the
intended beneficial outcome. These theories are important because they are often
cited as justification for segregating gang affiliates.
Deterrence theory contains two distinct pathways: specific and general
deterrence. The “specific deterrence” pathway refers to the sanctioning of
individual offenders via placement in restrictive housing for violating rules
and engaging in misconduct. Upon return to the general population, offenders
who have experienced the restrictions and loss of privileges will therefore
avoid repeating such behaviors in the future. The “general deterrence” pathway,
alternatively, emphasizes the message that is communicated to the prison
population at large when an offender is placed in restrictive housing. Greater
systemwide compliance with prison rules should be achieved because offenders
will refrain from disruptive behavior out of fear that they, too, could end up in
restrictive housing.
Incapacitation theory also contains specific and general pathways. “Specific
incapacitation” contends that removing disruptive inmates from the general
population will lead to greater order in prisons by virtue of high-rate offenders
being placed in restrictive housing. The key to specific incapacitation is
identifying and placing the most disruptive inmates in restrictive housing,
while also ensuring that these offenders are not replaced by equally disruptive
inmates. “General incapacitation” also adheres to the principle of selecting
specific inmates for restrictive housing, but the reason for this approach instead
emphasizes the social aspects in removing disruptive inmates. Given that
problematic inmates are likely to disrupt the everyday routines of prisons and
instigate problems with other inmates that lead to conflict and misconduct, their
removal may lead to a general improvement in prison order.
Normalization theory offers three pathways that overlap. Mears and colleagues
labeled these pathways “opportunities,” “prison staff,” and “bad apples.” The
logic from these normalization pathways is as follows: once prisons are free
from the “bad apples” who disproportionately cause problems, prisons will
operate differently under a “new normal,” which will create opportunities for
better programming and allow prison staff to focus their energies elsewhere.
Normalization theory bears some similarity to the general forms of deterrence
and incapacitation theories but emphasizes a combination of programming,
targeting the most disruptive inmates, and staff involvement in prison order.
The logic of these theories extends to gangs and gang members. Indeed, gang
members have long been considered the “bad apples” in prisons. Just as law
enforcement agencies focus their policing strategies on street gangs (e.g.,
specialized gang units, civil gang injunctions, focused deterrence), correctional
agencies target the gang population operating in prisons (Trulson, Marquart, &
Kawucha, 2006).

128 • National Institute of Justice | NIJ.gov

Shelden (1991) provided the first empirical study that established gang members
as a problematic population in prison. Sixty gang members were demographically
matched to 60 non-gang members in a medium-security institution in Nevada, and
disciplinary records from their current term of incarceration were compared. Gang
members committed more disciplinary offenses than other inmates, particularly
drug and fighting offenses. Indeed, 90 percent of the gang members committed
three or more disciplinary offenses, compared to 50 percent of the non-gang
inmates. And, 80 percent of the gang members committed a fighting offense,
compared to 42 percent of the non-gang members. Camp and Camp’s (1985,
p. 52) survey of prison systems revealed 20 gang-related homicides in 1983. These
homicides were concentrated in nine of the 33 prison systems reporting that gangs
were active in their facilities. These figures might appear small, but they translate
into a gang-related homicide victimization rate of 71 per 100,000 persons — much
higher than the overall homicide rates in state prisons (Noonan & Ginder, 2013)
and in the U.S. generally (National Research Council, 2008). In a study of 298
state prisons, Reisig (2002) found that rates of inmate homicide increased with the
percentage of inmates in a facility participating in prohibited groups such as gangs.
Altogether, the findings from the above studies suggest an explicit connection
between gangs and gang members and serious violence in U.S. prisons.
Numerous studies have since investigated the extent to which gang members
disproportionately engage in prison misconduct, particularly assaults on inmates
and staff. A recent meta-analysis (Pyrooz, Turanovic, Decker, & Wu, 2016) on
the relationship between gang membership and offending included 179 empirical
studies, with 42 of them focused on incarcerated samples. The results indicate
that gang membership maintains a robust — moderately strong and stable —
relationship with offending. They also reveal that the relationship between gang
membership and offending is just as strong among incarcerated samples as it is
among non-incarcerated samples.9 This indicates that the consequences associated
with gang members in prison settings extends beyond serious violence to other
forms of institutional misconduct.
These findings are part of the justification for responding to gangs and gang
members in prison through the use of restrictive housing. The theories outlined
by Mears and colleagues (Mears & Reisig, 2006; Mears & Watson, 2006) explicitly
call for targeting the right inmates when developing policies for restrictive
housing. The disproportionate rates of institutional misconduct by gang
members and the disorder associated with gangs make them prime targets of
such practices. Correctional officials have overwhelmingly endorsed the use of
segregation on gangs and gang members for the past 30 years (Camp & Camp,
1985, p. xvi). Knox’s (2005, 2012) surveys of gangs and STGs in U.S. jails and
9

Supplemental analysis of the data used in the meta-analysis was conducted for this white paper. The 42 studies
based on incarcerated samples produced 151 effect sizes; that is, instances where the relationship between
gang membership and misconduct was quantified. Nearly two-thirds of the effect sizes were statistically
significant, or differed from zero, which means that gang affiliates were rather consistently involved in more
misconduct than non-gang inmates.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 129

prisons revealed that between 55 and 67 percent of correctional officials reported
segregation as a strategy for controlling gangs.10
Restrictive housing is not only a common strategy that staff use to manage prison
gangs; it is also viewed as an effective strategy by correctional leaders. Winterdyk
and Ruddell’s (2010) survey of officials in 37 prison systems overseeing 1.19
million prison inmates confirmed this viewpoint. The survey listed a series
of gang and STG management strategies and asked respondents to rate their
effectiveness. Respondents identified segregation and isolation as the most
effective strategy: 75 percent reported segregation was “very effective” and
another 19 percent reported that it was “somewhat effective.” And, prison officials
also reported additional privilege restrictions — such as access to community,
communication, and visits — as effective ways to manage gangs. Finally, Mears
(2005) found agreement among nearly 600 wardens on the use of supermax
placement for gang members. Nearly half agreed or strongly agreed that gang
members should be placed in supermax housing, while 83 percent of wardens
endorsed its use for gang leaders.
The characteristics of gangs and gang affiliates are consistent with the logic
underlying restrictive housing — removing part or all of this problem group
from the general population could produce benefits consistent with deterrence,
incapacitation, and normalization theories. The majority of correctional officials
have endorsed the use of restrictive housing on gang affiliates and have viewed
it as an effective tool for managing this challenging population. This does not
guarantee that restrictive housing is used disproportionately or wantonly on
gangs or gang affiliates. It is therefore necessary to take a closer look at how
restrictive housing is used in prison systems to determine if gang affiliation
operates as a correlate of restrictive housing, and, if so, whether the sources of
any disparities are disciplinary, protective, or administrative. The latter practice,
which often involves considerable discretion and indeterminate placement, is the
source of much controversy that is inextricably linked with gangs.

Gang Affiliation as a Correlate of Restrictive Housing
What are the characteristics of inmates in restrictive housing? Are they different
from the general prison population? Despite such intense interest in the practice,
there remains a severe lack of knowledge about the demographic, psychological,
behavioral, and social correlates of placement into restrictive housing,
especially among social factors like gang affiliation. Indeed, a recent systematic

10

Unfortunately, Knox did not distinguish between jails and prisons or report the prison systems that were
represented in the study, which would be needed to determine the extent to which prison systems rely on
restrictive housing.

130 • National Institute of Justice | NIJ.gov

review of the research contained 16 studies, and only three of them included
information relevant to the relationship between gang affiliation and restrictive
housing (Labrecque, 2015b). There remain no national-level, comprehensive
investigations into the segregation practices used by correctional agencies on
gang affiliates, which is a major shortcoming of this literature. Nonetheless, there
are several streams of research that can shed light on this relationship.
It is also possible to examine the relationship between gang affiliation and
restrictive housing by drawing on the official reports of correctional agencies,
including California, Colorado, and Texas (California Department of
Corrections and Rehabilitation, 2013; O’Keefe, 2005; Texas Department of
Criminal Justice, 2014a, 2014b).11 In all three states, gang affiliates constitute
a minority of the prison population (table 1). Comparing the percentage of
gang affiliates in restrictive housing to that of non-gang inmates in restrictive
housing reveals a major disparity. The relative risk of placement in restrictive
housing is much greater for gang affiliates in all three states: 5.5 times greater
in Colorado, 16 times greater in Texas, and 71 times greater in California.
O’Keefe provided further analysis of the relationship between gang affiliation
and restrictive housing in Colorado. When other predictors of administrative
segregation were accounted for — such as violent offender, ethnicity, mental
health issues, prior punitive segregation — the risk associated with gang
affiliation reduced to 4.5, although it remained the best predictor of placement
in administrative segregation. However, the risk for placement in restrictive
housing was not even across gangs or gang affiliates. O’Keefe’s results indicated
that STG associates were just as likely to be in administrative segregation as in
the general population, while STG members were nearly eight times as likely as
non-members (including STG associates, suspects, and non-STG inmates) to be
in administrative segregation.
Caution should be exercised in interpreting these findings. First, the estimates
are dated. The California Department of Corrections and Rehabilitation (2014)
has changed its policy regarding the mandatory placement of gang associates in
security housing units (SHUs), and the administrative segregation practices of
the Colorado Department of Corrections have changed dramatically in recent
years (Baumgartel et al., 2015), with Executive Director Raemisch calling for
sweeping changes in its use (Mohr & Raemisch, 2015). Indeed, a recent report by
the Colorado Department of Corrections (2016) shows that the total population
in administrative segregation reduced from 1,505 in 2011 to 141 in 2015. Second,
the number of gang affiliates is likely underestimated, especially in California
11

The author conducted supplemental analyses to produce these findings, including prevalence and relative risk
statistics. Relative risk was computed as follows:
Risk Ratio = (RHgang/GPgang)/(RHnon-gang/GPnon-gang)
where RH represents the number of gang and non-gang inmates in restrictive housing and GP represents the
number of gang and non-gang inmates in the general population.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 131

Table 1. Gang Affiliation as a Correlate of Restrictive Housing in Three States
California1

Colorado2

Texas3

2013

2003

2014

134,160

16,171

150,361

3%

26%

6.6%

Percent of gang affiliates in
restrictive housing

76%

15%

35%

Percent of non-gang inmates
housed in segregation

1.1%

2.8%

2.2%

71

5.5

16

Year
Number of inmates
Percent of gang-affiliated inmates
in prison system

Relative risk of restrictive housing
(gang vs. non-gang)
1

Includes only inmates in security housing units (California Department of Corrections and Rehabilitation, 2013).

2

Includes only inmates in administrative segregation (O’Keefe, 2005).

3

Includes only inmates who are confirmed (not associated) members of the 12 recognized security threat groups
(Texas Department of Criminal Justice, 2014a, 2014b).

and Texas. In California, as the prison system with the “dubious distinction of
holding the largest group of gang-affiliated offenders” (Carlson, 2001, p. 16),
more than 3 percent of the prison population is likely to be a gang affiliate. If
this is the case, it would likely drive down the risk ratio. In Texas, the estimates
are based only on members of the 12 STGs recognized in the prison system; the
associates of these groups and affiliates of disruptive groups are not included,
which, like California, would likely lower the risk ratio. Nonetheless, table 1
represents the current state of the evidence on the relationship between gang
affiliation and restrictive housing using inmate-level administrative data, and
together the three states represent more than one-fifth of the inmate population
in the United States.
Facility-level data also shed light on this relationship, as illustrated in the
findings reported in Beck (2015) and the Ohio Department of Rehabilitation
and Correction (2014). Beck reported results based on the 2011-2012 National
Inmate Survey. While the survey did not include a measure of gang membership,
it did ask the 38,251 inmates included in the survey about the amount of gang
activity in the 233 facilities where they were housed over the past 12 months.
Approximately one in six inmates reported that there was “a lot” of gang activity
in their facility. Beck was then able to conduct a facility-level analysis of the
concentration of inmates who reported gang activity and the concentration of
inmates who reported spending time in restrictive housing. He found that gang
activity in facilities was strongly related to both the concentration of inmates

132 • National Institute of Justice | NIJ.gov

who have spent any time in restrictive housing as well as those who have
spent prolonged periods (30 days or more) in restrictive housing. To be sure,
if a facility has greater levels of gang activity, it is also expected to have higher
concentrations of inmates in restrictive housing. These findings are important
because they establish a link between gang activity and restrictive housing, and
do so using research methods that are entirely independent of official data.
The Ohio Department of Rehabilitation and Correction (2014) reported that
inmates with gang affiliations constitute 16.2 percent of its 50,000 prison
inmates. Although it is not possible to compute risk ratios comparable to those of
California, Colorado, and Texas, it is possible to compare the prevalence of gang
affiliation across the custody levels of 21 institutions. In 13 low-security facilities
(level 1 or 2), gang affiliates comprised no more than 15.8 percent of inmates.
In six medium-security facilities (level 3), gang affiliates comprised between
21 and 42 percent of inmates. In two high-security facilities (level 4 or 5), gang
affiliates comprised between 49 percent and 63 percent of inmates. Paralleling the
findings reported by Beck (2015), the concentration of gang members in a facility
increases with the custody or security level.
The last source of information on the relationship between gang affiliation
and restrictive housing is found in more sophisticated analyses of inmate-level
administrative data, which examine the characteristics of inmates in restrictive
housing to those in the general population (Labrecque, 2015a, 2015b). The
three studies included in Labrecque’s meta-analysis reveal that gang affiliation
constitutes a reliable predictor of administrative segregation, which is consistent
with his conclusion that “difficult to manage” inmates are those who end up
in segregation. Labrecque’s study of the Ohio Department of Rehabilitation
and Correction between 2007 and 2012, which was focused on the effects of
solitary confinement on institutional misconduct, also shed light about gang
affiliation as a correlate of restrictive housing. He found a modest relationship
between maintaining any past or present affiliation with a gang and placement
in restrictive housing, as well as a relationship with the duration an inmate
remained in restrictive housing.12
Based on the research presented above, there is ample evidence to suggest a
relationship between gang affiliation and restrictive housing and little evidence to
the contrary. This conclusion is based on the following: (1) inmates who affiliate
with gangs were clearly overrepresented in restrictive housing in California,
Colorado, and Texas; (2) a facility-level relationship between gang affiliation
and restrictive housing — as the concentration of gang affiliates and gang
12

Much like the meta-analytic findings, these results are bivariate; they do not take into account alternative factors
that could explain why gang affiliates might end up in restrictive housing (e.g., gang members are younger, have
greater criminogenic risks and needs, more violent criminal histories, see: Davis & Flannery, 2001). It would be
premature to draw firm conclusions based on this research, but it represents among the soundest analyses to
date, especially when the evidence is combined with the aforementioned studies.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 133

activity increased, so too did the custody levels and restrictive housing rates in
facilities; and (3) studies that compare the characteristics of inmates in restrictive
housing to inmates in the general population revealed that gang affiliation was a
reliable predictor of placement in restrictive housing. These findings may not be
representative of all U.S. prison systems, but they demonstrate rather consistently
that gang affiliation is related to placement in restrictive housing. This evidence
is important to those concerned with correctional policies and practices. What
these findings conceal, however, is how and why gang affiliates end up in
restrictive housing at disproportionately high rates. Accordingly, it is necessary
to examine the pathways that could lead inmates who affiliate with gangs into
restrictive housing.

Pathways Into Restrictive Housing Among Gang Affiliates
Inmates are generally placed in restrictive housing for protection, discipline,
or the threat they pose. There is good reason to suspect that much of the
overrepresentation of gang affiliates in restrictive housing is due to protective and
disciplinary reasons. Numerous scenarios reveal why current and former gang
members require placement in protective custody — being housed in a facility
where they are outnumbered by gang rivals, gang dropouts with outstanding
“debts” to the gang or who have violated gang codes of conduct, or gang affiliates
who have a personal conflict with fellow or ally gang members in the same
facility. In fact, it is not uncommon for gang members to request protective
custody (Fong & Buentello, 1991) or to be placed in protective custody upon
debriefing (Fischer, 2002). And, as discussed above, given the large disparity
between gang and non-gang inmates in their rates of institutional misconduct,
there is an obvious pathway into restrictive housing for disciplinary purposes. In
both instances, there is a clear explanation for the segregation of gang members:
they have either “earned” or “need” restrictive housing.
One of the most controversial issues surrounding the use of restrictive housing
is the segregation of inmates from the general population based solely on the
threat or safety risk they pose to prison staff, other inmates, or the institution at
large. This pathway is especially controversial for gang affiliates for at least two
reasons. First, placement into restrictive housing is usually indeterminate, that
is, until the threat wanes or is eliminated. Because gang affiliation is the primary
determinant of the threat, release to the prison’s general population requires
that gang members convince authorities that they are no longer affiliated with
a gang. This is a risky or uncertain endeavor. Second, the decision to place an
inmate into restrictive housing is made administratively, which often entails wide
discretionary latitude. In cases of gang affiliation, the discretion that poses the
greatest concern involves the validation of gang members (Toch, 2007). Both of
these issues will be addressed in more detail in the following sections.

13 4 • National Institute of Justice | NIJ.gov

There is little evidence on these issues, which suggests that the extent to which
the use of restrictive housing on gang members was “earned” (disciplinary),
“needed” (protective), or based on “status” (administrative) remains unclear. In
her review of the literature on this topic, Burman (2012) observed —
It is unknown, however, how many validated prison gang members are
segregated as a result of convictions for violent and assaultive offenses
— indicating a greater propensity for violence — as opposed to in-prison
investigations that revealed only membership with the overt behavior
(p. 125).
Several reports have addressed this issue by reviewing prison system policies
and practices, including Butler and colleagues (2013) and Jacobs and Lee (2012).
These reports provide important information about the potential for the use of
restrictive housing on gang affiliates based on an inmate’s social status rather
than the inmate’s problem behavior.
Butler and colleagues (2013) reviewed the policies of 42 state prison systems in
2010 to determine the admission criteria for inmates to be eligible for “long­
term administrative segregation” in supermax facilities or units.13 They identified
five primary factors that lead to administrative segregation, including (1) repeat
violent behavior, (2) escape risk, (3) riotous behavior, (4) threat to institutional
safety, and, most importantly, (5) STG. Affiliation with a gang or STG was an
adequate reason for placement in supermax confinement in 15 of the 42 (36
percent) states (see table 2). Although gang affiliation constituted a key form of
admission criteria for many states, it was the least identified category and under
no circumstances was it the only category that states used to determine who
is eligible for long-term segregation. Nonetheless, the findings from this study
indicate that, as of 2010, membership in a gang or STG alone was sufficient
to warrant placement in supermax facilities or units. Butler and colleagues
concluded that the low rate of identifying STG affiliation as an admission
criterion was unexpected, but they held that the conduct of gang members
would likely soon qualify them for supermax placement. This is not a subtle
distinction, as it emphasizes the “behavior” of gang affiliates over the “status” of
gang affiliation.
As part of their investigative reporting for Mother Jones magazine, Jacobs and Lee
(2012) obtained information about the gang validation and segregation practices
of 44 state correctional agencies throughout the United States. Their approach
differed from that of the Butler team (2013) in that, rather than review policies,
they surveyed agencies about their practices. There can be wide divergence

13

Eight states either declined to share their non-public policies related to this practice or indicated that they did
not employ the practice.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 135

Table 2. Gang Affiliation as a Segregation Determinant, 2010 and 2012
Authors:

Butler et al. (2013)

Jacobs and Lee (2012)

Year of data collection:

2010

2012

Question content:

Membership in a gang
or STG was included as
admission criterion for
long-term segregation

Membership in a gang
or STG was a possible
determinant of placement
in segregation

California

Yes

Yes

Indiana

Yes

Yes

Ohio

Yes

Yes

Pennsylvania

Yes

Yes

Tennessee

Yes

Yes

West Virginia

Yes

Yes

Hawaii

Missing

Yes

Texas

Missing

Yes

Connecticut

Yes

Missing

Illinois

Yes

Missing

New Mexico

Yes

Missing

Colorado

Yes

No

Kentucky

Yes

No

Mississippi

Yes

No

Nebraska

Yes

No

Washington

Yes

No

Florida

No

Yes

New Hampshire

No

Yes

New York

No

Yes

South Carolina

No

Yes

Wyoming

No

Yes

Alabama .

No

No

136 • National Institute of Justice | NIJ.gov

Alaska

No

No

Arkansas

No

No

Kansas

No

No

Maine

No

No

Massachusetts

No

No

Minnesota

No

No

Montana

No

No

New Jersey

No

No

North Carolina

No

No

Oklahoma

No

No

Oregon

No

No

Rhode Island

No

No

South Dakota

No

No

Vermont

No

No

Virginia

No

No

Arizona

No

Missing

Georgia

No

Missing

Idaho

No

Missing

Louisiana

No

Missing

Michigan

No

Missing

Missouri

No

Missing

Nevada

No

Missing

Delaware

Missing

No

Iowa

Missing

No

Maryland

Missing

No

North Dakota

Missing

No

Utah

Missing

No

Wisconsin

Missing

No

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 137

between policy and practice.14 The Jacobs and Lee survey included a question
that asked respondents: “Are validated gang/security threat group members or
associates placed in segregation as a result of validation (or only, for example, as
a result of rule violations)?” Among the 40 states that responded to this question,
Jacobs and Lee found that gang affiliation was a determinant of segregation in 13
of them (see table 2). Similar to the results of Butler and colleagues (2013), this
represents a minority of agencies.
Determining if the findings reported by Jacobs and Lee (2012) and the Butler
team (2013) are consistent with one another is important because if the results
show no overlap across the policies and practices of agencies, it would diminish
confidence in these findings. Alternatively, if there were consistent findings
using different methods among independent sources (journalists versus
researchers), it would offer strong confidence in the results. Table 2 shows that
there were convergent results 69 percent of the time, which reveals a good rate of
agreement between the two studies. However, there were 10 discrepancies. In five
states — Colorado, Kentucky, Mississippi, Nebraska, and Washington — gang
membership was not a determinant of segregation in the Jacobs and Lee study,
but it was in the Butler team’s study. Likewise, in five other states — Florida,
New Hampshire, New York, South Carolina, and Wyoming — gang membership
was a determinant of segregation in the Jacobs and Lee study, but it was not in
the Butler and colleagues’ study. Some of the discrepancy may be the result of
differences in the study methods and possible changes in restrictive housing
policies that occurred over the two-year period (e.g., Colorado). But the Jacobs
and Lee study also provided additional details about how respondents answered
the question. In Colorado, for example, a respondent told Jacobs and Lee that
inmates are placed in segregation “only for rule violations.” Alternatively, in New
Hampshire, inmates could be placed in segregation due to gang status “if the
offender is deemed to be a threat to the institution based on our investigation.”
Another take on this issue is found in Wyoming, where the respondent indicated
that “inmates are not automatically put in segregation due to gang/STG status”
but “engaging in recruitment or participating in STG-related activities” could
result in segregation.
Herein lies the complexity of correctional practices and policies as they apply to
the placement of gang affiliates in restrictive housing:
1. There is a set of rules that apply only to gangs (e.g., recruitment, gang
paraphernalia) and not to other types of inmate associations that, if violated,
could result in restrictive housing; and
2. There is a set of contingencies, such as affiliation in certain types of gangs
or leadership positions within a gang, for gang members to be placed in
restrictive housing.
14

A survey is limited to the institutional memory of the respondent, whereas a policy is systematic across the
agency. Jacobs and Lee received information from public information officers, or in some cases, prison wardens
(personal communication, November 17, 2015).

138 • National Institute of Justice | NIJ.gov

The remainder of this section will emphasize these contingencies, which are
especially important in the context of indeterminate placement in restrictive housing.
States where gang status is used as a determinant for placing inmates into
restrictive housing typically use this management strategy selectively. Not all
gangs and gang members are equal security risks, as demonstrated by Gaes and
colleagues (2002) in their study of the effect of gang affiliation status (inmates)
and gang affiliation type (gangs as groups) on institutional misconduct. For
example, the Texas Department of Criminal Justice automatically places certain
affiliates of certain gangs in administrative segregation. Among the 12 STGs
recognized and numerous disruptive groups monitored in the Texas system,
only seven of the STGs are automatically placed in administrative segregation.
Members of the remaining STGs may be placed in administrative segregation at
the discretion of correctional officials, while members of disruptive groups are
not automatically segregated.
The decision to segregate members of certain gangs highlights an important
distinction: Prison systems generally reserve administratively based restrictive
housing for inmates who are members of gangs that pose the greatest threat to
the institution. This threat is inextricably linked to a group’s level of organization
and, by extension, the persistent threat the group poses to prison operations.
Indeed, if the Texas Department of Criminal Justice were to segregate gangs
based on their sheer volume of misconduct, they would target the Tangos, as
discussed above (see footnote 5). The Tangos, however, are considered only a
disruptive group and lack the organizational structure to warrant management
beyond monitoring — for example, their leadership is localized to separate
facilities rather than to the entire prison system. The downgrading of the
Texas Mafia in the summer of 2015 from automatic segregation is also a telling
example. The threat assessment determined that the organizational structure of
the gang had weakened to the point of dysfunction, which in turn translated into
a weaker threat to the prison system.
Not all gang affiliates are placed in restrictive housing. In Texas, for example,
after a confirmation process, only members of STGs are automatically placed in
administrative segregation. Suspected and associated affiliates of these groups —
including the seven groups that are automatically segregated — are not placed
in administrative segregation unless they engage in violent behavior. This speaks
to the complexity of restrictive housing in the context of gangs across prison
systems: Decisions are based on status in some cases, behavior in others, and
sometimes a combination of the two.
A similar contingency existed previously for the California Department of
Corrections and Rehabilitation (2014). In the wake of the Ashker v. Governor of
California lawsuit in 2012, gang associates were no longer automatically placed
in restrictive housing (i.e., SHUs). The Ashker settlement, however, included
explicit language indicating that inmates are no longer placed “into a SHU,

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 139

Administrative Segregation, or Step Down Program solely on the basis of their
validation status” (p. 4). Regardless of whether an inmate is a member or an
associate of an STG-I or STG-II, behavior, not status, is the primary determinant
of SHU placement. Concerns remain about the behavior-based, gang-related
pathway into SHUs in California. For example, a review of the STG disciplinary
matrix reveals that lower-level violations — particularly sections 6 “serious” and
7 “administrative” — could be argued as representing status more than behavior
(California Department of Corrections and Rehabilitation, 2014).
There is little information about the duration of placement in restrictive housing
among gang affiliates. Labrecque (2015a) found that gang affiliates remained in
restrictive housing for longer periods in the Ohio Department of Rehabilitation
and Correction. About half of the 1,100 inmates in the Pelican Bay SHU spent 10
or more years in restrictive housing, and 78 of them had spent 20 or more years
(Rodriguez, 2015). Now, gang affiliates found guilty of SHU-eligible offenses
are placed in step-down programs that aim to offer a pathway — separate from
debriefing — out of gangs and SHUs. The Texas Department of Criminal Justice,
and other correctional agencies, also offer segregation diversion and gang
disengagement programs as well.

Pathways Out of Restrictive Housing Among Gang Affiliates
Historically, the only way out of long-term segregated housing for gang affiliates
is to “snitch (or debrief), parole or die” (Tachiki, 1995, p. 1128; see also Reiter,
2012; Hinds & Butler, 2015). Gang members held in restrictive housing for
disciplinary or protective purposes follow alternative routes to segregation —
usually fixed sentences for the former, indeterminate placement (until the safety
risk wanes) for the latter. One chief criticism of restrictive housing practices
based on threat and safety risk is that it is almost impossible to exit. Gang
members who want to exit restrictive housing find themselves in a challenging
predicament: Inform on the gang or remain in segregation. Hunt and colleagues
(1993) identified this among offenders recently released from a California prison:
[I]f for example a prisoner was in a high security unit, he often found himself
unable to get out unless he “debriefed”; i.e., provided information on other gang
members. Many respondents felt that this was an impossible situation because if
they didn’t snitch their chances of getting out were minimal. As one respondent
remarked: “They [the guards] wanted some information on other people ... [s]o
I was put between a rock and a hard place. So I decided I would rather do extra
time, than ending up saying something I would later regret” (p. 402).
The issues associated with debriefing are especially sensitive because inmates who
debrief earn the label of a “snitch” and become instant targets (see, e.g., Kurki &
Morris, 2001, p. 42; Blatchford, 2008). These issues are especially magnified for

14 0 • National Institute of Justice | NIJ.gov

inmates incorrectly classified as gang members or those merely on the fringe
of the gang, as detailed in Madrid v. Gomez, who are unable to provide
convincing information about the inner workings of the gang (Tachiki, 1995).
The path out of restrictive housing has changed in recent years. A recent report
by The Marshall Project revealed that more than 30 states have developed
step-down programs leading inmates out of restrictive housing (Chammah,
2016), and Baumgartel and colleagues (2015) report 33 agencies with step-down
programs. For example, the Michigan Department of Corrections’ Incentives in
Segregation step-down program has been in place since 2009 and outlines six
stages that inmates pass through on their way from segregation to lower custody
levels. With each step, inmates earn additional privileges, such as phone calls,
visits, commissary access, and other previously banned activities or items.
The department does not include gang affiliation as a segregation determinant
(see table 2), but other states have created comparable programs tailored to their
gang population.
Both Texas and California, the two states with the largest and arguably most
violent gang populations, provide gang-affiliated inmates with pathways out
of restrictive housing. In each state, there are two ways that inmates can either
avoid or exit restrictive housing. Texas recently introduced its Administrative
Segregation Diversion Program (ASDP), which gives members of STGs returning
to prison the opportunity to avoid segregation by participating in a six-month
program. The programming — cognitive intervention, substance abuse
treatment, anger management, and treatment for criminal addictive behavior —
is very similar to the Gang Renouncement and Disassociation (GRAD) program,
which was introduced in 1999 and modeled after a program in the Connecticut
Department of Corrections (Burman, 2012). One of the key differences between
these two voluntary programs is that enrollment in ASDP is a one-time deal
made available to gang affiliates at intake who would be placed automatically
in administrative segregation, while GRAD is made available to all members of
STGs post-intake.15 Regardless of entrée, these programs offer viable alternatives
to segregation. ASDP allows inmates to be housed in the general population after
six months, although there is a three-month period of monitoring. The GRAD
program allows inmates to be housed in the general population after six months,
and allows for social interaction in less than half of that time, followed by three
months of monitoring in the general population.
Nearly 200 inmates have completed the ASDP program and more than 4,600
have completed GRAD. However, there is little evidence for determining the
effectiveness of these programs. Burman (2012) reported that “approximately
11-12 offenders who successfully completed the GRAD process since 2000 have

15

Burman (2012, pp. 234-241) reported that release date, time in segregation, and position in a gang all factor
into the decision to allow gang members into the GRAD program.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 141

been reconfirmed as [STG] members” (p. 260). More recently, Texas reported
that 19 program participants have been reconfirmed as STG members (Pinkerton,
2014). The short existence and small sample size associated with ASDP means
that the program is at least a year away from a formal evaluation, but the GRAD
program — now operating for 15 years — is ripe for determining its efficacy in
leading inmates away from gangs. Although there are numerous outcomes in
need of evaluation, a gang member reconfirmation rate of 0.004 percent would
deem the GRAD program to be extremely successful, and, potentially, a national
model for prison systems with complex gang problems.
The California Department of Corrections and Rehabilitation recently expanded
the pathways out of restrictive housing. Previously, debriefing constituted the
primary way out of SHUs. Gang affiliates also underwent six-year inactivity
reviews by the Departmental Review Board. If found inactive (i.e., no STGrelated activity for six years), the inmate was transferred to general population
and then monitored for one year. In the aftermath of the Ashker v. Governor of
California case, debriefing remains a viable pathway out of SHUs while a new
pathway — the Step Down Program — replaces the inactivity review. Unlike
debriefing, the Step Down Program does not require inmates to complete
autobiographies or inform on their gang. Originally, it was proposed as a fiveyear, five-step process. After completing the first four steps — 12 months per
step — inmates transition into a year of “monitored” status in general population,
which is step 5. Due to the Ashker settlement, the duration of each step was cut in
half and step 5 was eliminated. The program can now be completed in two years,
or six months per step.
The Step Down Program shares some similarities with the Texas GRAD program.
There is an initial 12-month period of monitoring — steps 1 and 2 — where
inmates remain in standard SHU-like conditions. With each step, inmates earn
increasing privileges — the number of photographs allowed, the portion of
monthly canteen draw, the frequency of phone calls, receipt of personal packages,
social contact, and programming. The latter two become very apparent in steps
3 and 4, where programming occurs in small groups and inmates may interact
on the yard. California has not released much in the way of public statistics on
its Step Down Program. A brief report from September 2013 indicated that
there have been 543 case-by-case reviews of validated gang members in SHU
or administrative segregation units; of those, 65 percent were approved for (at
the time) step 5 or general population and another 28 percent placed in lower
steps in the Step Down Program. In term of debriefing, there is little systematic
information in the way of disengagement in California.16 It is therefore difficult
to determine how well California’s Step Down Program and other routes out of
restrictive housing have been implemented and whether they are successful.
16

In fact, only data from the Arizona Department of Corrections reveals information about disassociation from
gangs via debriefing (Fischer, 2002). About 14 percent of the inmates validated as gang members were
debriefed. Fischer found some positive results, but overall the findings were mixed in terms of whether
renouncement led to changes in misbehavior.

142 • National Institute of Justice | NIJ.gov

Regardless of whether the behavior or status of gang affiliates leads them into
restrictive housing, and whether programs are useful in getting them out, one
of the most fundamental questions of this practice is whether it achieves its
intended goals of reducing misconduct and disorder in prisons.

Gang Affiliation, Restrictive Housing, and the Reduction of
Misconduct and Disorder
The practice of segregating gang affiliates is both highly endorsed by correctional
authorities, who see it as a most effective solution to prison gang activity, and
highly criticized by activists and scholars, who claim the practice is empirically,
legally, or morally unsupported. Neither side has much in the way of evidence to
back their claims. In fact, despite such intense interest in the topic, to the best of
the author’s knowledge there have been only three studies to assess the effectiveness
of restrictive housing in reducing misconduct among inmates and disorder in
prisons (Fischer, 2002; Labrecque, 2015a; Ralph & Marquart, 1991). It is curious
that a practice as controversial and widespread as restrictive housing — particularly
segregation based on threat and safety — has not been subject to more formal
evaluations. Determining if restrictive housing is achieving its intended goals of
reducing misconduct, disrupting gang activity, increasing staff and inmate safety,
and decreasing prison disorder should be a high priority for correctional officials,
advocacy groups, and researchers interested in correctional practice.
Ralph and Marquart (1991) examined system-level trends in inmate violence
before, during, and after the “war years” in Texas prisons. In 1984 and 1985, there
were more murders in Texas prisons than in the prior 20 years combined. The
initial attempts of prison officials to regain control by targeting gang leadership
through segregation and transfers were ineffective. Prison officials then made
a drastic change: They moved all gang members into segregated housing in
September 1985. The prison system then witnessed a large-scale, immediate
decline in homicide along with a delayed, albeit substantively meaningful,
decline in assault. Ralph and Marquart concluded, “Although this method of
control seems to have eliminated many problems experienced by prison officials,
it is unknown what impact this has on the inmate and the community” (p. 48).
Others remained skeptical. Mears (2005) held that any firm conclusions about
the role of segregation in reducing violence in Texas prisons is premature. The
outcomes studied were narrow (limited to homicide and assault) when there
needed to be a broader range of determinants of success, and the data were of
“questionable utility” (Mears, 2005, p. 34). The findings of Ralph and Marquart
are largely descriptive and do not account for alternative explanations for the
decline of violence. It would require nonetheless a very convincing explanation to
contend that factors other than segregation were responsible for triggering such a
major decline in violence.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 14 3

Table 3. Pathways Out of Restrictive Housing in California and Texas
Texas

California

Administrative
Segregation
Diversion
Program (ASDP)

Gang
Renouncement
and
Disassociation
(GRAD)

Step Down
Program

Debriefing

Implementation year

2014

2000

2012

~1980s

Target population

Members of
STGs returning
to Texas
facilities who
want to exit
gangs and avoid
administrative
segregation

Members of
STGs who
want to
exit gangs
and leave
administrative
segregation

STG affiliates
(members and
associates)
sentenced
to security
housing unit
terms

STG affiliates
(members and
associates) who
want to exit
gangs and leave
the security
housing unit

Early stage monitoring

Admission into
the ASDP occurs
at intake

Offenders
undergo a
disassociation
monitoring
period of 12
months prior to
entry into the
GRAD program

Steps 1 and
2: "primarily
intended as
periods of
observation,"
restricted
movement,
in-cell
programming;
12 months
total (6 months
per step)

None

Program phases/steps

Six months
of classroom
setting,
group-based
programming
immediately
after admission
into Texas
Department of
Criminal Justice

Phase I: 3
months,
singlecelled, in-cell
programming,
recreation with
other offenders

Step 3: 6
months,
programming
in small
groups,
movement in
restraints

Interview phase:
gathering
information
about STGs
to determine
if offender
has dropped
out, including
autobiography

Phase II:
3 months,
double-celled,
classroom
programming,
less restricted
movement
Phase III:
general
population,
general
programming,
and work
assignments

14 4 • National Institute of Justice | NIJ.gov

Step 4: 6
months,
programming
in small
groups,
recreation
with other
offenders,
less restricted
movement

Observation
phase: 6 months,
transitional
housing unit with
other debriefing
offenders

Table 3. Pathways Out of Restrictive Housing in California and Texas (continued)
Texas

California

Administrative
Segregation
Diversion
Program (ASDP)

Gang
Renouncement
and
Disassociation
(GRAD)

Step Down
Program

Debriefing

Implementation year

2014

2000

2012

~1980s

General population

Last 3 months in
the program

Last 3 months
in the program

Upon
completion
of Steps 1-4

Upon completion
of observation
phase

Total time

6 months of
programming

12 months of
monitoring,
9 months of
programming

24 months

Varies

Participants

187

4,607

150

n/a

Sources of information

Burman (2012)
Personal communication
(December 3, 2015)
Texas Department of Criminal
Justice (2014)

Ashker v. Governor of California
(2015)
California Department of
Corrections and Rehabilitation
(2013, 2014)
Tachiki (1995)

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 14 5

Fischer (2002) conducted the most comprehensive study on gang affiliation,
restrictive housing, and misconduct. His work was based on data gathered from
Arizona Department of Corrections prisons and focused on two factors relevant
to the question at hand. First, the study determined if there was an incapacitation
effect for nine forms of violent and disruptive behavior associated with placing
431 validated gang members in a supermax unit. The results of the inmate-level
analysis confirm such an incapacitation effect: Violations fell by 30 percent, and
assault, drug, threat, fighting, and rioting violations fell anywhere from 53 to
97 percent. Other violations increased, however, including possessing weapons,
destroying property, tampering with equipment, and “other” violence. These
incapacitation findings from Fischer’s study remain the most convincing results to
date that segregation can reduce problematic behavior among members of gangs.
Second, Fischer also conducted a prison systemwide analysis of changes to
Arizona’s STG policies in 1995 and 1997. The gang policy initiative of July
1995 expanded the scope of the STG program (including documenting and
validating gangs) and intensified the penalties for STG behavior. The gang
policy of 1997 resulted in greater specification in the STG program (including
monitoring processes) and expanded the sanctions for membership in such a
group to include segregation for gang members who did not renounce their gang
affiliation and debrief with intelligence officers. Fischer examined violations
across three periods: (1) July 1994 to September 1995, (2) October 1995 to
September 1997, and (3) October 1997 to December 2000. By isolating the
periods in which the gang policies were implemented, Fischer was able to assess
whether the intended changes were achieved.
With each successive STG policy change, the Arizona Department of Corrections
observed an overall, systemwide 16 percent and 27 percent reduction in
violations for gang inmates, and 15 percent and 29 percent reduction in
violations for non-gang inmates in subsequent periods. These results are
encouraging, but it is important to note that gang members were higher-rate
offenders to begin with. Focusing on absolute violation rates shows that the
effects of changes to STG policies were especially strong for validated gang
members — the drop was four times greater for them than non-gang inmates
(p. 63). Fischer held that the implementation of these policies resulted in 21,984
fewer total violations, including 5,716 fewer among gang members. He concluded
that there was a systemwide benefit associated with the implementation of both
gang policies, one that was strongest among gang members but carried over
to non-gang members as well, which was predicted by the theories discussed
above. Fischer tempered his conclusion, however, by noting that there were many
security-related changes taking place in Arizona prisons, and those changes also
could have explained some of the overall reductions in violations.
Finally, Labrecque (2015a) conducted an individual-level, longitudinal study
of the effects of restrictive housing on subsequent misconduct in Ohio prisons
between 2007 and 2012. He examined both the placement of inmates in restrictive

14 6 • National Institute of Justice | NIJ.gov

housing and the duration of time inmates spent there, then determined if the
effects of restrictive housing were worse for gang affiliates (among other factors)
than non-gang inmates. Labrecque’s findings, while not uniform across the
prevalence and incidence of violent, nonviolent, and drug misconduct, generally
point in the same direction: Gang affiliates fared worse in terms of misconduct
when returning to the general population. Spending any time and spending
more time in restrictive housing corresponded with higher incidences of violent
and nonviolent misconduct among gang affiliates than non-gang inmates. These
findings are generally consistent with what is known about the effects of gang
affiliation on misconduct, but they also refute the notion that restrictive housing
may have specific deterrent effects for a problematic population such as gang
affiliates. Alternative theories — incapacitation and normalization — might
prove to be more defensible when subject to solid empirical investigations such
as those provided by Labrecque.

Gang Affiliation and Restrictive Housing — A Look Toward
the Future
The findings from the preceding sections constitute the empirical knowledge
base on the relationship between gang affiliation and restrictive housing. This
knowledge base remains meager in volume, with little empirical research to guide
the dialogue, which is unfortunate given how amplified this debate has become.
This section summarizes the key findings of this white paper, identifies critical
gaps in the field’s knowledge of gang affiliation and restrictive housing, and offers
productive directions for future research.

What Do We Know About Gang Affiliation and Restrictive Housing?
Several conclusions are possible — some preliminary, others firm — from
the available literature. First, as high-rate offenders and disproportionate
contributors to violent misconduct in prison, gangs and gang members fit
squarely into the logic — deterrence, incapacitation, and normalization theories
— underlying the use of restrictive housing in prison systems. To the extent that
the correlates of restrictive housing are based on discipline, protection, and safety
and threat risks, it would be remarkable if the affiliates of gangs were not the
targets of segregation policies and practices in prison systems.
Second, evidence suggests that inmates who affiliate with gangs are
overrepresented in restrictive housing. The evidence base to support this
assertion remains confined to select states, but there is also no evidence to
the contrary indicating that gang affiliation is not a correlate of placement in
restrictive housing, particularly for validated gang members.
Third, at least some of the overrepresentation of gang affiliates in restrictive
housing can be attributed to correctional policies and practices that automatically
Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 147

segregate certain types of gangs and gang members. Not all agencies employ
these practices, and no formal analysis has been conducted to differentiate
agencies that employ these practices from those that do not. It remains unclear
how much of the overrepresentation of gang affiliates in restrictive housing is due
to disciplinary, protective, or administrative segregation.
Fourth, the most controversial aspects of the relationship between gang affiliation
and restrictive housing pertain to the purported arbitrariness and lack of due
process in validating gang affiliates, the policies that automatically segregate gang
affiliates based on status and not behavior, and the indeterminate placement of
gang affiliates in restrictive housing.
Finally, the emergence of programs aiming to remove gang affiliates from
restrictive housing and encourage disengagement from gangs — particularly in
states such as California and Texas — reveals a shift from more punitive policies
to those that encourage integration into the general prison population housing.
However, there is very little evidence to determine if these programs are effective
at achieving their intended goals.

What Do We Need to Know About Gang Affiliation and
Restrictive Housing?
There are several aspects of the relationship between gang affiliation and
restrictive housing that are in need of concerted attention from policymakers,
practitioners, and researchers. This section identifies four key gaps in our
knowledge. Some of them will require substantial and sustained investment in
research infrastructure, while others may be accomplished more swiftly by using
archival data. In both cases, researcher-practitioner partnerships are key for
advancing this agenda.

Collect and analyze data that will establish an empirical foundation on the use
of restrictive housing among gangs and gang affiliates
Does segregating the affiliates of gangs from the general inmate population
reduce disorder in prisons and misconduct among inmates? This question is
directed toward prison systems that segregate gang affiliates based on their status.
It is equally important for prison systems that segregate gang affiliates based
on their behavior, especially if the violations involve actions unique to gangs
(e.g., possessing gang paraphernalia, participating in gang roll call). The recent
changes that the California Department of Corrections and Rehabilitation made
to its SHU-eligible offense policy illustrate the importance of the latter.
To date, there is scant evidence to support either status- or behavior-based
arguments for segregating gang affiliates. In fact, only three studies have
evaluated this line of inquiry — a macro, system-level study in Texas (Ralph

14 8 • National Institute of Justice | NIJ.gov

& Marquart, 1991); an individual- and system-level study in Arizona (Fischer,
2002); and an individual-level study in Ohio (Labrecque, 2015a). These studies
indicated that segregating gang affiliates reduces homicide, assault, and other
forms of violence, but that restrictive housing could introduce problems when
gang affiliates return to the general population. These findings provide a strong
foundation for future research. Archival data could be used to conduct betweeninmate and within-inmate effects of segregation on misconduct among gang and
non-gang groups.
Research on this topic should track a wide range of activities among inmates as
they move into and out of restrictive housing to isolate the independent effects of
this practice on inmates. This analytic logic can extend to facility-level analyses
as well, although detecting and isolating effects at aggregate units of analysis
might be difficult because of simultaneous changes that could undermine causal
relationships, such as those observed by Ralph and Marquart in Texas and
Fischer in Arizona. In addition to homicides and assaults, the research should
be concerned with a broad spectrum of outcomes, including contraband use
and seizures, yard disturbances, security incidents, and seemingly minor forms
of gang-related activities. In addition, this research should be a priority for
understanding how restrictive housing practices influence the gang affiliates
themselves, particularly their criminogenic risks and needs, as it relates to
institutional programming and re-entry success. This will provide the best
picture of the effects of placing gang affiliates in restrictive housing. In an era
when prison violence is at historically low levels, the push for policy changes
should be based on sound, empirical evidence. Whether or not gang status- or
behavior-based restrictive housing practices produce the intended effects, this
research should offer a blueprint for future correctional decision-making on this
highly charged issue.

Document national practices and trends on segregating gangs and gang
affiliates to understand and explain its use
When interested parties seek information about gang activity in prisons, there
are few places they can turn. Fleisher and Decker (2001, p. 2) described prison
as the “final frontier” in research on gangs. Indeed, the volume of the literature
on gangs in incarcerated settings pales in comparison to the street gang literature
(Pyrooz & Mitchell, 2015). Unlike national data sources on issues pertaining
to street gangs and gang membership, such as the National Youth Gang Survey
among law enforcement agencies or a variety of national-level surveys among
youth and young adults, “there is no centralized repository of data on prison
gang membership” (Gaston & Huebner, 2015, p. 329). As this white paper has
made clear, the literature on this topic is a collection of one-off studies rather
than an organized and systematic effort to document gang activity and restrictive
housing practices in U.S. prisons.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 149

Against this backdrop, it is impossible to paint a national portrait of the
relationship between gang affiliation and restrictive housing. Surveys such as
those conducted independently on restrictive housing (e.g., Baumgartel et al.,
2015) and those conducted on gangs (e.g., Winterdyk & Ruddell, 2010) are one
pathway forward. This is a reluctant recommendation, as these studies are not
easy to conduct. Well-developed surveys with methodologically sound research
designs require substantial resources to both design and administer. They
would also require negotiating the research policies of 51 prison systems, along
with wide and often conflicting terminology and definitions of gangs and gang
affiliates (Trulson et al., 2006).
An alternative pathway forward is to either develop or integrate repeated and
standardized surveys of (1) the population of prison systems, (2) nationally
representative samples of prison facilities, or (3) nationally representative
samples of prison inmates. These surveys would include a range of measures
related to restrictive housing, gangs, and gang affiliation. Rather than launching
a project of this scope from the ground up, a more feasible solution would be
to integrate measures of gangs and gang affiliation into the suite of existing
Bureau of Justice Statistics projects. For example, a single item added to the
2011-12 National Inmate Survey asking inmates about gang affiliation could
have provided the definitive evidence needed to reach conclusions about the
relationship between gang affiliation and restrictive housing. Other projects
could be used as well (e.g., National Prisoners Statistics). This approach would
yield foundational information about the frequency and trends in the use of
restrictive housing on gang affiliates. Such information would not only permit
the explanation of different practices and policies across prisons and prison
systems but would also allow these prisons and prison systems to learn from
each other — both the successes and failures — in managing a very challenging
population of inmates.

Subject gang and gang affiliate validation practices to independent scientific
assessments to establish their reliability and validity
There are many controversial issues associated with restrictive housing in U.S.
prisons, and the wholesale placement of gang affiliates in restrictive housing is
among the most pressing. Indeed, this practice contributed to the uprising that
led to the 2013 Pelican Bay hunger strike involving 30,000 inmates in California.
Therefore, this issue shines a bright light on the policies and practices followed to
validate gangs and gang affiliates in prison, which have been criticized by many
because of their implications for housing, programming, and other consequences
(Kassel, 1998; Shaiq, 2013; Sowa, 2012; Tachiki, 1995). Toch (2007) likened the
validation process to a witch hunt, suggesting that —
… a striking parallel to traditional inquisitorial tribunals can be found in [gang]
‘classification’ systems (such as those used in Arizona and California and by the

150 • National Institute of Justice | NIJ.gov

Federal Bureau of Prisons) deployed for gang members who are being considered
for consignment to ‘supermax’ control units (p. 274).
The issues are magnified when the validation process involves evidence derived
from gang debriefing — as illustrated in Madrid v. Gomez, debriefing is an
imperfect process. Gang affiliates in restrictive housing may inaccurately name
snitches, gang dropouts, rival gang members, or other vulnerable populations to
facilitate their return to general population. Incorrectly classified gang affiliates
then have no recourse through which to challenge the validation because
the information is confidential, nor can they debrief because they lack gang
intelligence. Validation can come across as arbitrary, yet it can have serious
consequences, opening up correctional agencies to costly litigation (Austin &
Irwin, 2011, pp. 134-135).
The practice of gang validation itself needs to be “validated.” Are the various
source items (see Hill, 2009) used to determine gang affiliation internally
consistent? Are these items related to each other? Do independent raters of the
source items similarly validate inmates as gang or non-gang affiliates? The source
items clearly have face validity, but there are serious questions about the criterionrelated validity of the gang affiliate validation. Do the source items predict other
forms of gang-related activity or future gang-related activity? And, do the source
items correlate with indicators that they should be related to theoretically?
California’s movement to a weighted validation system, where source items are
assigned different weights, represents an important shift in policy. Nonetheless, it is
unclear if there is an empirical basis for the weights linked to this new approach (or
the approach where source items are weighted equally).
More accurate determinations of gang affiliation will not only improve the
legitimacy of the practice but may also reduce institutional conflict. False
positives (i.e., wrongly classifying an inmate as a gang affiliate) and false
negatives (i.e., wrongly classifying an inmate as a non-gang affiliate) represent
serious threats to the safety of institutions. This applies to determining not only
who is a gang affiliate but also who has disengaged from gangs. A productive
start would involve subjecting gang validation practices to independent scientific
assessments of reliability and validity. Given just how great of an impact gang
validation can have on inmates’ lives, these practices must be open to rigorous
scientific scrutiny, and a threshold must be established if these practices are to
dictate placement in restrictive housing.

Evaluate the effectiveness of programs and policies that promote
disengagement from gangs and exit from restrictive housing to establish
best practices
Do programs designed to facilitate desistance from gangs and exits from
restrictive housing actually work? With programs in as many as 30 states,

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 151

including gang-targeted programs in states like California and Texas, step-down
processes and impacts must undergo rigorous scientific evaluation. Randomized
control trials should be the standard used to determine a program’s effectiveness.
Practitioners often find themselves uncomfortable withholding needed
programming from control groups. However, there are often more candidates
than there is funding and space for programming (e.g., Burman, 2012). In
instances where randomized control trials are impractical, quasi-experimental
designs can be attractive alternatives. Program participants can be matched on
a range of characteristics with inmates who are eligible for the program but, for
one reason or another, decline to participate.
When possible, natural experiments should be exploited for their intellectual
yields. California’s shift in 2012 to no longer automatically segregate STG-I
gang associates, and then in 2015 to no longer segregate STG-I gang members,
represents a natural experiment ripe for scientific investigation. Following the
logic of deterrence, incapacitation, and normalization theories, there are many
ways to assess if this policy shift has had positive or negative impacts on inmates
and institutions.
Researcher-practitioner partnerships are needed to foster knowledge on
restrictive housing (Frost & Monteiro, 2016). This is especially true with gang
research. Skarbek (2014, p. 10) noted that “the same walls that keep inmates
locked in also keep researchers out,” a point others have recognized for more
than two decades (Fong & Buentello, 1991). Although archival data are often
rich with information that could help evaluate programs, these data are
inflexible regarding specific items that are essential for determining program
efficacy. Especially for process evaluations, researchers need access to facilities
to interview prison staff and inmates. All data have limitations, but the luxury
of triangulating information on programs is not one to pass up. Researchers
can sometimes conduct analyses that are tangential to the goals of correctional
research departments overloaded with other tasks, thus revealing important
information about agency practices and the inmate population.
A final point to be made about program evaluations is that not conducting them,
or withholding their results, is a disservice to the correctional landscape. The
results of scientifically rigorous research must be made public and disseminated
widely. This is especially true for step-down and diversion programs designed
for gang affiliates. Information must be shared if researchers and stakeholders
are to learn what is and is not working in responding to gangs in prisons. Gang
research reveals that most programs in street settings fail to find any positive
effects, and some find programs backfire in unanticipated ways (Gravel,
Bouchard, Descormiers, Wong, & Morselli, 2013; Klein & Maxson, 2006). When
programs are successful, they have the potential to translate across contexts
(e.g., Operation Ceasefire); when they are not, they have the potential to retool
under more scientifically sound principles (e.g., Gang Resistance Education and
Training). Either way, establishing the legitimacy of a program is important to
the day-to-day activities of correctional staff and inmates in these programs.
152 • National Institute of Justice | NIJ.gov

Conclusions
Gangs are especially challenging populations to manage. The problems associated
with gangs in prison have not abated, and restrictive housing is one of the few
practices that correctional officials see as viable for managing the population.
This solution, however, has come under heightened scrutiny in recent years, due
in no small part to the vocal stand that gang affiliates have taken in places like
California. Restrictive housing is an imperfect solution to a complex problem.
Neither the wholesale placement of gang affiliates into restrictive housing based
merely on their status, nor the wholesale removal of gang populations from
restrictive housing, are practical solutions to this impasse. The limited empirical
knowledge base, unless addressed, will only allow these issues to fester and
grow. Much has been learned about the relationship between gang affiliation
and restrictive housing over the past three decades. What is needed is a dialogue
that is informed not by rhetoric or anecdote but by empirical evidence to guide
decisions about the future of restrictive housing and gangs in U.S. prisons.

References
American Corrections Association. (2014). Standards committee meeting
minutes. Presented at the ACA Winter Conference, Tampa, FL. Retrieved
from http://www.aca.org/ACA_PROD_IMIS/Docs/Standards%20and%20
Accreditation/SC%20Minutes_Tampa_Jan2014.pdf.
Austin, J., & Irwin, J. (2011). It’s about time: America’s imprisonment binge.
Boston, MA: Cengage Learning.
Baker, P. (2015, July 14). Obama calls for effort to fix a “broken system” of
criminal justice. The New York Times. Retrieved from http://www.nytimes.
com/2015/07/15/us/politics/obama-calls-for-effort-to-fix-a-broken-system-of­
criminal-justice.html.
Baumgartel, S., Guilmette, C., Kalb, J., Li, D., Nuni, J., Porter, D. E., & Resnik,
J. (2015). The ASCA-Liman 2014 national survey of administrative segregation
in prison. New Haven, CT: The Liman Program, Yale Law School, and the
Association of State Correctional Administrators.
Beck, A. J. (2015). Use of restrictive housing in U.S. prisons and jails, 2011–12.
NCJ 249209. Washington, DC: Bureau of Justice Statistics. Retrieved from
http://www.bjs.gov/content/pub/pdf/urhuspj1112.pdf.
Blatchford, C. (2008). The black hand: The story of Rene “Boxer” Enriquez and his
life in the Mexican Mafia. New York: Harper.
Burman, M. L. (2012). Resocializing and repairing homies within the Texas
prison system: A case study on security threat group management, administrative

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 153

segregation, prison gang renunciation and safety for all. Austin, TX: The University
of Texas at Austin.
Butler, H. D., Griffin, O. H., & Johnson, W. W. (2013). What makes you the
“worst of the worst?” An examination of state policies defining supermaximum
confinement. Criminal Justice Policy Review, 24(6), 676-694. http://dx.doi.org/
org/10.1177/0887403412465715.
California Department of Corrections and Rehabilitation. (2013). Security
housing units. Sacramento, CA: California Department of Corrections and
Rehabilitation. Retrieved from http://www.cdcr.ca.gov/stg/docs/Fact%20Sheet­
SHU%20with%20photos.pdf.
California Department of Corrections and Rehabilitation. (2014). Security threat
group (STG) information. Sacramento, CA: California Department of Corrections
and Rehabilitation. Retrieved from http://www.cdcr.ca.gov/stg/docs/Final_Text_
of_Adopted_Regulations_STG.pdf.
Camp, G. M., & Camp, C. G. (1985). Prison gangs: Their extent, nature, and
impact on prisons (pp. 1-259). Washington, DC: U.S. Department of Justice.
Carlson, P. M. (2001). Prison interventions: Evolving strategies to control
security threat groups. Corrections Management Quarterly, 5(1), 10.
Chammah, M. (2016, January 7). How to get out of solitary — one step at a
time. The Marshall Project (with The Atlantic). Retrieved January 14, 2016, from
https://www.themarshallproject.org/2016/01/07/how-to-get-out-of-solitary-one­
step-at-a-time.
Colorado Department of Corrections. (2016). SB 11-176 annual report:
Administrative segregation for Colorado inmates. Colorado Springs, CO: Colorado
Department of Corrections, Office of Planning and Analysis. Retrieved from
https://drive.google.com/file/d/0B8WLSXAb0Mg8Wkw2T2FBelRzOTA/
view?pref=2&pli=1&usp=embed_facebook.
Crouch, B. M., & Marquart, J. W. (1989). An appeal to justice: Litigated reform of
Texas prisons. Austin, TX: University of Texas Press.
Davis, M. S., & Flannery, D. J. (2001). The institutional treatment of gang
members. Corrections Management Quarterly, 5(1), 37-46.
DiIulio, J. J. (1990). Governing prisons: A comparative study of correctional
management. New York: Free Press.
Esbensen, F.-A., Winfree, L. T., He, N., & Taylor, T. J. (2001). Youth gangs and
definitional issues: When is a gang a gang, and why does it matter? Crime &
Delinquency, 47(1), 105-130. http://dx.doi.org/10.1177/0011128701047001005.

15 4 • National Institute of Justice | NIJ.gov

Fischer, D. R. (2002). Arizona Department of Corrections: Security threat group
(STG) program evaluation. Final report. Washington, DC: National Institute of
Justice, Office of Justice Programs, U.S. Department of Justice.
Fleisher, M. S., & Decker, S. H. (2001). An overview of the challenge of prison
gangs. Corrections Management Quarterly, 5(1), 1-11.
Fong, R. S., & Buentello, S. (1991). The detection of prison gang development: An
empirical assessment. Federal Probation, 55(1), 66-70.
Ford, M. (2015, June 18). Justice Kennedy denounces solitary confinement. The
Atlantic. Retrieved from http://www.theatlantic.com/politics/archive/2015/06/
kalief-browder-justice-kennedy-solitary-confinement/396320/.
Frost, N. A., & Monteiro, C. E. (2016). Administrative segregation in U.S. prisons.
Washington, DC: U.S. Department of Justice, Office of Justice Programs, National
Institute of Justice.
Gaes, G. G., Wallace, S., Gilman, E., Klein-Saffran, J., & Suppa, S. (2002). The
influence of prison gang affiliation on violence and other prison misconduct. The
Prison Journal, 82, 359-385. http://dx.doi.org/10.1177/003288550208200304.
Gaston, S., & Huebner, B. M. (2015). Gangs in correctional institutions. In S. H.
Decker & D. C. Pyrooz (Eds.), The handbook of gangs (pp. 328-344). Chichester,
West Sussex, England: John Wiley & Sons.
Goldstone, J. A., & Useem, B. (1999). Prison riots as microrevolutions: An
extension of state-centered theories of revolution. American Journal of Sociology,
104(4), 985-1029. http://dx.doi.org/10.1086/210134.
Goodman, P. (2008). It’s just Black, White or Hispanic: An observational study
of racializing moves in California’s segregated prison reception centers. Law &
Society Review, 42(4), 735-770.
Gravel, J., Bouchard, M., Descormiers, K., Wong, J. S., & Morselli, C. (2013).
Keeping promises: A systematic review and a new classification of gang control
strategies. Journal of Criminal Justice, 41(4), 228-242.
Griffin, M. L. (2007). Prison gang policy and recidivism: Short-term management
benefits, long-term consequences. Criminology & Public Policy, 6(2), 223-230.
Griffin, M. L., & Hepburn, J. R. (2006). The effect of gang affiliation on violent
misconduct among inmates during the early years of confinement. Criminal
Justice and Behavior, 33(4), 419-466.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 155

Haney, C., & Lynch, M. (1997). Regulating prisons of the future: A psychological
analysis of supermax and solitary confinement. NYU Review of Law & Social
Change, 23, 477.
Harrison, P. M., & Beck, A. J. (2003). Prisoners in 2002. Washington, DC: U.S.
Department of Justice, Office of Justice Programs, Bureau of Justice Statistics.
Retrieved from http://www.ncjrs.gov/App/abstractdb/AbstractDBDetails.
aspx?id=200248.
Hatcher, L. D. (2006). Security threat groups: It is more than managing gangs in a
local facility. Corrections Today, 68(2), 54-58.
Hill, C. (2009). Gangs/security threat groups. Corrections Compendium, 34(1), 23-37.
Hinds, M., & Butler, J. (2015). Solitary confinement: Can the courts get inmates
out of the hole? Stanford Journal of Civil Rights & Civil Liberties, 11, 331-372.
Huebner, B. M. (2003). Administrative determinants of inmate violence: A
multilevel analysis. Journal of Criminal Justice, 31(2), 107-117. http://dx.doi.
org/10.1086/21013410.1016/S0047-2352(02)00218-0.
Huff, C. R., & Barrows, J. (2015). Documenting gang activity: Intelligence
databases. In S. H. Decker & D. C. Pyrooz (Eds.), The handbook of gangs.
Chichester, West Sussex, England: John Wiley & Sons.
Huff, C. R., & Meyer, M. (1997). Managing prison gangs and other security threat
groups. Corrections Management Quarterly, 1(4), 10-18.
Hunt, G., Riegel, S., Morales, T., & Waldorf, D. (1993). Changes in prison culture:
Prison gangs and the case of the “Pepsi generation.” Social Problems, 40(3), 398-409.
Irwin, J. (1980). Prisons in turmoil. Boston, MA: Little, Brown.
Jacobs, J. B. (1974). Street gangs behind bars. Social Problems, 21(3), 395-409.
Jacobs, R., & Lee, J. (2012). Maps: Solitary confinement, state by state. Retrieved
October 15, 2015, from http://www.motherjones.com/politics/2012/10/map­
solitary-confinement-states.
Kassel, P. (1998). The gang crackdown in Massachusetts’ prisons: Arbitrary and
harsh treatment can only make matters worse. New England Journal on Criminal
& Civil Confinement, 24, 37.
King, R. D. (1999). The rise and rise of supermax: An American solution
in search of a problem? Punishment & Society, 1(2), 163-186. http://doi.
org/10.1177/14624749922227766.

156 • National Institute of Justice | NIJ.gov

Klein, M. W. (1971). Street gangs and street workers. Englewood Cliffs, NJ:
Prentice-Hall.
Klein, M. W., & Maxson, C. L. (2006). Street gang patterns and policies. New York:
Oxford University Press.
Knox, G. W. (1991). Findings from the 1991 Adult Corrections Survey: A
preliminary report from the National Gang Crime Research Center. Chicago, IL:
National Gang Crime Research Center. Retrieved from http://www.ngcrc.com/
ngcrc/page91w.htm.
Knox, G. W. (1995). A national assessment of gangs and security threat groups
(STGs) in adult correctional institutions: Results of the 1995 adult corrections
survey. Chicago, IL: National Gang Crime Research Center. Retrieved from
http://www.ngcrc.com/ngcrc/page95w.htm.
Knox, G. W. (1999). A national assessment of gangs and security threat groups
(STGs) in adult correctional institutions: Results of the 1999 adult corrections
survey. Chicago, IL: National Gang Crime Research Center. Retrieved from
http://www.ngcrc.com/ngcrc/page7.htm.
Knox, G. W. (2005). The problem of gangs and security threat groups (STG’s) in
American prisons today: Recent research findings from the 2004 Prison Gang
Survey. Chicago, IL: National Gang Crime Research Center. Retrieved from
http://www.ngcrc.com/ngcrc/corr2006.html.
Knox, G. W. (2012). The problem of gangs and security threat groups (STG’s) in
American prisons and jails today: Recent findings from the 2012 NGCRC National
Gang/STG Survey. Chicago, IL: National Gang Crime Research Center. Retrieved
from http://www.ngcrc.com/corr2012.html.
Kurki, L., & Morris, N. (2001). The purposes, practices, and problems of
supermax prisons. Crime and Justice, 28, 385-424.
Labrecque, R. M. (2015a). Effect of solitary confinement on institutional
misconduct: A longitudinal evaluation. Cincinnati, OH: University of
Cincinnati. Retrieved from https://www.ncjrs.gov/App/Publications/abstract.
aspx?ID=271153.
Labrecque, R. M. (2015b, October). Who ends up in administrative segregation?
A meta-analytic review. Presented at the National Institute of Justice Topical
Working Group on the Use of Administrative Segregation in the United States,
Arlington, VA.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 157

Lovett, I. (2015, September 1). California agrees to overhaul use of solitary
confinement. The New York Times. Retrieved from http://www.nytimes.
com/2015/09/02/us/solitary-confinement-california-prisons.html.
Lyman, M. D. (1989). Gangland. Springfield, IL: Charles C Thomas.
Mears, D. P. (2005). Evaluating the effectiveness of supermax prisons. Washington,
DC: U.S. Department of Justice, Office of Justice Programs, National Institute
of Justice. Retrieved from https://www.ncjrs.gov/app/Publications/abstract.
aspx?ID=233437.
Mears, D. P., & Reisig, M. D. (2006). The theory and practice of supermax
prisons. Punishment & Society, 8(1), 33-57.
Mears, D. P., & Watson, J. (2006). Towards a fair and balanced assessment of
supermax prisons. Justice Quarterly, 23(2), 232-270.
Mohr, G. C., & Raemisch, R. (2015, February 13). Restrictive housing: Taking
the lead. Corrections Today, March/April. Retrieved from http://www.aca.org/
ACA_PROD_IMIS/Docs/Corrections%20Today/2015%20Articles/March%20
2015/Guest%20Editorial.pdf.
Mumola, C. J. (2005). Suicide and homicide in state prisons and local jails.
Washington, DC: U.S. Department of Justice, Office of Justice Programs, Bureau
of Statistics. Retrieved from http://www.prisonpolicy.org/scans/bjs/shsplj.pdf.
Naday, A., Freilich, J. D., & Mellow, J. (2008). The elusive data on supermax
confinement. The Prison Journal, 88(1), 69-93.
National Gang Intelligence Center. (2011). 2011 National Gang Threat
Assessment: Emerging trends. Washington, DC: Federal Bureau of Investigation,
National Gang Intelligence Center. Retrieved from https://www.fbi.gov/stats­
services/publications/2011-national-gang-threat-assessment/2011-national­
gang-threat-assessment-emerging-trends.
National Institute of Corrections. (1991). Management strategies in disturbances
and with gangs/disruptive groups. Washington, DC: U.S. Department of Justice.
National Institute of Corrections. (1997). Supermax housing: A survey of current
practice. Longmont, CO: U.S. Department of Justice, Office of Justice Programs,
National Institute of Corrections Information Center.
National Research Council (2008). Understanding crime trends: Workshop report.
A. S. Goldberger & R. Rosenfeld (Eds.). (Vol. 10). Washington, DC: The National
Academies Press. Retrieved from http://www.nap.edu/catalog/12472.

158 • National Institute of Justice | NIJ.gov

Noonan, M., & Ginder, S. (2013). Mortality in local jails and state prisons, 2000­
2011. Washington, DC: U.S. Department of Justice, Office of Justice Programs,
Bureau of Justice Statistics. Retrieved from https://www.ncjrs.gov/App/
Publications/abstract.aspx?ID=264348.
Ohio Department of Rehabilitation and Correction. (2014). DRC Security Threat
Groups Correctional Institution Inspection Committee. Columbus, OH: Ohio
Department of Corrections. Retrieved from http://www.ciic.state.oh.us/docs/
STG%20Brief%202014.pdf.
O’Keefe, M. L. (2005). Analysis of Colorado’s administrative segregation. Colorado
Springs, CO: Colorado Department of Corrections, Office of Planning and
Analysis. Retrieved from http://www.exdoc1.state.co.us/sites/default/files/opa/
AdSegReport.pdf.
Pinkerton, J. (2014). Prisons study ways to reduce solitary confinement.
Houston Chronicle. Retrieved from http://www.houstonchronicle.com/
news/houston-texas/houston/article/Prisons-study-ways-to-reduce-solitary­
confinement-5321117.php.
Porter, B. (1982). California prison gangs: The price of control. Corrections
Magazine, 8(6), 6-19.
Pyrooz, D. C., Decker, S. H., & Fleisher, M. S. (2011). From the street to the
prison, from the prison to the street: Understanding and responding to prison
gangs. Journal of Aggression, Conflict and Peace Research, 3(1), 12-24. http://
dx.doi.org/10.5042/jacpr.2011.0018.
Pyrooz, D. C., & Mitchell, M. M. (2015). Little gang research, big gang research.
In S. H. Decker & D. C. Pyrooz (Eds.), The handbook of gangs (pp. 28-58).
Chichester, West Sussex, England: Wiley-Blackwell.
Pyrooz, D. C., Sweeten, G., & Piquero, A. R. (2013). Continuity and
change in gang membership and gang embeddedness. Journal of Research
in Crime and Delinquency, 50(2), 239-271. http://dx.doi.org/10.5042/jac
pr.2011.001810.1177/0022427811434830.
Pyrooz, D. C., Turanovic, J. J., Decker, S. H., & Wu, J. (2016). Taking stock of the
relationship between gang membership and offending: A meta-analysis. Criminal
Justice & Behavior, 43(3), 365-397.
Ralph, P. H., & Marquart, J. W. (1991). Gang violence in Texas prisons. The Prison
Journal, 71(2), 38-49.
Reisig, M. D. (2002). Administrative control and inmate homicide. Homicide
Studies, 6(1), 84-103. http://dx.doi.org/10.1177/1088767902006001005.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 159

Reiter, K. A. (2012). Parole, snitch, or die: California’s supermax prisons and
prisoners, 1997-2007. Punishment & Society, 14(5), 530–563. http://dx.doi.
org/10.1177/1462474512464007.
Rodriguez, S. (2013). As 30,000 join California prison hunger strike, corrections
officials issue guidelines on starvation. Retrieved from http://solitarywatch.
com/2013/07/09/as-30000-join-california-prison-hunger-strike-corrections­
officials-issue-guidelines-on-starvation/.
Rodriguez, S. (2015). In California prisons, hundreds removed from solitary
confinement — and thousands remain. Retrieved from http://solitarywatch.
com/2015/01/27/in-california-hundreds-have-been-removed-from-solitary­
confinement-and-thousands-remain/.
Shaiq, Z. (2013). More restrictive than necessary: A policy review of secure
housing units. Hastings Race and Poverty Law Journal, 10, 327.
Shames, A., Wilcox, J., & Subramanian, R. (2015). Solitary confinement: Common
misconceptions and emerging safe alternatives. New York, NY: Vera Institute.
Retrieved from http://www.vera.org/pubs/solitary-confinement-misconceptions­
safe-alternatives.
Shelden, R. G. (1991). A comparison of gang members and non-gang members
in a prison setting. The Prison Journal, 71(2), 50-60.
Skarbek, D. (2014). The social order of the underworld: How prison gangs govern
the American penal system. Oxford University Press.
Smith, P. S. (2006). The effects of solitary confinement on prison inmates: A brief
history and review of the literature. Crime and Justice, 34(1), 441-528.
Sowa, J. L. (2012). Gods behind bars: Prison gangs, due process, and the first
amendment. Brooklyn Law Review, 77, 1593.
St. John, P. (2015). California’s solitary settlement. Los Angeles Times. Retrieved
January 15, 2016, from http://documents.latimes.com/californias-solitary­
settlement/.
Tachiki, S. N. (1995). Indeterminate sentences in supermax prisons based
upon alleged gang affiliations: A reexamination of procedural protection and
a proposal for greater procedural requirements. California Law Review, 83(4),
1115-1149.
Tapia, M. (2014). Texas Latino gangs and large urban jails: Intergenerational
conflicts and issues in management. Journal of Crime and Justice, 37(2), 256-274.
http://dx.doi.org/10.1080/0735648X.2013.768179.

160 • National Institute of Justice | NIJ.gov

Texas Department of Criminal Justice. (2014a). Administrative segregation.
Huntsville, TX: Texas Department of Criminal Justice. Retrieved from https://
assets.documentcloud.org/documents/2089812/texas-adseg-programming.pdf.
Texas Department of Criminal Justice. (2014b). Annual report for fiscal year 2014.
Huntsville, TX: Texas Department of Criminal Justice. Retrieved from https://
www.tdcj.state.tx.us/documents/Annual_Report_2014.pdf
Texas Department of Public Safety. (2015). Texas gang threat assessment: A state
intelligence estimate. Huntsville, TX: Texas Joint Crime Information Center,
Intelligence & Counterterrorism Division.
Toch, H. (2007). Sequestering gang members, burning witches, and subverting
due process. Criminal Justice and Behavior, 34(2), 274-288. http://dx.doi.
org/10.1177/0093854806296663.
Trulson, C. R., Marquart, J. W., & Kawucha, S. K. (2006). Gang suppression and
institutional control. Corrections Today, 68(2), 26.
Useem, B., & Piehl, A. M. (2008). Prison state: The challenge of mass incarceration.
New York: Cambridge University Press.
Useem, B., & Reisig, M. D. (1999). Collective action in prisons: Protests,
disturbances, and riots. Criminology, 37(4), 735-760. http://dx.doi.
org/10.1111/j.1745-9125.1999.tb00503.x.
Vigil, D. A. (2006). Classification and security threat group management.
Corrections Today, 68(2), 32-34.
Watkins, A. M., & Melde, C. (2016). Bad medicine: The relationship between
gang membership, depression, self-esteem, and suicidal behavior. Criminal Justice
and Behavior, 43(8), 1107-1126. http://dx.doi.org/10.1177/0093854816631797.
Wells, J. B., Minor, K. I., Angel, E., & Carter, L. (2002). A study of gangs and
security threat groups in America’s adult prisons and jails. Indianapolis, IN:
National Major Gang Task Force.
Winterdyk, J., & Ruddell, R. (2010). Managing prison gangs: Results from a
survey of U.S. prison systems. Journal of Criminal Justice, 38(4), 730-736. http://
dx.doi.org/10.1016/j.jcrimjus.2010.04.047.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 161

Appendix A. California Department of Corrections and
Rehabilitation criteria used to certify security threat
groups (STGs)
1. Criteria used as part of an official threat assessment conducted by the Office of
Correctional Safety to certify a group:
2. Information from any federal, state, county, or city correctional or law
enforcement agency, identifying the propensity for violence or the disruptive
nature of the potential STG being considered for certification.
3. Consideration with regard to whether the group meets the definition of an STG.
4. History of STG behavior in the community.
5. Evidence that the group presents a potential threat to the security of the
institution and safety of staff and offenders.
6. History of threatening behavior to staff or offenders safety involving acts
such as riots, group disturbances, possession or manufacture of weapons,
assault/battery, trafficking of narcotics, extortion, and coercion of other
individuals or groups.
7. Documentation of violent and illegal activities, which may also include
planning, organizing, threatening, financing, soliciting, or committing
unlawful acts.
8. Group evolution, structure, formalized procedures or bylaws, and
membership characteristics.
9. Information concerning group meetings and membership criteria.
10. Chronology of events or other information evidencing a threat to
institutional security or safety of staff and offenders through group activities,
associations, and potential security alignments.
11. Tattoo, symbol, and graffiti documentation.
12. Group association evidence, including offender and staff interviews.
13. Information concerning the group’s philosophy and affiliations.
Based on a review of the evidence available at intake of the overall severity
of the threat that the group poses, groups are either “certified” as STG-Is or
“recognized” as STG-IIs.

162 • National Institute of Justice | NIJ.gov

Appendix B. California Department of Corrections and
Rehabilitation source items and weights for validating
gang affiliates
The following, in descending order of significance, along with the points
associated with the source item in parentheses, is used by the California
Department of Corrections and Rehabilitation (CDCR, 2014) for security threat
group (STG) validation:
1. Legal documents (7 points) evidencing STG conduct, including court
transcripts, probation officer reports, or other legal documents.
2. Tattoos/body markings (6 points) related to STGs.
3. Offenses (6 points) committed for the “benefit or promotion of, at the
direction of, or in association with an STG.”
4. Self-admission (5 points), written or verbal, of STG involvement.
5. Communications (4 points), written or verbal, involving STG activity.
6. Visitors (4 points) with whom the offender engages in “promoting, furthering,
or assisting” STG activities.
7. Information from non-CDCR agencies (4 points) indicating STG activity,
such as police or crime reports.
8. Staff information (4 points) that reasonably indicates STG activity, including
observations, rule infractions, and misconduct.
9. Photographs (4 points) with STG connotations, such as insignia, symbols, or
gang members, taken in the past four years.
10. Written materials (4 points/2 points) “evidencing STG activity,” such
as membership or enemy lists, in personal possession or not in
personal possession.
11. Debriefing reports (3 points) that include references to specific STG-related
behavior.
12. Informants (3 points), whether or not confidential, that provide reliable and
specific knowledge “evidencing STG affiliation.”
13. Association (3 points) with validated STG members that constitutes a
“pattern or history of encounters,” not just by chance, involving STG behavior
or business.
14. Symbols (2 points) with STG connotations, such as hand signs, clothing, graffiti.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 16 3

All inmates with non-zero points are considered gang affiliates. To be
documented as a member, an offender must meet three independent source
items, the items must sum to a combined value of 10 or more, and there must
be a direct link to an STG affiliate (current or former member, or associate). The
same procedures apply to being documented as an associate, the only difference
being that associates are involved “periodically or regularly” with members or
other associates. An STG investigator makes this distinction. Finally, suspects
have scored at least two points of validation but have not reached the combined
total of points or necessary number of source items. There are also gang dropouts,
a distinction that elicits a different gang classification status.

CHAPTER 5

The Relationship Between
Inmate Misconduct, Institutional
Violence, and Administrative
Segregation: A Systematic
Review of the Evidence
Benjamin Steiner, Ph.D.
University of Nebraska, Omaha
Calli M. Cain, M.A.
University of Nebraska, Omaha

Introduction

M

aintaining safe and orderly institutions is a high priority for prison
and jail administrators, but institutional misconduct and violence
often threaten the safety and order of an institution (DiIulio, 1987;
Gendreau, Goggin, & Law, 1997; Steiner & Wooldredge, 2009; Steiner, Butler,
& Ellison, 2014; Toch, Adams, & Grant, 1989). Segregation — transferring an
inmate from the general prison population to an isolation cell in a more secure
area of the prison — is a common response to inmate misconduct or violence
that corrections officials use to regulate inmate behavior and promote order and

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 165

safety within their institutions (Browne, Cambier, & Agha, 2011; DiIulio, 1987;
Reassessing Solitary Confinement II, 2014). However, the practice has recently
drawn the attention of civil rights advocates, the U.S. Congress, and President
Obama, who want to ensure that it is used equitably and that it is effective as
both a deterrent and a consequence of prisoner misconduct and violence (e.g.,
Amnesty International, 2014; Obama, 2015; U.S. Senate Committee on the
Judiciary, Subcommittee on the Constitution, Civil Rights and Human Rights,
2014). Yet, despite the increased national attention on segregation and the
frequency with which it is used (Reassessing Solitary Confinement II, 2014;
Riveland, 1999), few researchers have examined the factors that influence who
is placed in segregation (e.g., inmates with a history of violence) or whether it is
effective in reducing subsequent problem behaviors.
This white paper reviews the evidence pertaining to the use of segregation and
its effects on offenders’ subsequent behavior, and provides critical information
about the evidence for researchers and practitioners. It also identifies gaps in the
extant research that should be useful for expanding the evidence base related to
the use of segregation and its behavioral effects. The paper begins by discussing
the extent of misconduct, violence, and the use of segregation in correctional
facilities. Next, it reviews the leading theories about predictors that indicate when
corrections officials use segregation and its effects on offender behavior. The
study methodology and its results follow. The paper concludes with a discussion
of the implications of the findings and identifies directions for future research
related to the relationship between inmate misconduct, institutional violence,
and segregation.

The Extent and Predictors of Inmate Misconduct and
Violence in U.S. Correctional Facilities
Inmate misconduct includes deviations from the formal rules that govern and
regulate an inmate’s behavior in a correctional institution. Violence in a prison
or jail can take several forms: intrapersonal (e.g., self-harm), interpersonal
(e.g., offending, victimization), or collective violence (e.g., riots). The Bureau of
Justice Statistics (BJS) routinely collects national-level data from inmates and
correctional administrators to assess the extent of misconduct and violence in
U.S. correctional facilities. Based on data from its most recent assessments — the
2005 Census of Adult State and Federal Correctional Facilities (BJS, 2010) and
the 2014 Annual Survey of Jails (BJS, 2015a) — correctional facilities typically
experience high rates of misconduct (about 949 incidents of misconduct per
1,000 prison inmates). Average rates of inmate-on-inmate assaults are also two to
three times higher than arrest rates for assaults among adults in the U.S. general
population (compare to Federal Bureau of Investigation, 2006, 2015).

16 6 • National Institute of Justice | NIJ.gov

Incidents of misconduct and assault are not distributed randomly across
facilities, however. In most prisons and jails, rates of misconduct and assault are
below the national average for these facilities. An assault against a staff member
occurs in only 30 percent of prisons and 5 percent of jails. Less than 6 percent
of prisons experience a major disturbance — an incident involving serious
injury, property damage, or loss of control of a portion of the facility, requiring
extraordinary measures to regain control.
Based on self-report data from the 2004 Survey of Inmates in State and Federal
Correctional Facilities (BJS, 2007) and the 2002 Survey of Inmates in Local Jails
(BJS, 2012a), about 50 percent of prison inmates and 16 percent of jail inmates
engage in misconduct during their incarceration. Twelve percent of prison
inmates and 3 percent of jail inmates physically assault another inmate, while
3 percent of prison inmates and less than 1 percent of jail inmates physically
assault a staff member. These data also show that 15 percent of prison inmates
and 7 percent of jail inmates are intentionally injured by another person during
their incarceration. Other researchers have found that from 7 percent to 20
percent of prison inmates report that they were violently victimized in the
past six months (Wolff, Blitz, Shi, & Bachman, 2007; Wooldredge, 1994, 1998;
Wooldredge & Steiner, 2014). In addition, data collected during the 2011-2012
National Inmate Survey (BJS, 2015b) show that approximately 4 percent of
prison inmates and 3 percent of jail inmates were sexually victimized in the
past year (Beck, Berzofsky, Caspar, & Krebs, 2013). Although these estimates
reflect the prevalence of events over different periods of time, it is worth noting
that less than 2 percent of the U.S. population age 12 and older experienced a
physical assault in the past year, and less than 1 percent were sexually assaulted
(Catalano, 2005, 2006; Rand, 2009).
Despite the tenuous nature of some of the comparisons made above, the rate
of violence and the likelihood of experiencing violent victimization appear
to be higher in correctional facilities than in the U.S. population.1 The data
also suggest that rates of misconduct and institutional violence vary across
facilities. Researchers have also found that the likelihood of engaging in
misconduct, perpetrating violence, and experiencing violent victimization varies
across inmates (e.g., Beck & Harrison, 2010; Camp, Gaes, Langan, & Saylor,
2003; Steiner, 2009; Steiner & Wooldredge, 2008a, 2008b; Wolff et al., 2007;
Wooldredge, 1998; Wooldredge & Steiner, 2014). This finding suggests that
misconduct and violence are not random processes. That is, differences between
inmates and their facility environments contribute to variation in institutional
misconduct and violence.

1

An exception to this conclusion is the likelihood of being a homicide victim. The U.S. homicide rate was 5.6
homicides per 100,000 persons in 2002 (Federal Bureau of Investigation, 2003), while the homicide rate in state
prisons was four homicides per 100,000 inmates (Mumola, 2005).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 167

Considerable research has focused on understanding the various individualand facility-level sources of institutional misconduct and violence. Findings
from two recent systematic reviews show that the background characteristics
of inmates, their institutional routines and experiences, and the characteristics
of the facilities in which they are confined are each relevant to an explanation
of misconduct and violence (Steiner et al., 2014; Steiner & Cain, in press). The
factors associated with an increased risk of misconduct and violence perpetration
are age (younger), a history of antisocial behavior (e.g., drug use, prior criminal
justice involvement), antisocial peers or gang involvement, and mental health
problems. Other factors include a history of victimization (except for sexual
victimization) prior to incarceration, a higher security risk classification, and
a sentence of five years or less. Inmates with low self-control and inmates from
disadvantaged neighborhoods are at higher risk for misconduct, while inmates
incarcerated for a sex offense are less at risk. An inmate’s gender (male) and
incarcerating offense (i.e., a property offense) also are risk factors for violence
perpetration (Steiner et al., 2014; Steiner & Cain, in press).
The institutional routines and experiences linked to a higher risk of misconduct
and violence perpetration include a history of misconduct and a greater amount
of time served, whereas involvement in prison work is associated with less risk
of misconduct and violence. Inmates involved in religion-based programs also
are at lower risk for misconduct, but inmates involved in a prison program are at
higher risk (Steiner et al., 2014; Steiner & Cain, in press).
Regarding the predictors that an inmate may become a victim of violence while
incarcerated, pre-incarceration employment, a higher security risk classification,
and a history of victimization prior to incarceration (both physical and sexual)
are associated with increased risk for victimization in prison and jail. Some
institutional routines and experiences can reduce an inmate’s risk. Inmates
involved in recreational activities and housed in the general population — and
who perceive the institution to be safer and view the institutional staff more
favorably — are at lower risk of becoming victims of violence. However, inmates
with a history of misconduct are generally at greater risk of victimization (Steiner
& Cain, in press).
Facilities that are at risk for higher rates of institutional misconduct and violence
(both perpetration and victimization) among inmates include higher-security
institutions, those with larger populations, and those with higher densities of black
inmates or inmates younger than age 25. Minimum-security facilities typically have
lower levels of misconduct and violence. Institutions with higher concentrations
of inmates incarcerated for violent offenses have lower misconduct rates. In
addition, facilities with a higher density of inmates classified as maximum security,
a higher percentage of officers who perceive the facility’s rules as underenforced,
and facilities in the southern region of the U.S. typically have higher levels of
institutional violence (Steiner et al., 2014; Steiner & Cain, in press).

168 • National Institute of Justice | NIJ.gov

Taken together, the findings from these two reviews of predictors of inmate
misconduct and institutional violence highlight that there are numerous inmate
background characteristics (e.g., age, criminal history), institutional routines and
experiences (e.g., history of misconduct, time served), and facility characteristics
(e.g., inmate population, higher security level) that explain variation in
institutional misconduct and violence. The relevant background characteristics
could be useful in risk- or needs-assessment instruments. Institutional routines
and experiences reflect inmates’ behavior after incarceration or interventions
by correctional officials. Facility characteristics represent either structural
characteristics or characteristics of the inmate or staff population (e.g.,
proportion of black inmates). Information regarding an institution’s structural
characteristics (e.g., facility security level) might be used by officials involved
in facility planning and design. Officials could use information about the
composition of inmate or staff populations to help them place particular types of
inmates or staff appropriately within a facility.

The Extent of Segregation Use in U.S. Correctional Facilities
Segregation and solitary confinement are terms that describe different types of
restrictive housing used in prisons and jails (Browne et al., 2011). Disciplinary
segregation, for instance, is typically a short-term punishment for inmates who
violate the institution’s rules (misconduct). Inmates are sometimes housed in
segregation prior to a hearing about a rule violation or while officials determine the
best placement for an inmate who is new to the facility (temporary confinement).
Inmates deemed a threat to an institution’s security are often placed in long­
term administrative segregation to remove them from the institution’s general
population. Protective custody is segregated housing used to keep inmates
considered at risk for victimization or self-harm in the general population. Finally,
supermax confinement is long-term, restrictive housing typically used for inmates
who are deemed a threat to other inmates or staff (Browne et al., 2011). Conditions
of confinement in segregated housing vary across states and across types of
restrictive housing, but often include placement in cells that are more sterile than
those of the general prison population. These conditions may include limited time
outside the cell (e.g., one hour per day), restricted social contact with other inmates
or staff, limited access to recreational or program opportunities, and restricted
opportunities for visitation or telephone calls (Baumgartel, Guilmette, Kalb, Li,
Nuni, Porter, & Resnik, 2015; Browne et al., 2011).
Results from a national survey of state departments of corrections conducted by
the Liman Program at Yale Law School, in collaboration with the Association
of State Correctional Administrators, revealed that 2.5 percent of the custodial
population within each state (range = 0.1 percent-7.5 percent) is typically held
in segregation for administrative reasons, not including protection or discipline,
whereas 6.6 percent of the custodial population within each state is held in some

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 169

form of segregation (range = 2.1 percent-14.2 percent) (Baumgartel et al., 2015).
Other studies have yielded estimates indicating that from 5 percent to 8 percent
of the state prison population is held in some form of segregation (Shames,
Wilcox, & Subramanian, 2015).
Data from the 2011-2012 National Inmate Survey (BJS, 2015b) show that nearly
20 percent of prison inmates and 18 percent of jail inmates have lived in some
form of segregated housing in the past year (or since coming to their current
facility, if less than one year ago). Approximately 10 percent of prison inmates
and 5 percent of jail inmates spend 30 days or longer in segregation; however,
there is significant variability in the use of segregation across facilities. Less than
10 percent of inmates spend time in segregation in 39.6 percent of prisons and
24.8 percent of jails, whereas more than 25 percent of inmates spend time in
segregation in 28.1 percent of prisons and 24.2 percent of jails (Beck, 2015).
The evidence suggests that prison officials use segregation with some frequency,
and that a notable percentage of all prison and jail inmates spend time in some
form of segregation during their incarceration. Just as the extent of misconduct
and institutional violence varies among facilities, the use of segregation also
varies across facilities. Several theories, discussed below, reveal why prison and
jail officials impose segregation more frequently in response to certain rule
violations, for particular inmates, or in particular facilities. An understanding
of these theories contributes to a more reasonable explanation of why particular
incident, inmate, and facility characteristics may influence officials’ use of
segregation. Such understanding also may help to curb disparate treatment of
inmate groups that may result from officials’ decision-making regarding the use
of segregation. In addition, two competing theories, described below, address the
effects of segregation on inmates’ subsequent behavior. They also highlight why
particular behavioral consequences may result from placement in segregation.

Theories About How Corrections Officials Use Segregation
Violations of the rules of conduct (misconduct) in a correctional facility are
subject to disciplinary procedures that determine guilt and impose punishment
(Howard, Winfree, Mays, Stohr, & Clason, 1994; Flanagan, 1982). Disciplinary
hearings are closed to the public. Inmates are afforded only minimal procedural
protections during the adjudicatory stage of the disciplinary process (Crouch,
1985). Few inmates succeed in challenging the allegations against them
(Flanagan, 1982; Thomas, Mika, Blakemore, & Aylward, 1991). The punishment
phase within prisons and jails, however, is similar to criminal sentencing.
Much like criminal court judges, prison and jail officials can impose a range
of sanctions (including segregation) once an inmate has been found guilty of
misconduct. The severity of the misconduct typically determines the range of
sanctions that can be imposed (Metcalf, Morgan, Oliker-Friedland, Resnik,

170 • National Institute of Justice | NIJ.gov

Spiegel, Tae, Work, & Holbrook, 2013). Given the similarities between the
punishment process within correctional facilities and criminal sentencing, it
seems reasonable to discuss the use of disciplinary segregation by considering
the theoretical and empirical literature on decision-making during the criminal
sentencing process.
Researchers often use two theories to explain why certain factors affect judicial
decision-making: causal attribution theory (e.g., Albonetti, 1991; Farrell &
Holmes, 1991; Hawkins, 1987) and focal concerns theory (Steffensmeier, Ulmer,
& Kramer, 1998). Both are also used to explain the sanctioning decisions of
other justice system officials during administrative proceedings (e.g., parole
revocation), where offenders have fewer rights than criminal defendants (e.g.,
Huebner & Bynum, 2006; Lin, Grattet, & Petersilia, 2010; Steiner, Travis,
Makarios, & Meade, 2011).

Causal Attribution Theory
Some scholars theorize that justice system officials base sanctioning decisions
on their beliefs regarding individuals’ prospects for reform (e.g., Albonetti,
1991; Bridges & Steen, 1998). Yet, these officials rarely have all the necessary
information regarding individuals’ risk of future criminality. To reduce the
uncertainty involved in these decisions, officials develop patterned responses
shaped by their preconceptions of the personal and environmental causes of
criminal behavior (Albonetti, 1991; Bridges & Steen, 1998; Hawkins, 1987), an
approach known as causal attribution theory. They may sanction individuals
more harshly if they perceive the individuals’ actions as more the result of
personal, rather than environmental, factors because officials consider these
individuals more culpable and, therefore, at higher risk of reoffending (Bridges
& Steen, 1998). For instance, the more misconduct — and sanctions received
for misconduct — in inmates’ pasts, the more culpable they seem because
they have demonstrated an unwillingness to respond to previous sanctions.
Similarly, the overrepresentation of minorities, men, and younger individuals
in the incarcerated population may lead correctional officials to perceive these
individuals as more culpable, leading to harsher sanctions for this population
(see Wooldredge, Griffin, & Rauschenberg, 2005, for a parallel argument
pertaining to judges).

Focal Concerns Theory
Butler and Steiner (in press) applied the focal concerns theory to disciplinary
decision-making in prison; this perspective recognizes that, although justice
system officials have an interest in controlling crime, they make sanctioning
decisions with limited information about offenders’ prospects for reform
(Steffensmeier et al., 1998; Ulmer & Johnson, 2004; Wooldredge et al., 2005).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 171

To reduce the uncertainty involved in these decisions, officials develop a perceptual
shorthand based on stereotypes that they associate with specific individuals and
cases (Johnson, 2006; Steffensmeier et al., 1998). However, the focal concerns
perspective further theorizes that three domains of reference guide officials: (1) an
offender’s blameworthiness, (2) an offender’s risk to the community, and (3) the
practical consequences of the punishment imposed for both the individual and the
justice system (Lin et al., 2010; Steffensmeier et al., 1998).
Blameworthiness is associated with the retributive philosophy of punishment,
such that an offender’s punishment corresponds directly to his or her culpability
for the crime and the degree of injury inflicted (Johnson, 2006; Steffensmeier &
Demuth, 2000; Steffensmeier et al., 1998). Blameworthiness might be reflected by
the severity of the rule violation and the inmate’s experience with the disciplinary
process. Offenses designated as more severe, for example, generally involve a
greater level of culpability and the potential for more significant harm to the
victim or the institution (e.g., assault, possession, or manufacture of dangerous
contraband) (Flanagan, 1982).
Justice system officials’ concerns about an offender’s risk to the community
are linked to the incapacitation and deterrent functions of punishment, and
involve predictions about future dangerousness based on attributions associated
with characteristics of the case and the offender (Steffensmeier et al., 1998).
The perceptions of prison and jail officials regarding which inmates are at risk
for subsequent misbehavior could be shaped by case characteristics such as
the nature of offense(s) for which inmates are convicted (e.g., violent versus
nonviolent) and individual characteristics such as their violation history,
criminal history, and social (e.g., gang involvement) and demographic (e.g., age)
background. For instance, corrections officials might consider inmates involved
in gangs to be at higher risk of reoffending (e.g., Bales & Miller, 2012; Griffin &
Hepburn, 2006).
Practical consequences and constraints associated with the organization and
individuals also influence officials’ sanctioning decisions (e.g., Johnson, 2006;
Steffensmeier et al., 1998; Ulmer & Johnson, 2004). Corrections officials must
maintain working relationships in an interdependent justice system (e.g., Dixon,
1995; Eisenstein, Flemming, & Nardulli, 1988; Johnson, 2006; Ulmer & Johnson,
2004), and they also understand the consequences of imposing segregation on
particular individuals (e.g., those with dependent children) (Daly, 1987; Griffin
& Wooldredge, 2006; Koons-Witt, 2002). For example, they may be less likely
to place inmates who are parents in segregation because its more restrictive
visitation policies would further disrupt any remaining bond between an
incarcerated parent and his or her children (see Cochran & Mears, 2013, for a
discussion of the effect of incarceration on the parent-child bond).
Although the perspectives discussed above apply to disciplinary proceedings
in prisons and jails, they may also be relevant to decision-making related to

172 • National Institute of Justice | NIJ.gov

long-term administrative segregation or supermax housing. Inmates are housed
in segregation based on the threat they pose to the general prison population.
However, these decisions are often made with limited information about the
actual threat that an inmate poses to the facility. Prison and jail officials may
manage their uncertainty by developing patterned responses — a perceptual
shorthand — based on their preconceptions of higher-risk inmates. The practical
consequences and constraints of a specific prison or jail bureaucracy may also
affect placement decisions.

Theories About the Effects of Time in Segregation on
Inmates’ Subsequent Behavior
Two opposing theories frame discussions about segregated housing: deterrence
theory and labeling/deviance amplification theory. Officials use punitive
segregation to punish inmates by isolating them and restricting privileges,
believing that time spent in segregation will deter inmates from engaging
in misbehavior when they return to the general prison population. Officials
expect administrative segregation to both incapacitate inmates and deter them
from subsequent misbehavior (Mears, 2013). However, some theorists would
predict that sanctions such as punitive or administrative segregation might
amplify inmates’ tendencies to misbehave (Lemert, 1951; Marx, 1981; Smith &
Paternoster, 1990; Wilkins, 1964).

Deterrence Theory
Segregation is, essentially, a prison or jail within a correctional institution (Browne
et al., 2011) where inmates are isolated in a single cell, unable to participate in
programming, and offered limited social and recreational opportunities (Pizzaro
& Stenius, 2004; Riveland, 1999). It is an unpleasant experience that should
deter inmates from reoffending (see Nagin, Cullen, & Jonson, 2009, for a parallel
discussion pertaining to imprisonment). Deterrence theorists emphasize the
importance of the certainty, celerity, and severity of sanctions such as segregation.
Individuals’ estimates of the certainty of sanctions (such as segregation) and
the severity of those sanctions form the basis for their calculation of the costs of
offending (Becker, 1968; Grasmick & Bursik, 1990; Nagin & Paternoster, 1993).
The celerity with which sanctions are applied affects individuals’ associations of
sanctions with related behavior (Taxman, Soule, & Gelb, 1999).
Formal policy in most prisons and jails mandates that sanctions be applied with
similar certainty and celerity, regardless of the type of rule violated or sanction
imposed (Metcalf et al., 2013). However, segregation is more severe than other
sanctions, and longer sentences in segregated housing are more severe than
shorter terms. Thus, imposing segregation and longer periods of segregation

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 173

are expected to deter inmates from misconduct after they return to the general
prison population (see Nagin et al., 2009, for a parallel argument regarding
imprisonment), and could also deter offenders from reoffending when they are
released from prison and return to their home communities (Mears, 2013).

Labeling/Deviance Amplification Theory
In contrast to deterrence theory, proponents of the labeling/deviance
amplification perspective (hereafter, deviance theory) would expect exposure
to segregation to increase offenders’ likelihood of future misbehavior. Inmates
often cannot participate in prison programming such as self-help programs
while in segregated housing and are often labeled troublesome or problematic
(King, Steiner, & Breach, 2008; Riveland, 1999). In turn, inmates may begin to
identify with this label, associate with other problematic inmates, and ultimately
engage in subsequent misconduct (secondary deviance). The official response to
inmates who have served time in segregation may also intensify, increasing the
likelihood of subsequent misconduct (Lemert, 1951; Marx, 1981; Wilkins, 1964).
Thus, sending an inmate to segregation could increase his or her subsequent
misbehavior in prison or jail.
Long-term consequences also may result from the labeling and isolation
experienced by inmates exposed to segregation (Haney, 2003; Haney, Weill,
Bakhshay, & Lockett, 2016; Smith, 2006; Toch, 2001). In fact, prolonged
isolation may result in increased anger, frustration, mental health problems,
and an adaptive response to isolated conditions, each of which could amplify
the likelihood that offenders will recidivate (Haney, 2003; Haney et al., 2016;
Kupers, 2008; Smith, 2006). Such effects also may be particularly evident among
offenders exposed to longer periods of segregation.

An Examination of Existing Studies
The preceding discussion underscores the finding that rates of violence are
higher in correctional institutions than in the general U.S. population, and that
corrections officials often use segregation to control violence and misconduct.
Causal attribution theory and focal concerns theory provide explanations for
why prison and jail officials may impose segregation for certain rule violations,
for particular inmates, or in certain facilities. However, evidence about the
applicability of these theories to disciplinary or administrative proceedings in
prisons and jails is limited. Officials who use segregation assume that it deters
inmates’ subsequent misbehavior in prison and jails and may reduce recidivism,
but labeling theory would suggest the opposite effects. Few studies have
investigated which (if either) of these theories is empirically supported.

174 • National Institute of Justice | NIJ.gov

To synthesize the existing literature on these topics and identify gaps in the
extant research into the use and behavioral effects of segregation, a systematic
review of the scientific evidence follows, focusing on four areas:
• Predictors of officials’ use of segregation in correctional institutions.
• Effects of segregation on offenders’ subsequent behavior (e.g., misconduct).
• Gaps in the research pertaining to the relationship between inmate
misconduct, institutional violence, and segregation.
• Directions for future research into the relationship between inmate
misconduct, institutional violence, and segregation.

Research Design and Method
As discussed above, segregation describes several types of restrictive housing
used in prisons and jails (Browne et al., 2011). This study focuses on the
following types of segregation: disciplinary segregation, long-term and
temporary administrative segregation, protective custody, and supermax
confinement. Distinctions between the types of confinement are made when
possible because their goals differ (Browne et al., 2011).
Keyword searches were used to find relevant studies in databases such as
the National Criminal Justice Reference Service, PsychInfo/PsychLit, and
Sociological Abstracts. The National Institute of Corrections and the National
Institute of Justice libraries were searched for technical reports that have not
been published in peer-reviewed journals or books. Finally, the references of the
relevant articles and reports were examined to determine whether there were
additional studies that should be included.
The review was restricted to studies of segregation administered in correctional
institutions for adults. The literature review was further restricted to only
those studies that focused on offenders’ subsequent behavior as it related to
the effects of segregation. The criteria excluded studies of segregation used
in juvenile facilities and those that examined the effects of segregation on
nonbehavioral outcomes, such as mental health. The review also excluded studies
that provided only a description of the population sent to segregation or the
extent of a particular behavior found among individuals exposed to segregation
(e.g., McGinnis, Austin, Becker, Fields, Lane, et al., 2014). Finally, the review
was restricted to studies published after 1974, when the U.S. Supreme Court’s
decision in Wolff v. McDonnell altered the disciplinary process in most state and
federal correctional facilities (Babcock, 1981).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 175

Findings
The review revealed 26 studies, 10 of which focused on the predictors of officials’
use of segregation; the remaining 16 studies examined its behavioral effects.
Among the studies of the predictors of officials’ use of segregation, eight focused
on the use of disciplinary segregation. Of the remaining two, one study examined
the predictors of placement in long-term administrative segregation; the other
focused on supermax confinement.
Three studies examined the effect of disciplinary segregation on subsequent inmate
behavior, and three others assessed the relationship between the rate at which
disciplinary segregation is used and rates of misconduct and violence. Seven
studies focused on the effects of supermax confinement: four examined the effects
of supermax confinement on offender behavior and three focused on the effects
of supermax confinement on levels of violence. Finally, three studies assessed the
effects of any type of administrative segregation on offender behavior.

Predictors of How Corrections Officials Use Segregation
Disciplinary segregation is used in correctional facilities to respond to
inmates who commit a violent act or chronically violate the institutions’ rules
(misbehave), whereas long-term administrative segregation is for inmates who
cannot live in the general prison population because they pose a threat to the
safety of other inmates, staff, and the institution (Browne et al., 2011; Riveland,
1999). Ten studies focused on the predictors for placement in segregation, but
many of these studies are dated, and their samples and analytical techniques
vary considerably. Although the findings are discussed here, the literature base
is not sufficient to determine whether the appropriate factors, such as violent
misconduct, affect corrections officials’ decisions to place inmates in segregation
or whether these officials administer segregation equitably.
Most researchers discovered that inmates found guilty of more serious offenses
and inmates with longer histories of misconduct are more likely to be placed
in segregation (e.g., Butler & Steiner, in press; Crouch, 1985; Flanagan, 1982;
Lindquist, 1980; Schafer, 1986), although Howard and colleagues (1994) did not
find such effects. Some researchers found that younger inmates are placed in
segregation more frequently than are older inmates (Butler & Steiner, in press;
Flanagan, 1982; Lindquist, 1980), but Howard and colleagues (1994) found a
nonsignificant age effect. Butler and Steiner (in press), Lindquist (1980), and
Stephan (1989) found that women are less likely to be placed in segregation than
are men, whereas McClellan (1994) found the opposite to be true — that women
are more likely to be placed in segregation. Howard and colleagues (1994)
found a nonsignificant gender effect. No studies of disciplinary segregation
have uncovered significant race effects (Butler & Steiner, in press; Crouch, 1985;
Flanagan, 1982; Howard et al., 1994; Lindquist, 1980; Stephan, 1989). No other

176 • National Institute of Justice | NIJ.gov

variables were examined frequently enough to draw meaningful conclusions. It is
also worth reiterating that there was considerable variability in the samples and
analytical techniques used across these studies. Howard and colleagues (1994),
for instance, examined cases processed in a federally operated prison, whereas
Butler and Steiner (in press) and others examined cases from state-operated
prisons (Crouch, 1985; Flanagan, 1982; Lindquist, 1980; McClellan, 1994;
Stephan, 1989. Some analyses involved only one control variable (Flanagan, 1982;
Lindquist, 1980; McClellan, 1994), whereas Butler and Steiner (in press), Crouch
(1985), and Howard and colleagues (1994) included at least 10 variables in their
multivariate analyses.
O’Keefe (2007) examined the factors associated with placement in long-term
administrative segregation in Colorado. She found that men, Hispanic inmates,
those incarcerated for a violent offense, inmates with a mental illness, members
of security threat groups, those with a higher number of disciplinary infractions,
and those often placed in disciplinary segregation were more likely to be placed
in administrative segregation. Neither inmates’ risk/need scores nor their
education levels predicted segregation placement.
When Mears and Bales (2010) assessed the predictors of placement in supermax
confinement in Florida, they found that certain inmates were more likely to be
confined: younger inmates, those incarcerated for a violent offense, those with a
higher number of prior violent convictions or a higher number of prior escape
convictions, those who had served more time, and those with a higher number
of disciplinary infractions for violent behavior, defiant behavior, or contraband.
Inmates’ race, ethnicity, and number of prior convictions and disciplinary
infractions for threat behavior had no effect on supermax placement.
Although it is difficult to draw substantive conclusions from the limited
number of studies available, a few general patterns emerge. First, regardless of
the type of segregation used, officials base their decisions on inmates’ criminal
histories and institutional misbehavior. Inmates convicted of more serious
offenses (e.g., violent), those with more significant prior records, and those
with a history of serious (e.g., violent) or chronic institutional rule violations
are more likely to be placed in segregation. Second, most studies found that
racial and ethnic minorities are no more likely to be placed in segregation
than are white inmates. Taken together, these findings suggest that corrections
officials use segregation for inmates with a history of serious or chronic
institutional misbehavior, or for inmates who pose a risk to the safety of an
institution. However, some extralegal factors (e.g., age, mental illness) also
affect corrections officials’ use of segregation.

Effects of Segregation on Offenders’ Subsequent Behavior
The 16 studies of the effects of segregation on offenders’ behavior differ
significantly by research design and in the participants and outcomes examined.
Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 177

Although a number of studies are methodologically strong, the paucity of
research on this topic prevents any determination of whether segregation deters,
amplifies, or has no effect on offenders’ subsequent misbehavior (e.g., violence).
Nonetheless, the relevant findings are reviewed here.
Wolff, Morgan, and Shi (2013) examined the effect of disciplinary segregation
on aggression among men and women in prisons in a single state who would
be released within the next 24 months. They found that the more days that men
spent in segregation, the more their aggression increased; however, there was no
significant effect on aggression in women. Labrecque (2015) assessed the effect of
disciplinary segregation and its duration on different types of misconduct (e.g.,
violent, drug) among Ohio inmates who had spent at least one year in prison and
served time in disciplinary segregation. He found that neither placement nor
length of time in segregation affected inmates’ subsequent misconduct. He also
found that gang members and inmates with mental health problems who were
placed in segregation — or spent more time there — were more likely to commit
subsequent misconduct and to do so more frequently. Morris (2016) found that
when Texas inmates who had served at least three years in prison were placed in
disciplinary segregation for violent misconduct, the placement had no effect on
subsequent violent misconduct or the amount of time to their next violent act.
Other scholars have examined the effect of an institution’s rate of disciplinary
segregation use on its rates of misconduct and violence (Huebner, 2003; Steiner,
2009; Wooldredge & Steiner, 2015). For the most part, these studies show that
using disciplinary segregation does not have a significant effect on inmates’ rates
of misconduct or violence.
Researchers who have examined the effect of supermax confinement on
recidivism have found that placement in supermax confinement has no effect
on recidivism (Lovell & Johnson, 2004; Lovell, Johnson, & Cain, 2007; Mears &
Bales, 2009; Butler & Steiner, in press). However, Lovell and Johnson (2004) and
Mears and Bales (2009) found that offenders placed in supermax confinement in
Washington and Florida had higher rates of violent recidivism than did offenders
who did not spend time in segregation. Offenders released from prison directly
from supermax confinement in Washington were more likely to reoffend than
were offenders not exposed to supermax confinement (Lovell et al., 2007), but
offenders in Florida were not (Mears & Bales, 2009).
Briggs, Sundt, and Castellano (2003) examined the effect of opening supermax
prisons in Arizona, Illinois, and Minnesota on rates of systemwide inmate­
on-inmate and inmate-on-staff assaults. They found that opening the new
facilities had no effect on rates of inmate assaults on other inmates. Briggs and
colleagues (2003) found that the rate of inmate assaults on staff did not change
in Minnesota, whereas Arizona experienced a temporary increase and Illinois
experienced a permanent decrease. A subsequent study by Sundt, Castellano,
and Briggs (2008) revealed no change in the rate of inmate-on-inmate assaults in
Illinois, but the rate of inmate-on-staff assaults decreased. Dye (2010) examined

178 • National Institute of Justice | NIJ.gov

data collected from state and privately operated prisons via the 2000 Census of
State and Federal Correctional Facilities. She found that supermax prisons had a
greater number of suicides than did minimum-security prisons.
Motiuk and Blanchette (2001) assessed the effect of placement in any type of
administrative segregation on recidivism among Canadian offenders. They found
that those with time in administrative segregation were more likely to return
to custody after they were released. They were also more likely than offenders
not placed in segregation to return for a new offense. Lanes (2009) examined
the effects of protective custody and long-term administrative segregation
on self-injuring among Michigan inmates. He found that protective custody
had no effect on inmates’ self-injuring behavior, but that inmates housed in
administrative segregation were more likely to self-harm. When Kaba and
colleagues (2014) assessed self-harm among inmates in a New York jail, they
found that the likelihood of self-harm, and potentially fatal self-harm, was higher
among inmates exposed to solitary confinement.
In conclusion, the limited evidence suggests that exposure to any type of
segregation does not affect offenders’ misbehavior while in correctional facilities
(e.g., misconduct), and has little to no effect on their behavior after release from
prison (e.g., recidivism). A greater use of segregation (e.g., disciplinary housing,
supermax housing) is not associated with reductions in facility or systemwide
misconduct and violence. However, exposure to segregation increases the risk of
self-harm among inmates, and supermax facilities experience more suicides than
do minimum-security facilities.

Discussion
Institutional safety is a high priority for correctional administrators, but
inmate misconduct and violence threaten the safety and order of an institution
(DiIulio, 1987; Gendreau et al., 1997; Steiner & Wooldredge, 2009). Segregation
is an intervention used by correctional administrators in response to inmate
misconduct or violence. Administrators also use segregation to isolate violenceprone or problematic inmates from an institution’s general population (Browne
et al., 2011; Riveland, 1999). Yet, few studies have examined the factors that affect
the use of segregation or whether it is effective in reducing offenders’ subsequent
problem behaviors (e.g., misconduct, violence). To synthesize the extant research
pertaining to these topics and provide direction for future research, the reviewers
systematically assessed the evidence pertaining to the use of segregation and to
the effects of segregation on offenders’ subsequent behavior.
The reviewers found 26 studies: 10 focused on the predictors of officials’ use of
segregation, and 16 examined segregation’s effects on offender behavior. These
studies reveal that, regardless of the type of segregation used (e.g., disciplinary,
administrative), corrections officials place inmates convicted of more serious

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 179

offenses (e.g., violent), those with more significant criminal records, and those
with more rule infractions in segregation more frequently than inmates with less
significant criminal histories and fewer institutional infractions. The evidence
also showed that inmates who are members of racial and ethnic minority
groups are no more likely than white inmates to be placed in segregation. Thus,
corrections officials are using segregation primarily for inmates with histories
of serious or chronic institutional misbehavior and those who pose a risk to the
safety or security of an institution.
It is important to reiterate that these conclusions are derived from only 10 studies
that focused on different types of segregation. Moreover, many of the studies are
dated (e.g., Crouch, 1985; Flanagan, 1982; Howard et al., 1994; Lindquist, 1980;
McClellan, 1994; Schafer, 1986; Stephan, 1989) and are limited to relatively small
samples of men or women processed in one or two facilities (e.g., Crouch, 1985;
Lindquist, 1980; McClellan, 1994; Schafer, 1986). Prison and jail populations
have increased dramatically over the past several decades, and the percentage of
minority and female inmates is notably higher today than in the 1970s, 1980s, or
1990s (Blumstein & Beck, 2005). Therefore, it may no longer be appropriate to
generalize results gleaned from earlier work to inmates processed in contemporary
correctional environments. Examination of data collected from small samples
from one or two facilities also limits the ability to generalize study results and has
prevented researchers from including adequate control variables in their statistical
models. Moreover, some of the samples examined were selected for other purposes
(e.g., behavior patterns of long-term inmates); therefore, they are not representative
of the inmates who violated prison rules in the respective facilities. Finally, no
studies have examined the predictors of officials’ use of segregation in jails; all 10
studies have focused on prison officials’ decision-making.
The primary implication of the findings from this review of the research on
the predictors of officials’ use of segregation is that there simply have not been
enough studies related to this topic to draw any meaningful conclusions. As
noted above, only 10 studies of the predictors of officials’ use of segregation have
been conducted in the past 40 years. Many studies are outdated, lack scientific
rigor, and have limited generalizability. An understanding of the relationship
between institutional misconduct, violence, and segregation requires an
evidence base of the factors that influence corrections officials’ decisions to use
segregation. Although segregation is designed for inmates who have a history
of serious or chronic institutional misbehavior, or inmates who pose a risk to
the safety or security of an institution (Browne et al., 2011; Riveland, 1999),
it is unclear whether segregation is being used for only these inmates. The
equitable use of segregation is of high importance to correctional administrators,
civil rights advocates, the U.S. Congress, and President Obama, but it is
unclear whether corrections officials use it equitably. If corrections officials
use segregation inappropriately, or if disparate treatment of offender groups
results from its use, some offender groups may endure a harsher confinement
experience than others. This unfair treatment can also undermine the legitimacy

180 • National Institute of Justice | NIJ.gov

of a prison organization, which could influence inmates’ willingness to defy
prison rules and those of other legal authorities (Liebling, 2004; Sparks, Bottoms,
& Hay, 1996; Useem & Kimball, 1989).
This review finds that, for the most part, placement in any type of segregation
does not affect offenders’ likelihood of institutional misconduct or their
recidivism after their release from prison. In addition, using segregation
at a higher rate or opening a supermax facility has little effect on rates of
misconduct and violence across facilities or state prison systems. However,
placement in segregation does increase inmates’ risk for self-harm (e.g.,
mutilation), and supermax facilities tend to have more suicides than do
minimum-security facilities.
It is also worth noting that the conclusions drawn here were based on only 16
studies. These studies focused on different types of segregation (e.g., disciplinary
segregation versus supermax confinement), different outcomes (e.g., misconduct
versus self-harm), and different levels of analysis (e.g., individual versus facility).
In addition, some of the samples were collected for other purposes (e.g., offenders
under post-release supervision), so they may not accurately represent the inmate
populations needed for these studies. Moreover, only one study examined the
effects of segregation on jail inmates’ behavior.
Once again, 16 studies across 40 years is too small an evidence base to inform
practice. In particular, these studies were conducted at different levels of analysis,
focused on different types of segregation, and examined different outcomes. An
understanding of the relationship between institutional misconduct, violence, and
segregation requires an evidence base related to the behavioral effects of exposure
to different types of segregation. Corrections officials use segregation to reduce
inmates’ problem behaviors (e.g., misconduct), but it is unclear whether it achieves
this objective. The frequency with which segregation is used in prisons and jails
— coupled with the desire of correctional administrators, civil rights advocates,
the U.S. Congress, and President Obama for it to be used effectively — requires a
knowledge base of the behavioral effects of the different types of segregation.
The limited research into the factors that affect corrections officials’ use of
segregation or the effects of segregation on offender behavior means that it is
difficult to draw conclusions regarding the empirical support for the theories
discussed above. Some study findings of corrections officials’ decision-making
regarding the use of segregation seem to support both causal attribution theory
and the focal concerns perspective discussed in this paper. Offenders with more
significant criminal and misconduct histories, for example, would be considered
more blameworthy and a greater risk to institutional safety. On the other hand,
the findings regarding the effect of an inmate’s race and ethnicity would not
support these perspectives. Findings pertaining to the nonsignificant effects of
segregation on subsequent misbehavior also do not support deterrence theory
or labeling/deviance amplification perspectives. There are simply too few studies
of either the use or behavioral effects of segregation to draw any meaningful
Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 181

conclusions about the applicability of these theories. However, these perspectives
may be useful for guiding future research.
Several limitations of this study should be noted. Each potential limitation stems
from differences between the studies reviewed herein. Some limitations of the
studies included in the review are compounded by a synthesis of the findings
from those studies. First, some researchers who examined the predictors of
corrections officials’ use of segregation included a large number (e.g., “more
than 10”) of variables in their final models, while others examined only the
relative effects of a few variables. Differences in the number of variables included
in models could contribute to study differences in scientific rigor (e.g., tenthorder partial correlation versus a third-order partial correlation). Second, some
researchers who studied the effects of segregation on offender behavior used
rigorous, quasi-experimental designs to match segregated inmates to inmates
not exposed to segregation, while other researchers relied on multivariate
regression analysis to control for observed differences between these two groups.
Differences in the scientific rigor between these two types of studies may have
influenced the findings. Finally, the findings were derived from the results of
studies that involved different sample sizes and, occasionally, samples of different
subpopulations (e.g., men) of the relevant inmate population (jail or prison).
However, the findings from each study were evaluated equally in this review.
Setting aside the limitations discussed above, this white paper is the most
comprehensive review to date of the evidence related to the use and behavioral
effects of segregation. Although the review uncovered some consistent patterns
across the results of the studies included here, the data are not sufficient to
place much faith in the findings. To be sure, the most important finding is the
dearth of rigorous scientific research on these topics. The gaps in the research
pertaining to the use of segregation and its effects on offender behavior are clear;
additional research should be devoted to each of these topics. An understanding
of the link between institutional misconduct, violence, and segregation requires
knowledge of the types of offenses (e.g., violent) and offenders (e.g., those with
more significant criminal histories) that influence corrections officials’ decisions
regarding their use of segregation, and whether segregation affects inmates’
likelihood of perpetrating subsequent misconduct or violence. Once an evidence
base related to the predictors of the use and behavioral effects of segregation
is developed, future studies might pursue more nuanced topics related to the
practice. Some avenues for future research are discussed below.

Conclusions and Directions for Future Research
Prison and jail officials are responsible for protecting society from the
incarcerated population, but they are also legally responsible for protecting the
individuals confined in their institutions (Park, 2000). Incarcerating people in
prisons and jails, however, essentially gathers together people and property that

182 • National Institute of Justice | NIJ.gov

are potential targets and persons who are likely to offend (Wooldredge & Steiner,
2014). Indeed, rates of violence and violent victimization are higher in correctional
facilities than in the general U.S. population, and corrections officials often
respond to this violence and misconduct (or the threat of it) by using segregation
to maintain order and safety. At any time, between 5 percent and 8 percent of the
state prison population is held in some form of segregation (Baumgartel et al.,
2015; Shames et al., 2015). At least 20 percent of prison inmates and 18 percent of
jail inmates spend time in segregation during their incarceration (Beck, 2015), yet
scholars know very little about the incidents, inmates, and facility characteristics
that affect the use of segregation or whether segregation is effective in reducing
problem behaviors (e.g., misconduct, recidivism).
This systematic review of the evidence found that, regardless of the type of
segregation designated (e.g., disciplinary, administrative), corrections officials
use segregation primarily for inmates with a history of serious (e.g., violent) or
chronic institutional misconduct or those who pose a risk to the institutions’
safety or security. The researchers also found evidence that placement in
segregation has no effect on offenders’ subsequent misbehavior (e.g., misconduct,
recidivism), but that offenders placed in segregation are more at risk for selfharm. These findings should be interpreted with caution, however, given the
relatively small number of studies conducted on these topics. The researchers also
observed numerous limitations of these studies and between-study differences in
the research methods used, which further lessens any confidence in the results.
However, the review did uncover some important gaps in the evidence base from
which the following questions were identified for future research.

1. What are the predictors of corrections officials’ use of segregation?
Over the past 40 years, only 10 studies have examined the factors that influence
corrections officials’ decisions to use segregation; eight of those studies focused
on disciplinary segregation. Most of this research is dated and has limited
generalizability.2 This review finds a clear need for research into the factors that
influence corrections officials’ decision-making related to the use of segregation.
The equitable use of segregation is of high importance to correctional
administrators, civil rights advocates, the U.S. Congress, and President Obama,

2

Prison and jail populations have increased dramatically over the past several decades (Blumstein & Beck,
2005; Western, 2006), so it may no longer be appropriate to generalize the results to inmates processed in
contemporary prison and jail environments. Many of the studies also involved analysis of data collected from
small samples drawn from one or two facilities, which limits the generalizability of study results and prevents
researchers from including adequate control variables in their statistical models or from examining potential
facility-level effects. Furthermore, some samples were selected for other purposes (e.g., the examination of
behavior patterns of long-term inmates), so they do not represent the population of inmates who violated prison
rules in the respective facilities.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 183

but it is unclear whether corrections officials use segregation equitably. For
instance, segregation is designed for inmates who have a history of serious or
chronic institutional misbehavior or inmates who pose a risk to institutional
safety (Browne et al., 2011; Riveland, 1999), but there is not enough evidence to
determine whether segregation is being reserved for these inmates. If corrections
officials use segregation improperly, or if disparate treatment of offender groups
results from its use, then some inmate groups may endure a harsher confinement
experience than will others. This unfair or disparate treatment can also
undermine the legitimacy of a correctional organization, which could influence
inmates’ willingness to defy institutional rules and those of other legal authorities
(Liebling, 2004; Sparks et al., 1996; Useem & Kimball, 1989).
Moving forward, researchers should focus on the factors that affect corrections
officials’ use of each type of segregation: disciplinary segregation, long-term
administrative segregation, protective custody, temporary confinement, and
supermax confinement. Researchers should also collect reasonably sized,
representative samples of inmates who violate the rules of a corrections facility.
Existing research should guide the selection of predictor variables to be
considered in future studies, but given the parallels between the disciplinary
process in correctional facilities and the criminal sentencing process, researchers
might also draw on the theories discussed above, which have often been
applied to criminal sentencing (e.g., focal concerns theory). Furthermore, these
perspectives have already been used to explain sanctioning decisions made by
other justice system officials during administrative proceedings (e.g., parole
revocation) where offenders enjoy fewer rights than criminal defendants,
including prison disciplinary proceedings (e.g., Butler & Steiner, in press;
Huebner & Bynum, 2006; Lin et al., 2010; Steiner et al., 2011). Applying causal
attribution theory or the focal concerns perspective to the punishment process
in correctional institutions permits an assessment of whether these theories are
general to decision-making related to criminal punishment. The theories also
contribute to a more reasonable explanation of why particular incidents, inmates,
and facility characteristics may influence officials’ use of segregation. Such an
understanding could lead to strategies designed to curb disparate treatment of
inmate groups that may result from officials’ decision-making regarding the use
of segregation.
Finally, researchers might consider conducting multisite studies to increase the
generalizability of their results and to permit analysis of facility-level effects,
but only if inmates from enough facilities are included. Research on criminal
courts has uncovered various sentencing outcomes across court organizations
(Eisenstein et al., 1988; Ulmer, 2012; Ulmer & Johnson, 2004; Wang & Mears,
2010), and similar variation in case outcomes may exist across prisons (Butler &
Steiner, in press).

18 4 • National Institute of Justice | NIJ.gov

2. Do an inmate’s gender, race, or ethnicity moderate the effects of
predictors of corrections officials’ use of segregation?
Prison and jail populations have increased dramatically over the past several
decades. The percentage of racial and ethnic minorities and female inmates is
notably higher today than in the 1970s, 1980s, or 1990s (Blumstein & Beck, 2005;
Western, 2006). Yet, no studies have examined the moderating effects of gender,
race, and ethnicity on the influences of corrections officials’ use of segregation,
even though such examinations are relatively common in studies of judicial
sentencing decisions (e.g., Baumer, 2013; Brennan, 2006; Brennan & Spohn,
2009; Griffin & Wooldredge, 2006; Steffensmeier & Demuth, 2006; Wooldredge
et al., 2005). There may be race, ethnic, or gender group differences in the
effects of particular case (e.g., violent offense) or individual characteristics (e.g.,
misconduct history) on officials’ decisions to use segregation. Indeed, researchers
of criminal sentencing outcomes have uncovered these hidden disparities (e.g.,
Griffin & Wooldredge, 2006; Steffensmeier & Demuth, 2006; Wooldredge et al.,
2005). However, evidence that an inmate’s gender, race, and ethnicity can have a
moderating effect on corrections officials’ use of segregation is nonexistent.
The equitable use of segregation is of high importance to correctional
administrators, civil rights advocates, the U.S. Congress, and President Obama.
An examination of whether an inmate’s gender, race, or ethnicity moderates the
effects of the predictors of segregation use would shed light on possible hidden
disparities in corrections officials’ use of segregation. Determining whether
these disparities exist is important for understanding the relationship between
institutional misconduct, violence, and segregation. For instance, if committing
a violent misconduct is associated with placement in segregation, then this
relationship should be similar in magnitude for men and women, and across
different racial and ethnic groups. If the magnitude of the violent misconductsegregation relationship differs across these groups, then there are unwanted
disparities in the treatment of some inmate groups. Such unfair or disparate
treatment of inmate groups can undermine the legitimacy of a correctional
organization, which could influence inmates’ willingness to defy institutional
rules and those of other authorities (Sparks et al., 1996; Useem & Kimball 1989).

3. What are the effects of segregation on inmate behavior?
Over the past four decades, only 16 studies were conducted regarding the effects
of segregation on inmate behavior. They examined the effects of different types
of segregation (e.g., disciplinary, long-term administrative segregation), different
outcomes (e.g., recidivism, self-harm), and different levels of analysis (e.g.,
individual, state). In some, the samples were collected for other purposes, so they
do not accurately represent the inmate populations needed for these studies.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 185

Institutional misconduct and violence threaten the safety and order of a
correctional institution. Corrections officials frequently respond to these
behaviors by placing inmates in segregation. A notable percentage of inmates are
held in segregation at any given time, and a significant number of inmates spend
some time in segregation during incarceration (Baumgartel et al., 2015; Beck,
2015; Shames et al., 2015); yet, few studies have assessed whether segregation
is effective in reducing offenders’ problem behaviors. There is a clear need for
studies that examine the effects of different types of segregation (e.g., disciplinary
segregation, long-term administrative segregation) on offender behavior.
An experimental design for such a study is not possible for ethical reasons —
inmates cannot be randomly assigned to segregation (Mears, 2008, 2013; O’Keefe,
2008). Instead, researchers should examine large samples of inmates who are
eligible for segregation, and should consider using rigorous quasi-experimental
designs that involve matching techniques (e.g., propensity score matching) to
estimate the effects of segregation on offender behavior. Researchers should also
focus on different behavioral outcomes, including those that tap subsequent
criminality (e.g., misconduct, recidivism), along with other measures of
institutional violence (e.g., self-harm).
An understanding of the behavioral effects of segregation would help to determine
whether the theories discussed above (e.g., deterrence) could be applied to its use
in prisons and jails. Knowledge about whether segregation — or different lengths
of segregation — will deter or amplify an offender’s problem behaviors (e.g.,
misconduct, violence) would be immediately useful to corrections officials who
rely on segregation to regulate inmate behavior in their facilities. If segregation
is effective in curbing inmate misconduct, for example, then corrections officials
might use it more often. Similarly, if five days in segregation has the same
misconduct-reducing effect as 10 days, then corrections officials might reduce the
time that inmates spend in segregation. These are critical questions for both theory
and practice, but the current evidence base is insufficient for drawing conclusions.

4. Are there inmate and facility characteristics that moderate the
effects of segregation on inmate behavior?
Labrecque (2015) observed that the effects of disciplinary segregation and length
of time spent there were affected by inmate mental health and gang involvement.
That is, inmates with mental health problems and inmates involved with
gangs who were placed in segregation, or spent longer periods in segregation,
were more likely to commit subsequent misconduct and to do so with greater
frequency. Other studies of possible moderating effects are nonexistent, however.
Researchers who have examined the effects of criminal justice sanctions (e.g.,
arrest, prosecution) on recidivism have discovered that sanctions are more
effective with individuals who are good risks or those who have a greater stake
in conforming (e.g., Berk, Campbell, Klap, & Western, 1992; Dejong, 1997;

186 • National Institute of Justice | NIJ.gov

Sherman, Smith, Schmidt, & Rogan, 1992; Thistlethwaite, Wooldredge, & Gibbs,
1998; Wooldredge & Thistlethwaite, 2002). It seems reasonable, therefore, that
particular groups of inmates (in addition to those with mental health problems
and gang involvement) may be at higher risk for subsequent misbehavior or
maladjustment from being placed in segregation. However, inmates with other
characteristics may have a lower risk of misbehavior or maladjustment after time
in segregation. In addition, a multisite study could uncover facility characteristics
that influence the effects of segregation on inmate behavior. All of these avenues
are fruitful areas for future inquiry. Researchers might begin by examining the
potential moderating effects of the causes and correlates of inmate misconduct
and violence identified by the recent systematic reviews of the literature
discussed above (Steiner et al., 2014; Steiner & Cain, in press).
Information pertaining to differential effects of segregation on inmate behavior
could be important for informing both theory and practice. From a theoretical
perspective, a finding that inmate or facility characteristics moderate the
behavioral effects of segregation would align with researchers’ findings related to
conditional deterrent effects in the broader criminological literature
(e.g., Berk et al., 1992; Dejong, 1997; Sherman et al., 1992; Thistlethwaite et
al., 1998; Wooldredge & Thistlethwaite, 2002). Practitioners could use this
information to determine which inmates might be more suitable for segregation,
rather than an alternative sanction. The importance of institutional safety to
corrections officials underscores the need for research that not only identifies
whether sanctions such as segregation are effective in curbing misconduct and
violence, but also whether these sanctions are more effective or less effective for
particular inmates or in particular facility environments.

5. What are the predictors of officials’ use of segregation in jails, and
of the effects of jail segregation on individuals’ subsequent behavior?
Although each of the questions for future research identified above applies to
prisons, each also applies to jails, and the need for this research in jails merits
additional discussion (see also Haney et al., 2016). This is not to say that research
in jails should take priority over research in prisons. It is only to reiterate that
the findings from this review suggest that research into segregation use in jails
is considerably more scarce than similar research into its use in prisons. In fact,
only one study included in this review focused on jail inmates.
Approximately one-third of the U.S. incarcerated population is confined in
local jails. The volume of people who enter jails each year is far greater than
the number who enter prisons (Glaze & Kaeble, 2014). Nearly one-fifth of jail
inmates spend time in some form of restrictive housing during incarceration
(Beck, 2015); yet, this review uncovered no studies of the predictors of officials’
use of segregation in jails and only one study of the behavioral effects of jail
segregation. To paraphrase Haney and colleagues (2016:131), segregated housing

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 187

in jails is among the least-studied components of the criminal justice system,
even though jails affect more people per year than do prisons and segregation
is experienced by a significant percentage of people incarcerated in jails. More
research on this topic is clearly needed. Study findings pertaining to the use of
segregation in jails and the effects of jail segregation on individuals’ subsequent
behavior (e.g., misconduct, violence) could provide useful insights for
practitioners working within jails.
In conclusion, institutional misconduct and violence threaten the safety and
order of a correctional facility, making both behaviors high priorities for
correctional administrators (DiIulio, 1987; Gendreau et al., 1997; Steiner &
Wooldredge, 2009; Steiner et al., 2014; Toch & Grant, 1989). Corrections officials
often use segregation in response to inmate misconduct and violence, expecting
the experience to regulate inmate behavior and promote order and safety
within their institutions (Browne et al., 2011; DiIulio, 1987; Reassessing Solitary
Confinement II, 2014). If segregation is used in response to serious (e.g., violent)
or chronic institutional misconduct and is effective in curbing institutional
misconduct and violence, then there is a logical relationship when an inmate’s
misconduct and violence (or a propensity to engage in these behaviors) initiates
the use of segregation, which in turn reduces these problem behaviors. If other
factors (e.g., age, mental health) initiate the use of segregation or exposure to
segregation amplifies an inmate’s risk for perpetrating misconduct or violence,
then the relationship is less straightforward.
This systematic review of the literature finds that corrections officials use
segregation primarily for inmates who have a history of serious (e.g., violent) or
chronic institutional misconduct, or for those who pose a risk to the safety or
security of an institution. It also reveals that exposure to segregation has no effect
on offenders’ subsequent misbehavior (e.g., misconduct, violence, recidivism),
but does increase their risk for self-harm. It is also clear that very few studies
have examined the factors that affect the use of segregation or its effectiveness in
reducing inmates’ problem behaviors (e.g., misconduct). These are clear gaps in
the knowledge base related to the relationship between institutional misconduct,
violence, and segregation. This review identifies important questions for future
research in light of the emphasis that correctional administrators place on
institutional safety (Gendreau et al., 1997; Steiner & Wooldredge, 2009; Steiner
et al., 2014), how frequently segregation is used to promote safety (Beck, 2015),
and how much additional cost is involved in housing inmates in segregation
relative to housing the general prison population (e.g., American Civil Liberties
Union of Texas, 2015; Briggs et al., 2003; Butler, Steiner, Makarios, & Travis, in
press; Metcalf, Resnik, & Quattlebaum, 2015; Shames et al., 2015). The equitable
and effective use of segregation is also a high priority for civil rights advocates,
the U.S. Congress, and President Obama (e.g., Amnesty International, 2014;
Reassessing Solitary Confinement II, 2014; Obama, 2015; U.S. Senate Committee
on the Judiciary, Subcommittee on the Constitution, Civil Rights and Human
Rights, 2014), and so the need for this research is great.

188 • National Institute of Justice | NIJ.gov

References
Albonetti, C. A. (1991). An integration of theories to explain judicial discretion.
Social Problems, 38(2), 247-266.
American Civil Liberties Union of Texas. (2015, February). A solitary failure: The
waste, cost and harm of solitary confinement in Texas.
Amnesty International. (2014). Entombed: Isolation in the U.S. federal prison system.
Babcock, W. (1981). Due process in prison disciplinary proceedings. Boston
College Law Review, 22(5), 1009-1096.
Bales, W. D., & Miller, C. H. (2012). The impact of determinate sentencing on
prisoner misconduct. Journal of Criminal Justice, 40(5), 394-403.
Baumer, E. P. (2013). Reassessing and redirecting research on race and
sentencing. Justice Quarterly, 30(2), 231-261.
Baumgartel, S., Guilmette, C., Kalb, J., Li, D., Nuni, J., Porter, D., Resnik, J. (2015).
Time-in-cell: The ASCA-Liman 2014 national survey of administrative segregation
in prison. Public Law Research Paper No. 552. Yale Law School 1(3). Available at
http://ssrn.com/abstract=2655627 or http://dx.doi.org/10.2139/ssrn.2655627.
Beck, A. J. (2015). Use of restrictive housing in U.S. prisons and jails, 2011–12.
NCJ 249209. Washington, DC: Bureau of Justice Statistics.
Beck, A., Berzofsky, M., Caspar, R., Krebs, C. (2013). Sexual victimization in
prisons and jails reported by inmates, 2011-12. NCJ 241399. Washington, DC:
Bureau of Justice Statistics.
Beck, A., & Harrison, P. M. (2010). Sexual victimization in prisons and jails reported
by inmates, 2008-09. NCJ 231169. Washington, DC: Bureau of Justice Statistics.
Beck, A., & Johnson, C. (2012). Sexual victimization reported by former state
inmates, 2008. NCJ 237363. Washington, DC: Bureau of Justice Statistics.
Becker, H. S. (1968). On labeling outsiders. Deviance: The Interactionist
Perspective, 13-17.
Berk, R. A., Campbell, A., Klap, R., & Western, B. (1992). The deterrent effect
of arrest in incidents of domestic violence: A Bayesian analysis of four field
experiments. American Sociological Review, 57(5), 698-708.
Blumstein, A., & Beck, A. J. (2005). Reentry as a transient state between liberty
and recommitment, 56. New York, NY: Cambridge University Press.
Brennan, P. K. (2006). Sentencing female misdemeanants: An examination of the
direct and indirect effects of race/ethnicity. Justice Quarterly, 23(1), 60-95.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 189

Brennan, P. K. (2009). The joint effects of offender race/ethnicity and sex on
sentencing outcomes. In M. Krohn, A. Lizotte, & G. Penly Hall (Eds.), Handbook
on crime and deviance (pp. 319-347). New York, NY: Springer.
Brennan, P. K., & Spohn, C. (2009). The joint effects of offender race/ethnicity
and sex on sentence length decisions in federal courts. Race and social problems,
1(4), 200-217.
Bridges, G. S., & Steen, S. (1998). Racial disparities in official assessments of
juvenile offenders: Attributional stereotypes as mediating mechanisms. American
Sociological Review, 63(4), 554-570.
Briggs, C. S., Sundt, J. L., & Castellano, T. C. (2003). The effect of supermaximum
security prisons on aggregate levels of institutional violence. Criminology, 41(4),
1341-1376.
Browne, A., Cambier, A., & Agha, S. (2011). Prisons within prisons: The use of
segregation in the United States. Federal Sentencing Reporter, 24(1), 46-49.
Bureau of Justice Statistics. (2007). Survey of inmates in state and federal
correctional facilities, 2004, ICPSR-4572. Washington, DC: U.S. Department of
Justice. Ann Arbor, MI: Inter-University Consortium for Political and Social
Research (distributor).
Bureau of Justice Statistics. (2010). Census of adult state and federal correctional
facilities, 2005, ICPSR-24642-v2. Washington, DC: U.S. Department of Justice.
Ann Arbor, MI: Inter-University Consortium for Political and Social Research
(distributor).
Bureau of Justice Statistics. (2012a). Survey of inmates in local jails, 2002, ICPSR­
04359-v2. Washington, DC: U.S. Department of Justice. Ann Arbor, MI: InterUniversity Consortium for Political and Social Research (distributor).
Bureau of Justice Statistics. (2015a). Annual survey of jails, 2014, ICPSR-36274-v1.
Washington, DC: U.S. Department of Justice. Ann Arbor, MI: Inter-University
Consortium for Political and Social Research (distributor).
Bureau of Justice Statistics. (2015b). National inmate survey, 2011-2012, ICPSR­
35009-v1. Washington, DC: U.S. Department of Justice. Ann Arbor, MI: InterUniversity Consortium for Political and Social Research (distributor).
Butler, H. D., & Steiner, B. (in press). Examining the use of disciplinary
segregation within and across prisons. Justice Quarterly.
Butler, H. D., Steiner, B., Makarios, M., & Travis, L. (in press). Assessing the
effects of exposure to supermax confinement on offender post-release behaviors.
The Prison Journal.

19 0 • National Institute of Justice | NIJ.gov

Camp, S. D., Gaes, G. G., Langan, N. P., & Saylor, W. G. (2003). The influence
of prisons on inmate misconduct: A multilevel investigation. Justice Quarterly,
20(3), 501-533.
Catalano, S. (2005). Criminal victimization, 2004. Special report. NCJ 210674.
Washington, DC: Bureau of Justice Statistics.
Catalano, S. (2006). Criminal victimization, 2005. Special report. NCJ 214644.
Washington, DC: Bureau of Justice Statistics.
Cochran, J. C., & Mears, D. P. (2013). Social isolation and inmate behavior: A
conceptual framework for theorizing prison visitation and guiding and assessing
research. Journal of Criminal Justice, 41(4), 252-261.
Crouch, B. M. (1985). The significance of minority status to discipline severity in
prison. Sociological Focus, 18(3), 221-233.
Daly, K. (1987). Discrimination in the criminal courts: Family, gender, and the
problem of equal treatment. Social Forces, 66(1), 152-175.
DeJong, C. (1997). Survival analysis and specific deterrence: Integrating
theoretical and empirical models of recidivism. Criminology, 35(4), 561-576.
DiIulio, Jr., J. J. (1987). Governing prisons: A comparative study of correctional
management. New York, NY: The Free Press.
Dixon, J. (1995). The organizational context of criminal sentencing. American
Journal of Sociology, 100(5), 1157-1198.
Dye, M. H. (2010). Deprivation, importation, and prison suicide: Combined
effects of institutional conditions and inmate composition. Journal of Criminal
Justice, 38(4), 796-806.
Eisenstein, J., Flemming, R. B., & Nardulli, P. F. (1988). The contours of justice:
Communities and their courts. Boston, MA: Little, Brown.
Farrell, R. A., & Holmes, M. D. (1991). The social and cognitive structure of legal
decision-making. The Sociological Quarterly, 32(4), 529-542.
Federal Bureau of Investigation. (2003). Crime in the United States, 2002.
Federal Bureau of Investigation. (2006). Crime in the United States, 2005.
Federal Bureau of Investigation. (2015). Crime in the United States, 2014.
Flanagan, T. J. (1982). Discretion in the prison justice system: A study of
sentencing in institutional disciplinary proceedings. Journal of Research in Crime
and Delinquency, 19(2), 216-237.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 191

Glaze, L. E., & Kaeble, D. (2014). Correctional populations in the United States,
2013. NCJ 248479. Washington, DC: Bureau of Justice Statistics.
Grasmick, H. G., & Bursik Jr., R. J. (1990). Conscience, significant others, and
rational choice: Extending the deterrence model. Law and Society Review, 24(3),
837-861.
Gendreau, P., Goggin, C. E., & Law, M. A. (1997). Predicting prison misconducts.
Criminal Justice and Behavior, 24(4), 414-431. doi: 10.1177/0093854897024004002.
Griffin, M. L., & Hepburn, J. R. (2006). The effect of gang affiliation on violent
misconduct among inmates during the early years of confinement. Criminal
Justice and Behavior, 33(4), 419-466.
Griffin, T., & Wooldredge, J. (2006). Sex-based disparities in felony dispositions
before versus after sentencing reform in Ohio. Criminology, 44(4), 893-923.
Hawkins, D. F. (1987). Beyond anomalies: Rethinking the conflict perspective on
race and criminal punishment. Social Forces, 65(3), 719-745.
Haney, C. (2003). Mental health issues in long-term solitary and “supermax”
confinement. Crime and Delinquency, 49(1), 124-156.
Haney, C., Weill, J., Bakhshay, S., & Lockett, T. (2016). Examining jail isolation: What
we don’t know can be profoundly harmful. The Prison Journal, 96(1), 126-152.
Howard, C., Winfree, L. T., Mays, G. L., Stohr, M. K., & Clason, D. L. (1994).
Processing inmate disciplinary infractions in a federal correctional institution:
Legal and extralegal correlates of prison-based legal decisions. The Prison
Journal, 74(1), 5-31.
Huebner, B. M. (2003). Administrative determinants of inmate violence: A
multilevel analysis. Journal of Criminal Justice, 31(2), 107-117.
Huebner, B. M., & Bynum, T. S. (2006). An analysis of parole decision making using a
sample of sex offenders: A focal concerns perspective. Criminology, 44(4), 961.
Johnson, B. D. (2006). The multilevel context of criminal sentencing: Integrating
judge- and county-level influences. Criminology, 44(2), 259.
Kaba, F., Lewis, A., Glowa-Kollisch, S., Hadler, J., Lee, D., Alper, H, Selling, D.
… Venters, H. (2014). Solitary confinement and risk of self-harm among jail
inmates. American Journal of Public Health, 104(3), 442-447.
King, K., Steiner, B., & Breach, S. R. (2008). Violence in the supermax: A selffulfilling prophecy. The Prison Journal, 88(1), 144-168.

192 • National Institute of Justice | NIJ.gov

Koons-Witt, B. A. (2002). The effect of gender on the decision to incarcerate before
and after the introduction of sentencing guidelines. Criminology, 40(2), 297-328.
Kupers, T. A. (2008). What to do with the survivors? Coping with the long-term
effects of isolated confinement. Criminal Justice and Behavior, 35(8), 1005-1016.
Labrecque, R. (2015). The effect of solitary confinement on institutional
misconduct: A longitudinal evaluation. Doctoral dissertation. University of
Cincinnati, Division of Criminal Justice.
Lanes, E. C. (2009). The association of administrative segregation placement
and other risk factors with the self-injury-free time of male prisoners. Journal of
Offender Rehabilitation, 48(6), 529-546.
Lemert, E. (1951). Social pathology. New York, NY: McGraw-Hill.
Liebling, A. (2004). Prisons and their moral performance: A study of values,
quality, and prison life. New York, NY: Oxford University Press.
Lin, J., Grattet, R., & Petersilia, J. (2010). Back-end sentencing and
reimprisonment: Individual, organizational, and community predictors of parole
sanctioning decisions. Criminology, 48(3), 759-795.
Lindquist, C. A. (1980). Prison discipline and the female offender. Journal of
Offender Counseling, Services and Rehabilitation, 4(4), 305-318.
Lovell, D., & Johnson L. C. (2004). Felony and violent recidivism among
supermax prison inmates in Washington state: A pilot study. Seattle, WA:
University of Washington.
Lovell, D., Johnson, L. C., & Cain, K. C. (2007). Recidivism of supermax
prisoners in Washington state. Crime and Delinquency, 53(4), 633-656.
Marx, G. T. (1981). Ironies of social control: Authorities as contributors to deviance
through escalation, nonenforcement and covert facilitation. Social Problems, 28(3),
221-246.
McClellan, D. S. (1994). Disparity in the discipline of male and female inmates in
Texas prisons. Women & Criminal Justice, 5(2), 71-97.
McGinnis, K., Austin, J., Becker, K., Fields, L., Lane, M., Maloney, M., Marcial,
M., … Felix. T. (2014). Federal Bureau of Prisons: Special housing unit review and
assessment. Arlington, VA: CNA Analysis and Solutions.
Mears, D. P. (2008). An assessment of supermax prisons using an evaluation
research framework. The Prison Journal, 88(1), 43-68.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 193

Mears, D. P. (2013). Supermax prisons: The policy and the evidence. Criminology
and Public Policy, 12(4), 681-719.
Mears, D. P., & Bales, W. D. (2009). Supermax incarceration and recidivism.
Criminology, 47(4), 1131-1166.
Mears, D. P., & Bales, W. D. (2010). Supermax housing: Placement, duration, and
time to reentry. Journal of Criminal Justice, 38(4), 545-554.
Metcalf, H., Morgan, J., Oliker-Friedland, S., Resnik, J., Spiegel, J., Tae, H., Work,
A. R., & Holbrook, B. (2013). Administrative segregation, degrees of isolation, and
incarceration: A national overview of state and federal correctional policies. New
Haven, CT: Liman Public Interest Program.
Metcalf, H., Resnik, J., & Quattlebaum, M. (2015, January). Isolation and
reintegration: Punishment circa 2014. New Haven, CT: The Seventeenth Annual
Liman Colloquium.
Morris, R. G. (2016). Exploring the effect of exposure to short-term solitary
confinement among violent prison inmates. Journal of Quantitative Criminology,
32(1), 1-22.
Motiuk, L. L., & Blanchette, K. (2001). Characteristics of administratively
segregated offenders in federal corrections. Canadian Journal of Criminology and
Criminal Justice, 43(1), 131-143.
Mumola, C. J. (2005). Suicide and homicide in state prisons and local jails. Special
report. NCJ 210036. Washington, DC: Bureau of Justice Statistics.
Nagin, D. S., Cullen, F. T., & Jonson, C. L. (2009). Imprisonment and reoffending.
In M. Tonry (Ed.), Crime and Justice: A Review of Research, 38, pp. 115-200.
Chicago, IL: University of Chicago Press.
Nagin, D. S., & Paternoster, R. (1993). Enduring individual differences and
rational choice theories of crime. Law and Society Review, 27(3), 467-496.
Obama, B. (2015, July 14). Remarks by the President at the NAACP conference
in Philadelphia.
O’Keefe, M. L. (2008). Administrative segregation from within: A corrections
perspective. The Prison Journal, 88(1), 123-143.
Park, J. J. (2000). Redefining Eighth Amendment punishments: A new standard
for determining the liability of prison officials for failing to protect inmates from
serious harm. Quinnipiac Law Review, 20, 407-807.
Pizarro, J. M., & Stenius, V. M. K. (2004). Supermax prisons: Their rise, current
practices, and effect on inmates. The Prison Journal, 84(2), 248-264.

19 4 • National Institute of Justice | NIJ.gov

Rand, M. R. (2009). Criminal victimization, 2008. NCJ 227777. Washington, DC:
Bureau of Justice Statistics.
Reassessing solitary confinement II: The human rights, fiscal, and public
safety consequences. Hearings before the Senate Judiciary Subcommittee on
the Constitution, Civil Rights, and Human Rights, 113th Congress (2014)
(Testimony of G. Mohr on behalf of the Association of State Correctional
Administrators, Feb. 25, 2014).
Riveland, C. (1999). Supermax prisons: Overview and general considerations.
Washington, DC: National Institute of Corrections.
Schafer, N. E. (1986). Discretion, due process, and the prison discipline
committee. Criminal Justice Review, 11(2), 37-46.
Shames, A., Wilcox, J., & Subramanian, R. (2015). Solitary confinement: Common
misconceptions and emerging safe alternatives. New York, NY: Vera Institute of Justice.
Sherman, L. W., Smith, D. A., Schmidt, J. D., & Rogan, D. P. (1992). Crime,
punishment, and stake in conformity: Legal and informal control of domestic
violence. American Sociological Review, 57(5), 680-690.
Smith, P. S. (2006). The effects of solitary confinement on prison inmates: A brief
history and review of the literature. Crime and Justice, 34, 441-528.
Smith, D. A., & Paternoster, R. (1990). Formal processing and future
delinquency: Deviance amplification as selection artifact. Law and Society
Review, 24(5), 1109-1131.
Sparks, J. R., Bottoms, A. E., & Hay, W. (1996). Prisons and the problem of order.
Oxford, UK: Oxford University Press.
Steffensmeier, D., & Demuth, S. (2000). Ethnicity and sentencing outcomes in
US federal courts: Who is punished more harshly? American Sociological Review,
65(5), 705-729.
Steffensmeier, D., & Demuth, S. (2006). Does gender modify the effects of race–
ethnicity on criminal sanctioning? Sentences for male and female white, black,
and Hispanic defendants. Journal of Quantitative Criminology, 22(3), 241-261.
Steffensmeier, D., Ulmer, J., & Kramer, J. (1998). The interaction of race, gender,
and age in criminal sentencing: The punishment cost of being young, black, and
male. Criminology, 36(4), 763-798.
Steiner, B. (2009). Assessing static and dynamic influences on inmate violence
levels. Crime and Delinquency, 55(1), 134-161.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 195

Steiner, B., Butler, H. D., & Ellison, J. M. (2014). Causes and correlates of prison
inmate misconduct: A systematic review of the evidence. Journal of Criminal
Justice, 42(6), 462-470.
Steiner, B., & Cain, C. M. (in press). Assessing the predictors and consequences of
institutional violence: A systematic review of the evidence. White Paper submitted
to the National Institute of Justice, Office of Justice Programs.
Steiner, B., Travis III, L. F., Makarios, M. D., & Meade, B. (2011). Short-term
effects of sanctioning reform on parole officers’ revocation decisions. Law and
Society Review, 45(2), 371-400.
Steiner, B., & Wooldredge, J. (2008a). Comparing state versus facility effects on
crowding in U.S. correctional facilities. Crime and Delinquency, 54(2), 259-290.
Steiner, B., & Wooldredge, J. (2008b). Inmate versus facility effects on prison rule
violations. Criminal Justice and Behavior, 35(4), 438-456.
Steiner, B., & Wooldredge, J. (2009). The relevance of inmate race/ethnicity
versus population composition for understanding prison rule violations.
Punishment and Society, 11(4), 459-489.
Stephan, J. (1989). Prison rule violators. NCJ 120344. Washington, DC: Bureau of
Justice Statistics.
Sundt, J. L., Castellano, T. C., & Briggs, C. S. (2008). The sociopolitical context of
prison violence and its control: A case study of supermax and its effect in Illinois.
The Prison Journal, 88(1), 94-122.
Taxman, F. S., Soule, D., & Gelb, A. (1999). Graduated sanctions: Stepping into
accountable systems and offenders. The Prison Journal, 79(2), 182-204.
Thistlethwaite, A., Wooldredge, J. D., & Gibbs, D. (1998). Severity of dispositions
and the likelihood of domestic violence recidivism. Crime and Delinquency,
44(3), 388-398.
Thomas, J., Mika, H., Blakemore, J., & Aylward, A. (1991). Exacting control
through disciplinary hearings: “Making do” with prison rules. Justice Quarterly,
8(1), 37-57.
Toch, H. (2001). The future of supermax confinement. The Prison Journal, 81(3),
376-388.
Toch, H., Adams, K., & Grant, J. D. (1989). Coping: Maladaptation in prisons.
New Brunswick, NJ: Transaction.
Ulmer, J. T. (2012). Recent developments and new directions in sentencing
research. Justice Quarterly, 29(1), 1-40.

19 6 • National Institute of Justice | NIJ.gov

Ulmer, J. T., & Johnson, B. (2004). Sentencing in context: A multilevel analysis.
Criminology, 42(1), 137-178.
Useem, B., & Kimball, P. (1989). States of siege: U.S. prison riots, 1971-1986. New
York, NY: Oxford University Press.
Wang, X., & Mears, D. P. (2010). A multilevel test of minority threat effects on
sentencing. Journal of Quantitative Criminology, 26(2), 191-215.
Western, B. (2006). Punishment and inequality in America. New York, NY: Russell
Sage Foundation.
Wilkins, L. T. (1964). Social deviance. Englewood Cliffs, NJ: Prentice-Hall.
Wooldredge, J. D. (1994). Inmate crime and victimization in a southwestern
correctional facility. Journal of Criminal Justice, 22(4), 367-381.
Wooldredge, J. D. (1998). Inmate lifestyles and opportunities for victimization.
Journal of Research in Crime and Delinquency, 35(4), 480-502.
Wooldredge, J., Griffin, T., & Rauschenberg, F. (2005). (Un)anticipated effects
of sentencing reform on the disparate treatment of defendants. Law and Society
Review, 39(4), 835-874.
Wooldredge, J., & Steiner, B. (2014). A bi-level framework for understanding
prisoner victimization. Journal of Quantitative Criminology, 30(1), 141-162.
Wooldredge, J., & Steiner, B. (2015). A macro-level perspective on prison inmate
deviance. Punishment and Society, 17(2), 230-257.
Wooldredge, J., & Thistlethwaite, A. (2002). Reconsidering domestic violence
recidivism: Conditioned effects of legal controls by individual and aggregate
levels of stake in conformity. Journal of Quantitative Criminology, 18(1), 45-70.
Wolff, N., Blitz, C. L., Shi, J., Siegel, J., & Bachman, R. (2007). Physical violence inside
prisons: Rates of victimization. Criminal Justice and Behavior, 34(5), 588-599.
Wolff, N., Morgan, R. D., & Shi, J. (2013). Comparative analysis of attitudes
and emotions among inmates: Does mental illness matter? Criminal Justice and
Behavior, 40(10), 1092-1108.

Case Cited
Wolff v. McDonnell, 418 U.S. 539 (1974).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 197

CHAPTER 6

Mental Health Effects of
Restrictive Housing
Reena Kapoor, M.D.
Yale School of Medicine
Robert Trestman, M.D., Ph.D.
University of Connecticut School of Medicine

Introduction

T

he practice of housing prisoners in solitary confinement — typically
defined as 23 hours a day of in-cell restriction with minimal social
contact — has come under increased scrutiny in recent years. Human
rights advocates have described the practice as torture (United Nations,
2011), while corrections officials have historically asserted that the practice is
necessary to maintain the safety and security of prisons. The debate has become
so contentious that opponents cannot even agree about what term to use to
describe the practice, with prison officials preferring “segregation” to the more
emotionally charged term “solitary confinement.” In practice, correctional
systems throughout the U.S. use many terms to describe the same phenomenon:
“administrative segregation,” “enhanced supervision,” “behavior modification,”
“secure housing,” “special housing,” “restrictive housing,” “supermax,” “intensive
management,” “close supervision,” and several others (Liman & Association of
State Correctional Administrators, 2015, p. 1). In this paper, “restrictive housing”

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 19 9

will be used to include all circumstances in which prisoners are removed from
the general population of the institution and confined to their cells for more than
22 hours per day.
A recent report on administrative segregation practices estimates that 80,000 to
100,000 prisoners are in restrictive housing settings in the U.S. on any given day
(Liman & ASCA, 2015). Prisoners are generally placed in restrictive housing
for one of three reasons: (1) for their own protection (protective custody), (2)
because they pose an ongoing security threat (administrative segregation), or
(3) as a disciplinary sanction for violating prison rules (disciplinary custody).
Each of these categories is theoretically distinct, but they are all used to enhance
institutional safety, and the prisoners’ conditions of confinement are often very
similar. A typical restrictive housing cell is approximately 6 feet by 10 feet and
includes a bed (or two), sink, toilet, and desk. The cell may or may not have a
window, and prisoners may or may not be able to control conditions such as
lighting, temperature, and noise. Outside recreation occurs in a small, fencedin area. Prisoners have minimal face-to-face contact with each other but often
communicate by yelling or passing notes under doors.
The debate about the effects of restrictive housing is wide-ranging, but a central
focus in recent years has been on whether the practice is psychologically harmful
to prisoners. Popular media sources have documented numerous biographical
accounts of prisoners’ harrowing experiences in solitary confinement (Gawande,
2009; Guenther, 2012). In addition, mental health professionals and historians
have written several reviews about the psychological effects of solitary
confinement, concluding that the practice causes psychological harm (Grassian,
2006; Haney, 2003; Shalev, 2008; Grassian & Friedman, 1986; Scharff-Smith,
2006). Some scholars and advocates have considered this body of evidence
conclusive, stating that there is no longer any question that solitary confinement
causes serious and long-lasting psychological damage (Arrigo & Bullock,
2008; Haney, 2003; Grassian, 2006; Human Rights Watch, 2015; ACLU, 2014).
However, other scholars have been more circumspect (O’Keefe, 2007; Bonta
& Gendreau, 1990; Gendreau & Labrecque, 2015), pointing to the dangers of
drawing conclusions from studies with imperfect scientific methods (e.g., small
sample sizes, lack of control groups), questionable applicability to modern
prisons (e.g., sensory deprivation experiments in laboratories), and potential for
bias (e.g., conducting studies in the context of litigation).
Indeed, restrictive housing is a notoriously difficult practice to study. Restrictive
housing units all have some characteristics in common — social isolation,
changes in sensory stimulation, and confinement beyond the experiences of the
general prison population (Zubek, Bayer, & Shephard, 1969) — but the degree
to which each of these characteristics is present in a given facility or housing
unit varies greatly between institutions. Therefore, scientific studies conducted
in different housing units around the country — let alone the world — may
actually be studying very different conditions, and extrapolating the results from

20 0 • National Institute of Justice | NIJ.gov

one setting to another may be erroneous. These differences make it challenging
for scholars, administrators, or legislators to draw broad-based conclusions
with certainty. Currently, there is simply no way to standardize conditions
in restrictive housing units and conduct the type of large-scale, randomized
controlled studies that would be optimal from a scientific perspective. Thus, we
are left with imperfect data from which to draw conclusions and make decisions
about the appropriate management of prisoners.
This white paper will review the current state of scientific evidence about the
psychological effects of restrictive housing. First, it describes the method by
which the review was conducted. Second, it discusses critiques of the literature
and the challenges that face researchers attempting to conduct rigorous scientific
investigations of restrictive housing. It examines the evidence about several
factors relevant to mental health in restrictive housing:
• The purpose, duration, and conditions of confinement.
• Access to mental health care.
• Development or exacerbation of psychiatric symptoms.
• The influence of age, gender, intellectual disability, and mental illness.
• Rates of self-injury, psychiatric hospitalization, and institutional misconduct.
It then reviews the consensus statements of major mental health professional
organizations. Finally, it identifies gaps in the current knowledge and
recommends future research and policy changes.

Method of Review
This review includes articles in English-language, peer-reviewed medical, legal,
and social science journals; book chapters; and published dissertations that
present empirical data related to mental health and restrictive housing. Studies
were first identified using PubMed, the Social Science Research Network (SSRN),
and Google using combinations of the search terms “solitary confinement,”
“administrative segregation,” “supermax,” “psychological effect,” “psychiatric
effect,” and “mental health.” Bibliographies of key articles (Haney, 2003; Grassian,
2006; Scharff-Smith, 2006; Shalev, 2008; Gendreau & Labrecque, 2015; Frost &
Monteiro, 2016) were examined to identify additional relevant studies. Papers
were excluded from the review if —
• No original data were presented.
• Only biographical or anecdotal evidence was presented.
• The research was not conducted in a prison setting.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 201

• The research was not conducted on humans.
• The findings were published only in lay media or in advocacy group literature.
To examine evidence that applies to modern American prisons, this review
includes studies conducted in prisons and published after 1980. However, older
studies were included if no modern studies existed in a particular content area
(e.g., a 1972 study that showed electroencephalogram (EEG) changes in the
brains of subjects held in solitary confinement).
This type of narrative review has its limitations. Some have argued that metaanalysis of the scientific data is a more reliable method of interpreting the
existing literature (Gendreau & Labrecque, 2015). However, only two such
meta-analytic studies about the effects of administrative segregation have been
performed (Gendreau & Labrecque, 2015; Morgan, 2016), and several areas of
inquiry related to mental health and restrictive housing deserve consideration
but have not been assessed using meta-analysis. Thus, this paper includes a
descriptive review of some studies in addition to the results of meta-analyses.

Research and Data Limitations
Research about the psychological effects of restrictive housing follows a common
pattern of scientific inquiry; that is, the quality of research studies improves
over time. Early scholars in the 1950s and 1960s relied on clinical observations
of prisoners and research in related areas (e.g., sensory deprivation or prisoners
of war) to generate hypotheses about the effects of solitary confinement. To
test these hypotheses, small-scale studies were conducted from the 1980s to
2000s, often generating conflicting results based on research design (Suedfeld,
Ramirez, Deaton, & Baker-Brown, 1982; Grassian, 1983; Zinger, Wichmann, &
Andrews, 2001; Andersen et al., 2000). Over the past five years, due in part to
increased litigation and popular interest in solitary confinement, larger and more
methodologically rigorous studies have been conducted (O’Keefe et al., 2013;
Kaba et al., 2014; Gendreau & Labrecque, 2015). These more recent studies have
not always clarified the specific effects of restrictive housing on mental health,
but they have significantly increased the amount of information available about
restrictive housing practices.
The limitations of the current scientific data have been well documented
elsewhere (Gendreau & Labrecque, 2015; O’Keefe, 2007; Scharff-Smith 2006;
Zinger et al., 2001; Metzner, 2015), so this paper states them only briefly here:
• Policies, procedures, and conditions of confinement vary widely between
institutions. No two restrictive housing units are exactly alike, so scientific
conclusions in one setting may not apply to other settings, even when both
are called, for example, “administrative segregation.” Further complicating

202 • National Institute of Justice | NIJ.gov

matters, many published studies of restrictive housing do not specifically
describe the conditions of confinement, so it is impossible for a reader to
know how the results of one study compare to another’s.
• There are no standard definitions in the literature for terms such as mental
illness, harm, benefit, short term, and long term. For example, when one
study concludes there is no evidence of harm from short-term isolation
to inmates with mental illness, another could conclude just the opposite,
depending on the definitions used. Unfortunately, most of the published
studies do not define the terms they use, leaving the reader to speculate on the
authors’ intended meaning. When the terms are defined, the definitions vary.
For example, some studies base inclusion criteria on psychiatric symptoms
or diagnosis, while others include any prisoner who receives mental health
services in the correctional institution. Similarly, “short term” might refer to
seven days, or it might refer to periods as long as three months. Finally, harm
can be conceptualized broadly — the development of any new psychological
symptoms — or narrowly, such as psychiatric hospitalization or serious
suicide attempts.
• In many cases, study designs are limited. Many studies of restrictive housing
have small sample sizes (Suedfeld & Roy, 1975; Brodsky & Scogin, 1988), high
attrition rates (Zinger et al., 2001; Miller, 1994), and use volunteer prisoners
(Gendreau, Freedman, Wilde, & Scott, 1972). Some studies do not include
control groups (Jackson, 1983; Grassian, 1983), which prevents comparison
to the potential effects of other relevant conditions of confinement. Other
studies were conducted on prisoners involved in class-action litigation against
the prison at the time of examination (Grassian, 1983), posing inevitable
concerns about report bias. While none of these limitations render the studies’
conclusions invalid, they do necessitate the use of caution when generalizing
their results.
• Correlation can be confused with causation. Studies that use a crosssectional design or do not include control groups have the potential to
conflate correlation and causation. For example, if a study finds that
individuals with lung cancer own cigarette lighters at a much higher rate
than those without lung cancer, this does not mean that lighters cause lung
cancer. In the case of restrictive housing, if individuals in restrictive settings
exhibit higher rates of mental illness, this does not necessarily mean that the
housing placement itself caused the symptoms. To reach that conclusion,
symptom rates in control groups must be compared, and intervening factors
must be considered.
• Prisons and mental health treatment have both changed substantially over
the relevant period of study. Many of the relevant studies about restrictive
housing were conducted from the 1980s to 2010s, a period when American
prisons experienced massive growth (Bureau of Justice Statistics, 2014). At the

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 203

same time, psychiatry underwent two major revisions of its diagnostic criteria
as presented in its Diagnostic and Statistical Manual of Mental Disorders,
taking the manual from its third edition (DSM-III) (American Psychiatric
Association, 1980) to its fifth edition (DSM-V) (American Psychiatric
Association, 2013). These changes call into question the relevance of older
studies conducted in prisons, particularly when the exact conditions of
confinement or definitions of mental disorders were not delineated.
• The political and social context surrounding solitary confinement is highly
charged. When considering important social and political questions, such as
how our society treats prisoners, a fair and thorough examination of scientific
evidence is both essential and difficult to achieve. Parties tend to be polarized,
and scholars can be tempted to align with one side or the other, potentially
introducing unintended bias into what should be an objective inquiry.

Literature Review
This review begins by examining the prevalence of mental illness in inmates
who live in restrictive housing. Next, it considers the psychological effects of
restrictive housing by examining data about the effect of specific characteristics
of confinement, such as the length of confinement and single versus shared
cells, and characteristics of the individual, such as age, gender, or pre-existing
mental illness. It also examines behavioral outcomes in restrictive housing, such
as suicide rates and institutional misconduct. Finally, it considers the long-term
effects of isolation, including the persistence of psychological symptoms or the
development of new disorders such as post-traumatic stress disorder (PTSD).

Prevalence of Persons With Mental Illness
in Restrictive Housing
Research has consistently demonstrated that prisoners in restrictive housing
settings have higher rates of diagnosed mental disorders, higher rates of
psychiatric symptoms (as measured by symptom rating scales), and more severe
psychiatric symptoms than inmates in the general prison population. In contrast
to the 10 percent to 15 percent prevalence of mental illness in prisons generally
(Bureau of Justice Statistics, 2002; Lamb & Weinberger, 1998), the prevalence of
serious mental illness in restrictive housing has been estimated at approximately
30 percent in several different studies. Estimates range from 15 percent to
62 percent, depending on the definition of mental illness and the assessment
method used (Hodgins & Côté, 1991; Lovell, Cloyes, Allen, & Rhodes, 2000;
Andersen et al., 2000; Zinger et al., 2001; Wynn & Szatrowski, 2004; O’Keefe,
2007; Labrecque, 2015). Lower rates are associated with more restrictive
definitions, such as serious mental illness, while higher rates are associated

20 4 • National Institute of Justice | NIJ.gov

with more inclusive definitions such as “on the mental health caseload.” Some
diagnoses are overrepresented in restrictive housing populations, such as
schizophrenia and bipolar disorder (Hodgins & Côté, 1991), depression and
adjustment disorders (Andersen et al., 2000), and attention-deficit hyperactivity
disorder (O’Keefe, 2007). Lanes (2011) demonstrated that prisoners who
engage in self-injury are likely to be placed in restrictive housing; the likelihood
increased in tandem with the severity of the self-injury.
Studies also indicate that prisoners in restrictive housing demonstrate high rates
of symptoms beyond just the diagnosis of serious mental illness. Grassian (1983)
and Haney (2003) have both described prisoners in restrictive housing settings as
having high levels of anxiety, anger, sleep problems, perceptual distortions, and
somatic symptoms. Although these studies were qualitative and based on clinical
interviews, other studies have found similar results using different methods.
Lovell (2008) found that 20 percent of the restrictive housing prisoners had a
serious mental illness, but an additional 25 percent had evidence of “marked
psychological dysfunction, psychological breakdowns, or brain damage.” Cloyes
and colleagues (2006) administered a structured rating scale of psychiatric
symptoms and concluded that prisoners in restrictive housing displayed evidence
of “moderate psychosocial dysfunction,” with a significant portion of the
study sample displaying evidence of serious dysfunction. After administering
structured rating scales, the O’Keefe team (2013) also found that prisoners in
restrictive housing demonstrated higher rates of psychological symptoms than
general population prisoners.
Several studies have also examined the personality characteristics of prisoners
in restrictive housing. Comparing inmates in restrictive housing and the general
prison population, Motiuk and Blanchette (1997) found those in restrictive
housing had significantly more cognitive and personality problems. For example,
compared to prisoners in general population settings, the restrictive housing
prisoners were more impulsive, had difficulty solving interpersonal problems,
set unrealistic goals, demonstrated low frustration tolerance and disregard for
others, and had narrow and rigid thinking. Similarly, O’Keefe (2007) found that
prisoners in administrative segregation displayed more thinking disorders and
suspicious hostility than general population prisoners.
Taken together, the research clearly indicates that prisoners in restrictive housing are
a disturbed group, with disproportionately high rates of diagnosed mental illness,
psychological symptoms (whether diagnosed or not), and maladaptive personality
traits. As described below, the degree to which these problems pre-existed the
inmate’s placement in restrictive housing is unclear, but the high prevalence of
individuals with significant mental health concerns in restrictive housing has
been amply demonstrated.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 205

Effect of Particular Characteristics of Confinement
Purpose of Confinement
Does it matter why an individual is placed in restrictive housing? For example,
are individuals who volunteer for protective custody healthier and less
symptomatic than those who are placed in restrictive housing for disciplinary
purposes? And does the prisoner’s knowledge and attitude about the
confinement, such as knowing how long it will last and perceiving the placement
as legitimate, make a difference?
No published studies have directly addressed these questions, but a few provide
relevant data for consideration. For example, the literature about sensory
deprivation demonstrates that individuals’ expectations about what they will
experience in confinement can significantly alter the symptoms they report;
individuals who are not told that confinement can cause distress will report fewer
negative psychological experiences (Gendreau & Labrecque, 2015; Grassian &
Friedman, 1986). Conversely, if individuals believe the experience is likely to
be benign, one early study found evidence that they may enjoy it (Goldberger,
1966). Although far from conclusive, these studies indicate that a prisoner’s
perception of restrictive housing may significantly affect his or her psychological
experience. Most prisoners have heard colloquial descriptions of “solitary,” “the
hole,” and “the box,” all of which have negative connotations and can possibly
lead to more negative perceptions of the experience of restrictive housing.
A few studies have examined differences between voluntary (i.e., protective
custody) and involuntary (i.e., disciplinary or administrative) prisoners in
restrictive housing. No differences were found in the psychological functioning
of voluntary and involuntary prisoners (Zinger et al., 2001). Motiuk and
Blanchette (1997) found that prisoners in voluntary restrictive housing were
more likely to report prior victimization, but their psychological characteristics
were no different from involuntary prisoners in restrictive housing. Miller
and Young (1997) compared prisoners in administrative custody, disciplinary
custody, and general population, finding that those in disciplinary custody most
frequently reported feelings of withdrawal, anger, and hostility. Brodsky and
Scogin (1988) interviewed 69 men in two different protective custody units.
They found that two-thirds of them had significant psychopathology, and the
psychological symptoms were worse in the unit without adequate space, light,
and programming. Although they did not compare their subjects with those
in administrative or disciplinary segregation, their findings indicated that the
conditions of confinement may be just as (or more) important than the purpose
of confinement in determining psychological outcomes.

20 6 • National Institute of Justice | NIJ.gov

Duration of Confinement
There are two theories about how the duration of restrictive housing relates
to mental health. In the first theory, individuals experience the greatest
psychological discomfort — anxiety, fear, depression, anger — in the first few
days of confinement. Over time, they adjust to the conditions and display fewer
symptoms. In the second theory, individuals are “driven mad” by isolation,
becoming more anxious, aggressive, and delirious the longer they are kept in
restrictive housing. Although anecdotal reports tend to support the latter theory,
the empirical evidence is much more mixed.
Most researchers have not found that individuals developed increased
psychological symptoms or significant behavioral changes during short-term
placement in restrictive housing. Prisoners who volunteered to spend seven
days in solitary confinement showed EEG changes, but no behavioral differences
(Gendreau et al., 1972). Another study found no changes in psychological
testing results during five days of confinement (Weinberg, 1967, as reported in
Suedfeld et al., 1982). Walters, Callagan, and Newman (1963) found no mental
deterioration when volunteer prisoners spent four days in isolation. Ecclestone,
Gendreau, and Knox (1974) studied physiological markers of stress, such as heart
rate, blood pressure, and plasma cortisol levels, finding no significant changes
in levels during 10 days of isolation. Zinger and colleagues (2001) studied
prisoners in restrictive housing and general population for a longer period — 60
days — and found that psychological symptoms decreased in both groups, with
no differences in suicidal ideation. Labrecque (2015) examined institutional
misconduct rates, comparing those who spent less than 15 days in segregation
with those who spent more than 15 days and found no differences between
the groups. The Colorado study (O’Keefe et al., 2013) found that psychological
symptoms decreased over the first 90 days of confinement in isolated prisoners
with and without mental illness.
In contrast, data from well-designed studies on the long-term effects of restrictive
housing are sparse and conflicting. Several researchers have each described
compelling case studies of prisoners whose psychological functioning severely
deteriorated over time (Grassian, 1983; Haney, 2003; Jackson, 1983). However,
without control groups or standardized measures, their findings came under
scrutiny (O’Keefe, 2007; Bonta & Gendreau, 1990). Andersen and colleagues
(2000) did lend empirical support to the case reports, using a longitudinal study
design and finding that prisoners developed more depression and adjustment
disorders over the four-month study period. The O’Keefe team (2013) evaluated
prisoners prospectively over one year and found an overall decrease in
psychological symptoms. In their meta-analysis, Gendreau and Labrecque found
a small but statistically significant increase in psychological symptoms over time,
but the duration of the prisoners’ stay in solitary confinement varied. Overall,

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 207

the data are mixed, and there is currently no clear answer to the question of
whether any particular duration of restrictive housing is safe or harmful from a
psychological standpoint.

Degree of Social Isolation
No modern studies have examined questions such as, “Does placing two inmates
together in a double cell cause less psychological distress than isolating one
inmate in a single cell?” Humans are undoubtedly social creatures, and historical
accounts of 19th-century penitentiaries that routinely used solitary confinement
(i.e., the Pennsylvania and Auburn models) provide a compelling narrative about
prisoners driven insane by isolation (Scharff-Smith, 2006). However, modern
restrictive housing units differ significantly from the conditions of 19th-century
prisons and it is unclear to what extent historical findings are relevant to modern
correctional practice.
It is clear, however, that the risk of suicide in single cells in restrictive housing is
substantial. Reeves and Tamburello (2014) found that all but one of the suicides
in the New Jersey Department of Corrections over a five-year period occurred
in a single cell, concluding that placement in a single cell in restrictive housing
carries 400 times greater risk of suicide than a general population double cell.
A study of the California prison system found that 73 percent of the completed
suicides in a six-year period occurred in a single cell, with 45 percent occurring
in administrative segregation (Patterson & Hughes, 2008). The issue of suicide
is discussed in more detail below (see “Rates of Self-Injury and Suicide”); the
data are mentioned here merely to indicate that the degree of social isolation
in restrictive housing units may be an important variable in predicting
psychological outcomes. Further study is needed to assess this question.

Physical Plant Characteristics
Physical plant characteristics, such as cell size, recreation yard size, food quality,
amount of natural light, and noise levels each have a potential impact on
psychological functioning in restrictive housing settings. For example, despite
being called “solitary,” many restrictive housing units are actually very noisy
because prisoners yell back and forth between cells to communicate with each
other. Anecdotal reports indicate that this type of constant background noise has
a detrimental effect on mental health (Childress, 2014). To date, no published
studies have systematically examined the issue of how particular physical plant
characteristics affect psychological functioning in restrictive housing settings.
Brodsky and Scogin (1988) did find greater levels of psychological distress in a
group of prisoners isolated in a unit without adequate light or space, providing
preliminary data that these conditions may be important determinants of
psychological health.

208 • National Institute of Justice | NIJ.gov

Degree of In-Cell Activity
Many prisoners in restrictive housing units are allowed to keep televisions,
radios, books, and other sources of sensory stimulation in their cells (Metzner,
2002), presumably because such items are thought to keep prisoners occupied
and decrease emotional distress. No systematic studies have been conducted
to determine whether access to such items has an effect on psychological
functioning in restrictive housing. Likewise, no studies have assessed the
impact of physical stimulation (exercise in the recreation yard or in cell) on
psychological functioning. One small study of a Kentucky prison did examine
three groups of prisoners with varying degrees of restriction: general population,
administrative segregation, and disciplinary segregation (Miller, 1994). The study
found that psychiatric symptoms were proportional to the degree of restriction,
with the most symptoms found in the disciplinary custody group housed in
single cells, not allowed to smoke, and with the least access to commissary items.

Access to Mental Health or Other Programming
Correctional facilities vary widely in the amount of programming and mental
health services provided to prisoners in restrictive housing, which likely reflects
the divergent views of corrections professionals about the role of mental health
problems in restricted prisoners’ behavior. In a 2014 survey of state correctional
systems, some jurisdictions reported that mental health concerns for segregated
prisoners are “significant” or “100 percent,” while others reported that mental
health plays a “minimal” role in segregated prisoners’ behavior (Liman & ASCA,
2015, p. 57). Similarly, the survey found that some systems divert prisoners who
have a diagnosed mental illness out of administrative segregation, while others
indicated that grouping such prisoners together in administrative segregation
improves access to mental health staff (Liman & ASCA, 2015, p. 57).
Given the lack of consensus about the role of mental illness in problematic
behavior that occurs during restrictive housing placement, it is not surprising
that access to mental health care in that setting varies greatly, ranging from
essentially none to more than 20 hours a week of structured or unstructured
activity. Mental health involvement in restricted prisoners’ care can include —
• Prescreening by a nurse or mental health professional to exclude prisoners
with medical or mental health contraindications from placement in
restrictive housing.
• Cell-side “wellness checks” by medical or mental health staff, conducted with
varying frequency (daily to weekly, depending on correctional policies and
degree of isolation).
• Cell-side administration of psychotropic medication by a nurse.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 209

• In-cell mental health programming, typically consisting of workbook
assignments, journals, or other reading materials.
• Out-of-cell individual evaluations by a mental health professional (ranging
from several times per week to every two to three months), depending on the
clinical status of the prisoner and institutional policies.
• Group activities with a social or nontherapeutic purpose (unstructured activity),
such as watching movies or playing games.
• Group activities with a therapeutic purpose (structured activity), such as
group psychotherapy or interacting with a therapy animal, typically conducted
using either shackles or “therapeutic modules” (telephone booth-sized cages)
to secure prisoners.
In some jurisdictions, prisoners with mental illness have access to these services
but remain in the same physical location as other restricted prisoners. Their daily
routines, including access to recreation, showers, commissary items, and phone
calls, remain unchanged. However, some large correctional systems have developed
specialized residential programs for prisoners with significant mental health
concerns and repeated disciplinary infractions. These programs use principles of
cognitive behavioral therapy or dialectical behavioral therapy to provide tangible,
short-term incentives for prosocial behavior to improve outcomes for prisoners
who have failed to progress through traditional disciplinary custody programs.
One example is the Secure Residential Treatment Units in the Pennsylvania
Department of Corrections (Commonwealth of Pennsylvania, 2015). Rather than
relying solely on sanctions for bad behavior, the specialized programs provide
positive reinforcement for good behavior, such as the ability to earn daily “points”
that can be spent on rewards at the end of the week. Common rewards include
snack foods, hygiene products, and extra phone time.
Although most mental health clinicians would recommend greater access to
programming for prisoners in restrictive housing, no published studies have yet
examined whether access to any particular type of mental health treatment has
an impact on psychological health. Similarly, no studies have examined whether
other types of programming — educational, religious, vocational, recreational —
have any impact. Some scholars have raised concerns that, even when out-of-cell
programming is offered in administrative segregation, in most jurisdictions less
than 25 percent of inmates actually participate (Liman & ASCA, 2015, p. 48).
These knowledge gaps leave us with two important areas for future study:
(1) how best to design mental health programming for restricted prisoners, and
(2) how the policies and programs are being implemented at the ground level.
Some preliminary data do indicate that improving mental health services
for restricted prisoners has positive outcomes. Kupers and colleagues (2009)
reported substantial reductions in the use of force and inmate assaults after the
Mississippi Department of Corrections improved its mental health services and

210 • National Institute of Justice | NIJ.gov

revised its restrictive housing practices. These results are promising, but they
must be interpreted with caution. The authors did not measure psychological
functioning as an outcome, and the changes implemented were much
more sweeping than simply providing access to mental health services for
restricted prisoners.

Relationship Between Staff and Prisoners
Scholars have documented that correctional employees assigned to restrictive
housing units often have negative attitudes toward segregated prisoners (Wormith,
Tellier, & Gendreau, 1988; Carriere, 1989). Prisoners are aware of this. They
complain just as much about staff attitudes, such as lack of respect and the
humiliation it can lead to, as they do about sensory deprivation or social isolation
in restrictive housing (Suedfeld et al., 1982). Although the issue has not been
systematically studied, some scholars postulate that treating prisoners fairly and
humanely, even in conditions of relative isolation, may have a significant mitigating
effect on the psychological harms of restrictive housing (Gendreau & Bonta, 1984;
Suedfeld et al., 1982; Gendreau & Labrecque, 2015).

Effect of Individual Characteristics
Individuals Without Pre-Existing Mental Illness
Many studies about the psychological effects of restrictive housing report their
findings without distinguishing individuals with pre-existing psychiatric illness
from those without. This, of course, makes it difficult to assess the question of
whether psychologically healthy individuals respond differently to segregated
confinement than those with mental illness. Common sense and clinical judgment
would lead to the belief that differences in response could be significant, but the
small amount of existing data do not allow definitive conclusions.
Early studies of restrictive housing excluded individuals with pre-existing
mental or physical disorders, providing some information about how “normal”
individuals respond to confinement. These studies (Suedfeld et al., 1982;
Ecclestone et al., 1974; Gendreau & Bonta, 1984) found no adverse effect of
solitary confinement for healthy individuals over relatively short periods.
The Colorado study results (O’Keefe et al., 2013) were similar, finding that
individuals without pre-existing mental illness who were placed in administrative
segregation had higher initial rates of psychological symptoms than those in
general population, but the symptoms decreased over time. When scholars
interviewed men in the community who had spent time in restrictive housing
during their incarcerations in New York, they found that 70 percent said they
felt safer in confinement than in general population, though they still viewed the
experience negatively (Valera & Kates-Benman, 2015).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 211

By contrast, others who performed clinical assessments of prisoners described
a distinct psychiatric syndrome, , secure housing unit (SHU) syndrome, that
can affect individuals regardless of a pre-existing mental illness diagnosis
(Grassian, 1983; Haney & Lynch, 1997). Individuals with this syndrome will
become progressively more anxious, irritable, confused, aggressive, and selfinjurious over time, but their symptoms will dissipate rapidly after release from
segregated confinement. Andersen and colleagues (2000) studied a similar
question, assessing prisoners without known psychiatric illness who were placed
in restrictive housing immediately upon entering the correctional facility. Using
repeated symptom assessments over four months, the study found that prisoners
in solitary confinement were more likely to develop psychiatric disorders than
those in general population (28 percent compared to 15 percent). A follow-up
study (Andersen et al., 2003) found that psychiatric symptoms decreased over time
in the control group but remained stable in the restrictive housing group. Kaba
and colleagues (2014, 2015) examined rates of self-injury and suicide in isolated
prisoners, finding that individuals without mental illness were more likely than
those with serious mental illness to engage in non-lethal acts of self-injury. The
authors hypothesized that inmates who did not have a diagnosed mental illness
engaged in self-injury as a means to change housing conditions, not that the
confinement caused new psychiatric problems.
Taken together, these conflicting results do not lead to a clear picture of how a
“normal” person responds to restrictive housing. In 2015, scholars attempted to
use rigorous scientific methods to shed light on the question, performing two
meta-analyses of the effects of restrictive housing. After narrowing a sample of 150
studies to just 15 (Gendreau & Labrecque, 2015) or 14 (Morgan, 2016) that met
inclusion criteria, both meta-analyses concluded that administrative segregation
has a small but significant negative impact on psychological functioning, with
the greatest changes being in the domains of anxiety and depression. The metaanalyses did not comment specifically on differences between populations with
and without mental illness, but they do provide a useful baseline from which to
compare outcomes in individuals with mental illness.

Individuals With Pre-Existing Mental Illness
As described in the “Prevalence of Mental Illness” section above, individuals
with serious mental illness are overrepresented in restrictive housing
populations, likely because they engage in disruptive behaviors and accrue
institutional misconduct reports. In some correctional institutions, particularly
those without well-developed systems of mental health care, individuals with
mental illness may be viewed as unmanageable or particularly dangerous in
general population. Additionally, some prisoners genuinely seem to prefer
being in restrictive housing, finding general population too stimulating or
threatening. These prisoners will sometimes deliberately commit infractions
when they are nearing release from restrictive housing, seemingly for the sole
purpose of remaining in confinement.

212 • National Institute of Justice | NIJ.gov

Given the widespread consensus among mental health professionals and human
rights advocates that individuals with serious mental illness should be excluded
from prolonged solitary confinement, there is surprisingly little empirical evidence
that demonstrates an exacerbation of psychiatric symptoms in restrictive housing
for individuals with a mental illness diagnosis. A few studies have concluded that
most inmates displaying symptoms of mental illness in restrictive housing were
diagnosed prior to entry. For example, Hodgins and Cote (1991) found that 86
percent and 64 percent of prisoners diagnosed with mental illness in two different
restrictive housing units were diagnosed prior to placement. Andersen et al.
(2000) also found high rates of psychiatric symptoms at the time of placement in
solitary confinement. Inmates with mental illness diagnoses spend much longer
in restrictive housing than those without such a diagnosis; one study found that
prisoners with serious mental illness spent an average of 38 months in isolation,
compared to 5 months for prisoners who did not have a diagnosed mental illness
(Correctional Association of New York, 2004).
The data about how persons with mental illness respond after being placed in
confinement are mixed. An early study found positive behavioral change in
four inmates with schizophrenia who were placed in restrictive housing for
seven to 10 days. The inmates displayed decreased aggression, violence, selfinjury, and psychotic symptoms for two years after release from confinement
(Suedfeld & Roy, 1975). Of course, a study with such a small sample size must be
interpreted with caution. A much larger study was conducted in the New York
City jails (Kaba et al., 2015). It examined the relationship between the timing
of mental illness diagnosis and restrictive housing placement and hypothesized
that more diagnoses would be made over time. However, the findings did not
support this hypothesis; the diagnoses clustered in a normal distribution around
“Day 0” of restrictive housing placement. Prisoners were most often diagnosed
with adjustment disorders and antisocial personality disorders, suggesting that
they came to mental health attention because they were distressed about being
placed in restrictive housing, not because psychotic or mood symptoms were
exacerbated by confinement. The findings of the Colorado study (O’Keefe et al.,
2013) were similar, concluding that inmates with mental illness experienced the
greatest severity of symptoms just after placement in administrative segregation,
with a decrease in symptoms over 12 months.
The case studies of Grassian (1983) and Haney (2003) reach the opposite
conclusion, stating that restrictive housing places prisoners with mental illness at
great risk of decompensation over time. The empirical literature does lend some
support to this theory. Kaba and colleagues (2014) determined that placement
in solitary confinement increases the risk of suicide attempts and self-injury for
all prisoners, even after release from confinement. Prisoners with serious mental
illness are at particular risk of engaging in potentially lethal acts of self-injury
while in solitary confinement (9.8 times the general prison risk). Despite finding
that psychological symptoms decreased overall in prisoners with serious mental

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 213

illness, O’Keefe and colleagues (2013) noted that 7 percent of the group with
serious mental illness experienced an increase in symptoms. Another study that
examined institutional infractions committed by inmates in restrictive housing
found that a smaller percentage of the inmates with serious mental illness
committed infractions, but those who did, did so repeatedly (Smith, Labrecque,
& Gendreau, 2015). Overall, these findings suggest that some prisoners with
mental illness adapt well to restrictive housing, but a significant minority may
experience catastrophic results, including additional disciplinary infractions and
potentially lethal suicide attempts.

Age
Prisoners in restrictive housing settings average approximately 30 years of age,
compared to 35 years for the general prison population (O’Keefe, 2007; Lovell,
Cloyes, Allen, & Rhodes, 2000; Cloyes et al., 2006), suggesting that young age
and psychological immaturity are risk factors for such placement. Kaba and
colleagues (2015) support this theory, finding that prisoners under age 21 are
five times as likely to be placed in solitary confinement as prisoners over age 21.
Younger prisoners, particularly those 18 years and younger, have a significantly
higher risk of suicide in prison, though it is not clear whether restrictive housing
elevates this risk (Kaba et al., 2014).
Experts in child mental health agree with the prohibition on placing juveniles
in solitary confinement endorsed by the United Nations (United Nations
General Assembly, 1990), the U.S. Department of Justice (2016), and President
Obama (2016), as set forth in the “United Nations Rules for the Protection of
Juveniles Deprived of their Liberty, 1990, Section 67.” The American Academy
of Child and Adolescent Psychiatry (AACAP, 2012) recommends that, except
in extraordinary circumstances, juveniles should not be placed in restrictive
housing. Although the data are limited, they support the theory that, the younger
the child, the greater the potential harm from placement in restrictive housing.

Gender
The majority of research on the effects of restrictive housing has been conducted
on men. Qualitative reviews of women in restrictive housing (Korn, 1988;
Martel, 1999) are similar to those documented by Grassian (1983), describing
depression, anger, hallucinations, and withdrawal. Women placed in restrictive
housing do share some traits in common with men, particularly their high rates
of institutional maladjustment and criminogenic risk (Thompson & Rubenfeld,
2013). However, a small number of quantitative studies point to potentially
significant differences in the way men and women experience restrictive
housing. O’Keefe (2007) found that women make up a disproportionately small
percentage of the prisoners in restrictive housing. Although women are more
likely than men to have psychiatric diagnoses, they are less likely to be placed

214 • National Institute of Justice | NIJ.gov

in restrictive housing, indicating that women prisoners with mental illness may
be offered treatment, while men are punished. Suedfeld and colleagues (1982)
studied women in a quasi-restricted setting (they ate meals together and spent
some time out of their cells) and concluded that women use different coping skills
than men to tolerate the experience. They fantasize, daydream, and recall books
they have read and movies they have seen. When examining outcomes for women
in restrictive housing, Labrecque (2015) found that, in contrast to men, rates of
institutional misconduct for women decreased by more than 20 percent after they
were placed in confinement. These data are preliminary, and further investigation
of women in restrictive housing is needed before conclusions can be drawn.

Intelligence and Cognitive Functioning
Some evidence suggests that individuals in restrictive housing have lower
intelligence scores and more cognitive problems than prisoners in general
population. One small study (Zinger et al., 2001) found that prisoners in restrictive
housing had a mean IQ score (89.70) that was 8 points lower than the general
prison population, though still within the normal range. Studies have also shown
that prisoners in restrictive housing display less flexible thinking and are less able to
solve problems than non-restricted prisoners (Motiuk & Blanchette, 1997; O’Keefe,
2007). A review of medical charts found that 30 percent of prisoners in restrictive
housing had documented evidence of traumatic brain injuries (Lovell, 2008). These
studies suggest that individuals with intellectual disabilities may be overrepresented
in restrictive housing, and they may be less able to cope with the conditions of
confinement than the average prisoner. However, no published studies have
systematically examined this issue.

Prior Experience in Restrictive Housing
Studies have shown that, on any given day, many of the prisoners in restrictive
housing settings have been there before, because the same individuals tend to
commit rule violations repeatedly (Lovell, 2008; Zinger et al., 2001). One could
hypothesize that experienced prisoners find restrictive housing less distressing,
but one could also hypothesize that the effects of isolation are cumulative and
cause more problems over time. No published studies have addressed this issue.

Behavioral Outcomes
Self-Injury and Suicide
Research clearly indicates that restrictive housing placement, particularly in
a single cell, is significantly correlated with prisoner suicide. Studies from
large correctional systems have shown that a disproportionate number of
suicides occur in restrictive housing units (Way, Miraglia, Sawyer, Beer, &

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 215

Eddy, 2005; White, Schimmel, & Frickey, 2002; Patterson & Hughes, 2008;
Reeves & Tamburello, 2014), with estimates ranging from 30 percent to 65
percent. Another study found that 14 percent of men who had been placed in
restrictive housing reported attempting suicide while there (Valera & KatesBenman, 2015), and a large-scale meta-analysis concluded that placement
in administrative segregation has a moderate effect on self-injury (Morgan,
2016). Kaba and colleagues (2014) found that the risk of suicide while in
restrictive housing is more than six times greater than in general population,
but the risk was also increased after the prisoner had been released (two times
greater than those who had never been placed in restrictive housing). Reeves
and Tamburello (2014) concluded that placement in a single cell in restrictive
housing carries a risk of suicide that is more than 400 times higher than that of
the general prison population.
In another study conducted in New York, scholars reported that suicides in
restrictive housing units occurred, on average, after 63 days of confinement
(Way, Sawyer, Barboza, & Nash, 2007). The authors advocate for enhanced
observation of prisoners during the first eight weeks of confinement in
restrictive housing, when most suicides occurred. There is also some evidence
that suicide rates increase incrementally as the degree of isolation increases. In
Italy, suicide rates in short-term restrictive housing were 239 percent higher
than in the general prison population and 439 percent higher in long-term
restrictive housing (such as an American supermax facility) (Roma, Pompili,
Lester, Girardi, & Ferracuti, 2013). Of course, risk factors other than restrictive
housing placement per se can contribute to the elevated incidence of suicide
in that setting. Mental illness, history of suicide attempts, and young age
are all associated with increased risk of prison suicide (Fazel, Cartwright,
Norman-Nott, & Hawton, 2008; Kaba et al., 2014), and they are also associated
with increased risk of placement in restrictive housing. Some authors have
attempted to disentangle these factors, finding that placement in restrictive
housing does independently increase suicide risk (Kaba et al., 2014).

Psychiatric Hospitalization
Very little evidence about psychiatric hospitalization rates for prisoners placed
in restrictive housing has been published. One Danish study (Sestoft, Andersen,
Lilleback, & Gabrielsen, 1998) found that individuals who remained in restrictive
housing longer than four weeks were 20 times more likely to be hospitalized
for psychiatric reasons, compared with non-restricted prisoners. A study of
prisoners in Marion, Illinois, found a much different result; only 3.1 percent of
the prisoners in restrictive housing were transferred to a psychiatric hospital over
a 10-year period (Ward & Werlich, 2003).

216 • National Institute of Justice | NIJ.gov

Institutional Misconduct
A comprehensive review of the relationship between restrictive housing and
institutional misconduct has recently been published (Labrecque, 2015). It is
mentioned briefly here because misconduct in prisoners can sometimes be a
proxy for psychological health. When experiencing symptoms of mental illness,
some prisoners turn inward, exhibiting withdrawal and self-injury. Others
turn outward, becoming hostile, aggressive, or violent. If restrictive housing
worsens prisoners’ psychological health, one would expect rates of institutional
misconduct after placement in restrictive housing to increase. Labrecque’s study
did not support this hypothesis, finding no increase in the rates of violent,
nonviolent, or drug-related misconduct after placement in restrictive housing.
Morris (2016) reached a similar conclusion using a different method, finding no
difference in the rates of violent misconduct between prisoners who had been
placed in short-term (15 days) restrictive housing and a control group.

Long-Term Psychological Effects
Most of the literature about the long-term psychological effects of restrictive
housing is descriptive or biographical, painting compelling portraits of
individuals who were fundamentally altered by solitary confinement and bear
deep scars from the experience long after it has ended. Few published studies
have systematically addressed this topic. Grassian’s sample of 14 prisoners (1983)
reported that their symptoms resolved rapidly after release from confinement,
but Grassian has also described longer lasting effects from restrictive housing
(2006). Valera and Kates-Benman (2015) performed a qualitative study of men
in the community who had spent time in restrictive housing, finding that most
of them described “getting used to it” over time. Presently, there are no published
studies that answer such important questions as whether prisoners who spent
time in restrictive housing develop PTSD as a result of the experience. Likewise,
no studies address whether restrictive housing prisoners experience long-term
changes in psychosocial functioning following release into the community (e.g.,
getting a job, reconnecting with friends and family, finding stable housing).
However, some authors have examined criminal justice outcomes, finding
preliminary evidence that prisoners who were placed in restrictive housing have
higher rates of recidivism (Motiuk & Blanchette, 2001; Lovell, Johnson, & Cain,
2007; Smith, Gendreau, & Labrecque, 2015).

Consensus of Mental Health Professional Organizations
Several organizations of healthcare professionals have published position
statements on the placement of prisoners with mental illness in restrictive
housing. These include the American Psychiatric Association (APA, 2012),
American Academy of Child & Adolescent Psychiatry (AACAP, 2012), American

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 217

College of Correctional Physicians (ACCP, 2013) (formerly the Society of
Correctional Physicians), and American Public Health Association (APHA,
2013). In addition, the APA updated its guidelines, Psychiatric Services in
Correctional Facilities, in 2015, and included a section on mental illness and
segregation. The National Commission on Correctional Health Care (NCCHC)
published a position statement on solitary confinement in April 2016 that
expands on its Standards for Mental Health Services in Correctional Facilities
(2015), the Commission’s guidelines for managing segregated prisoners.
The position statements and guidelines address different aspects of restrictive
housing, but they all agree that the practice places prisoners at risk, and care must
be taken to protect their health and well-being. Notably, the statements do not
call for abolishing restrictive housing altogether. They are fairly conservative in
their approach, focusing on the exclusion of particularly vulnerable populations
— juveniles and those with serious mental illness — and limiting the amount of
time prisoners spend in isolation. The American Psychological Association and
National Association of Social Workers (NASW), the professional organizations
whose members perform the bulk of prison mental health care, have not published
official positions on restrictive housing, though they have provided testimony
before Congress and published articles raising concerns about its use (American
Psychological Association, 2012; NASW, 2014). Some scholars have suggested that
mental health professionals have not gone far enough (Appelbaum, 2015), and they
should join the numerous advocacy groups involved in the movement to abolish
solitary confinement. To date, no mental health organizations have done so, though
the NCCHC comes close, stating that placement in isolation for more than 15 days
is cruel, inhumane, and degrading (2016).
Current recommendations from mental health professional organizations’
position statements and published guidelines include the following:
• Mental health professionals should have input into the prison disciplinary
process. ACCP (2013) and APA (2016) agree that prisoners should not be
placed in isolation as a punishment for behavior that is solely the result of
mental illness. Mental health professionals can inform the disciplinary process
about mitigating factors and, in some cases, divert prisoners from entering
disciplinary segregation by referring them instead to mental health housing or
other therapeutic settings.
• All prisoners being considered for restrictive housing placement should
be screened for mental health conditions that contraindicate placement
or require accommodation. The NCCHC takes somewhat contradictory
positions on this issue. The mental health standards it published in 2015
recommend reviewing the prisoner’s medical record prior to placement in
restrictive housing. However, a more recent position statement from the
organization indicates that “health staff must not be involved in determining
whether adults or juveniles are physically or psychologically able to be placed

218 • National Institute of Justice | NIJ.gov

in isolation” (NCCHC, 2016). APA (2016), on the other hand, states that acutely
suicidal or psychotic prisoners should not be placed in restrictive housing, and
APHA (2013) recommends that isolating prisoners for therapeutic purposes
should occur only when ordered by a health care professional.
• Individuals with serious mental illness should be excluded from prolonged
confinement in restrictive housing. Of the mental health organizations, APHA
and NCCHC take the most expansive position, with both calling for exclusion
of individuals with a serious mental illness from solitary confinement. APA
and ACCP are more restrained, allowing for some individuals with a serious
mental illness to be placed in restrictive housing but stating that, except in
rare cases, they should not be kept in that setting beyond four weeks. ACCP
(2013) and APA (2016) define serious mental illness to include prisoners with
all psychotic disorders (schizophrenia, schizoaffective disorder, delusional
disorder, brief psychotic disorder, schizophreniform disorder, substanceinduced psychotic disorder, and unspecified schizophrenia-spectrum
disorder), bipolar disorders, and major depressive disorder. Other illnesses,
such as PTSD, dementia, and personality disorders, may be considered serious
if they cause significant functional impairment.
• Individuals with intellectual disabilities should be excluded from prolonged
confinement in restrictive housing. ACCP (2013) includes intellectual disability
(called by its older name, “mental retardation”) in its list of conditions that
should exclude an inmate from restrictive housing longer than four weeks. The
other organizations do not specifically comment on this population.
• Juveniles should be categorically excluded from prolonged restrictive
housing. NCCHC, APHA, and AACAP (2012) recommend that individuals
younger than age 18 should not be placed in restrictive housing. The other
organizations do not comment on this population.
• Individuals in restrictive housing should have access to necessary mental
health treatment. APA and NCCHC make clear that correctional facilities
remain responsible for meeting the serious medical and mental health needs
of prisoners held in restrictive housing. This includes access to medication,
psychiatric assessments, and counseling. NCCHC provides guidelines (2015)
about how significant mental health findings should be documented and
conveyed to custody officials (when necessary). If prisoners with a serious
mental illness are kept in restrictive housing, APA (2016) recommends that
they be provided with 10 hours a week of unstructured activity in addition
to the necessary out-of-cell therapeutic activities. The NCCHC position
statement (2016) indicates that inmates in isolation are entitled to health care
that is consistent with community standards.
• All individuals in restrictive housing should be monitored closely by mental
health professionals. APA (2016) recommends regular rounds by a qualified

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 219

mental health professional (generally defined as a psychologist, psychiatrist,
psychiatric nurse, or social worker). NCCHC (2015) recommends that this
monitoring should occur at a frequency based on the degree of isolation:
daily for inmates in extreme isolation (those with little or no contact with
other individuals), every three days for those with limited contact with other
individuals, and weekly for those allowed routine social contact with other
inmates while remaining separated from the general population (e.g., inmates
in protective custody). NCCHC (2016) recommends monitoring at least on a
daily basis.
• Restrictive housing policies and procedures should allow prisoners with
acute mental health needs to be transferred to an appropriate treatment
setting. NCCHC (2015, 2016) and APA (2016) recommend that, when mental
health professionals identify signs of deteriorating mental health in prisoners,
they should communicate these findings to custody officials promptly. They
should also take steps to meet the prisoner’s therapeutic needs, including
transfer to a different setting if necessary.
• Correctional systems should develop alternatives to prolonged restrictive
housing. APA (2016) recognizes that alternatives to restrictive housing are
limited at this time, and the scientific data about its psychological effects
are rapidly evolving. NCCHC (2015) recommends that mental health
professionals should keep custody officials informed about the latest scientific
information and work with them to develop and evaluate alternatives to
restrictive housing, particularly for prisoners with a serious mental illness.

Knowledge Gaps
As noted in the “Literature Review” section above, the current literature
about mental health and restrictive housing leaves many important questions
unanswered. In fact, there are very few areas in which the data are clear and
compelling. It is clear that prisoners in restrictive housing are more disturbed
than the general prison population, with higher rates of diagnosed mental illness
and more severe symptoms. It is also clear that suicides occur disproportionately
in restrictive housing settings, both because higher-risk prisoners are placed
there and because of additional risks conferred by the setting. Finally, there is no
convincing evidence that restrictive housing provides any therapeutic benefit,
with many studies finding psychological harm and the Colorado study (O’Keefe
et al., 2013) concluding that prisoners with mental illness in that setting may
recover less rapidly than their peers in general prison population.
The finer points of the harm-vs.-benefit debate about restrictive housing are
still a gray area. For example, even if one accepts that restrictive housing has a
small but significant negative psychological impact (as the recent meta-analyses
suggest), it is not known which particular conditions of confinement are most

220 • National Institute of Justice | NIJ.gov

implicated: social isolation? noise and light levels? poor staff attitudes? The
relationship between these individual factors and psychological outcomes simply
has not been studied systematically, leaving a major gap in our understanding
about restrictive housing.
Another significant area for future study is the effect that access to mental health
programming has on psychological outcomes in restrictive housing. Some states,
such as Mississippi, Michigan, and Maine (Kupers et al., 2009; Chammah, 2016),
have created step-down programs for prisoners transitioning between restrictive
housing and general population, but the psychological impact of these programs
has not been systematically assessed. Likewise, some states are beginning to
implement the American Psychiatric Association’s recommendation to provide
prisoners who have a serious mental illness with at least 10 hours a week of
out-of-cell programming, but the effect has not yet been evaluated. This is a key
component for future research, as data about outcomes will help guide future
policy decisions.
Finally, further study of the long-term psychological effects of restrictive housing
is necessary. There is essentially no data about how prisoners released from
restrictive housing fare once they are released into the community. Do they have
difficulty, as the anecdotal literature suggests, reintegrating with society? Do
they develop higher rates of PTSD than prisoners who were not in restrictive
housing? And does release from restrictive housing straight into the community
— as one might hypothesize — cause greater psychological distress than a
gradual transition from restrictive housing to general population and then to the
community? All of these questions should be studied, as discussions about risks
and benefits of restrictive housing should not be limited to its immediate effects.

Policy Implications: Reconciling Research With Real Life
The national debate about the psychological effects of solitary confinement is
sometimes framed as a “chicken and egg” question: Are people with mental
illness preferentially placed in solitary confinement, or does solitary confinement
cause mental illness? This question does not necessarily have one answer; both
statements can be true. Indeed, with more knowledge about restrictive housing,
the corrections and mental health fields are beginning to see that both statements
are true. Individuals with mental illness break institutional rules and engage
in disruptive behaviors, causing them to be placed in isolation at greater rates
than individuals without mental illness. Once in isolation, they may deteriorate
further, developing increased symptoms of anxiety, aggression, and self-injury.
Of course, not all individuals react the same way to the conditions of restrictive
housing. Human beings display great variation in their responses to any
environmental stimulus, so why would restrictive housing be any different? Some
prisoners may prefer to be in the less-stimulating conditions of confinement,

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 221

finding the decreased interpersonal contact comforting. Others will crave human
contact and seek it out, sometimes in maladaptive ways such as self-injury or
destruction of property. It should come as no surprise that researchers using
different study groups and different methods have reached different conclusions
about how prisoners respond to restrictive housing, as they may simply be
accurately reporting about one small part of a complex whole.
When considering how individuals respond to restrictive housing, it is helpful to
conceptualize prison coping skills in a hierarchical manner. Figure 1 illustrates
some examples of how inmates cope with being in prison, moving from healthy
skills (top of figure) to unhealthy skills (bottom).
When prisoners are placed in restrictive housing, it is often because they are
using coping strategies in the middle of the hierarchy: engaging in sexual
relationships, maintaining gang affiliations, or using illicit drugs. They no
longer have access to those coping mechanisms after being placed in restrictive
housing. Under their new circumstances, some are able to move up in the
hierarchy of coping by writing letters, drawing, or working on legal challenges
to their conditions of confinement. Other prisoners are not able to muster

Figure 1. Hierarchy of Coping Skills in Prison
School

Jobs Therapy Vocational training
Creative arts
Religion
Visits and letters to family Prison journalism

Preparing for reentry,
maintaining connections
with outside world

Lawsuits
Organized protests
Hunger strikes

Changing conditions
of confinement

Sexual relationships Commissary
Gang affiliation Television
Illicit drug use

Easing the time,
creating relationships
inside prison

Cutting Flooding
Inserting
Exposing
Feces smearing

Masochism,
provocation, sadism

Regression into
primitive behaviors

Suicide
Self-annihilation

222 • National Institute of Justice | NIJ.gov

the psychological resources they need. Instead, they move downward in the
hierarchy to much more regressed behaviors: cutting themselves, flooding their
cells, inserting objects into body cavities, and making suicide attempts.
Unfortunately, neither mental health clinicians nor prison officials have a
reliable method of determining in advance which prisoners will do well in
isolation and which will not. The risk created by this limitation is substantial,
and prisoners may be harmed. This risk of harm, combined with the lack of
convincing evidence that restrictive housing achieves greater safety and security,
requires serious consideration about whether solitary confinement (at least for
the purpose of administrative segregation or punishment) serves any useful
purpose. Changes to restrictive housing practices will not happen overnight, but
substantial reform is encouraged.

Research Directions
As described in the “Knowledge Gaps” section above, further study of many
aspects of restrictive housing is necessary. Several high-priority areas are
suggested below:
• Future research about restrictive housing should be conducted in accordance
with established scientific principles, with clearly delineated methods,
variables, and outcome measures.
• Once outcome measures are defined, the characteristics of prisoners and
characteristics of confinement that result in particular outcomes (both
positive and negative) should be studied and delineated.
• The effect of mental health treatment and out-of-cell programming on
the psychological symptoms and psychosocial functioning of prisoners in
restrictive housing should be studied systematically.
• Prisoners placed in restrictive housing should be evaluated for any long-term
psychological and functional outcomes of this housing.

Clinical Practice
The practice guidelines established by the American Psychiatric Association,
American College of Correctional Physicians, National Commission on
Correctional Health Care, American Academy of Child & Adolescent Psychiatry,
and American Public Health Association should be supported. In addition,
numerous factors not raised in the professional organization guidelines but
supported by research are also important:

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 223

• Efforts should be made to ensure that interactions between staff and prisoners
in restrictive housing settings are fair and respectful. Humiliation and
degradation should not be part of the experience.
• Restrictive housing programs should include both positive and negative
incentives for prisoners who change their behavior in positive ways;
such programs should not rely solely on punishment and deprivation as
management tools.

Systems Change
Restrictive housing units often serve as the “treatment setting” of last resort
in correctional systems without adequately developed mental health systems.
Therefore, the following changes to correctional systems are recommended:
• Alternatives to restrictive housing units should be developed, particularly for
prisoners with mental illness. Adequate funding should be allocated for their
design, implementation, and evaluation.
• Prison systems should implement evidence-based affirmative programming
that develops prosocial skills in prisoners.
• Correctional staff should be trained in techniques for preventing or defusing
critical situations that would otherwise lead to placing prisoners into
restrictive housing.
• Access to mental health services should be expanded for all prisoners, as
providing proactive treatment has the potential to decrease behaviors that
result in restrictive housing placement.

Conclusion
The relationship between restrictive housing and mental health is complex,
with many more nuances than are initially apparent to the casual reader of the
solitary confinement literature. Many questions remain unanswered. However,
this literature review has raised enough questions about the psychological
effects of restrictive housing to warrant a large-scale reassessment of our current
correctional practices. In particular, the disproportionate number of prisoners
with mental illness who are placed in restrictive housing is troubling, and the
setting itself confers significant risk of suicide and self-injury. Even if scholars
disagree about how and why these poor outcomes occur, no one can deny
that they do occur. To move forward, corrections officials and mental health
professionals must work together to create systems of care that improve the
health of prisoners while also maintaining institutional safety.

224 • National Institute of Justice | NIJ.gov

To their credit, a number of state correctional systems have begun the difficult
process of developing and implementing alternatives to traditional restrictive
housing practices, particularly for individuals with mental illness. Systematic
assessment of these new programs is critical. Using established scientific methods
to conduct the assessments is essential to minimize the potential for bias or
error, particularly in an area as controversial as restrictive housing. Through
this combination of clinical innovation and rigorous scientific investigation of
outcomes, real progress in the field is possible.

References
American Academy of Child & Adolescent Psychiatry. (2012). Solitary
confinement of juvenile offenders. Retrieved January 10, 2016, from http://www.
aacap.org/AACAP/Policy_Statements/2012/Solitary_Confinement_of_Juvenile_
Offenders.aspx.
American Civil Liberties Union. (2014). Briefing paper: The dangerous overuse of
solitary confinement in the United States. Retrieved January 16, 2016, from https://
www.aclu.org/report/dangerous-overuse-solitary-confinement-united-states.
American College of Correctional Physicians. (2013). Restricted
housing of mentally ill inmates. Retrieved January 16, 2016, from http://
societyofcorrectionalphysicians.org/resources/position-statements/restricted­
housing-of-mentally-ill-inmates.
American Psychiatric Association. (1980). Diagnostic and statistical manual of
mental disorders (3rd ed.). Washington, DC: Author.
American Psychiatric Association. (2012). Position statement on segregation of
prisoners with mental illness. Retrieved January 10, 2016, from http://www.dhcs.
ca.gov/services/MH/Documents/2013_04_AC_06c_APA_ps2012_PrizSeg.pdf.
American Psychiatric Association. (2013). Diagnostic and statistical manual of
mental disorders (5th ed.). Washington, DC: Author.
American Psychiatric Association. (2016). Psychiatric services in correctional
facilities (3rd ed.). Arlington, VA: APA.
American Psychological Association. (2012, October). Psychologist testifies on
the risks of solitary confinement. Monitor on Psychology, 43, 10.
American Public Health Association. (2013). Solitary confinement as a public
health issue. Retrieved January 10, 2016, from http://apha.org/policies-and­
advocacy/public-health-policy-statements/policy-database/2014/07/14/13/30/
solitary-confinement-as-a-public-health-issue.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 225

Andersen, H. S., Sestoft, D., Lillebaek, T., Gabrielsen, G., Hemmingsen, R., & Kramp,
P. (2000). A longitudinal study of prisoners on remand: Psychiatric prevalence,
incidence, and psychopathology in solitary vs. non-solitary confinement. Acta
Psychiatrica Scandinavica, 102, 19-25.
Andersen, H. S., Sestoft, D., Lillebaek, T., Gabrielsen, G., Hemmingsen, R.
(2003). A longitudinal study of prisoners on remand: Repeated measures of
psychopathology in the initial phase of solitary versus nonsolitary confinement.
International Journal of Law and Psychiatry, 26, 165-177.
Appelbaum, K. (2015). American psychiatry should join the call to abolish
solitary confinement. Journal of the American Academy of Psychiatry and the Law,
43, 406-15.
Arrigo, B. A., & Bullock, J. L. (2008). The psychological effects of solitary
confinement on prisoners in supermax units: Reviewing what we know and
recommending what should change. International Journal of Offender Therapy
and Comparative Criminology, 52(6), 622-640.
Bonta, J., & Gendreau, P. (1990). Reexamining the cruel and unusual punishment
of prison life. Law and Human Behavior, 14(4), 347-372.
Brodsky, S. L., & Scogin, F. R. (1988). Inmates in protective custody: First data on
emotional effects. Forensic Reports, 1, 267-280.
Bureau of Justice Statistics. (2002). Mental health and treatment of inmates and
probationers. Washington, DC: U.S. Department of Justice.
Bureau of Justice Statistics. (2014, September). Prisoners in 2012: Trends in
admissions and releases, 1991-2012 (revised). Retrieved January 10, 2016, from
http://www.bjs.gov/index.cfm?ty=pbdetail&iid=4842.
Carriere, K. D. (1989). Protective custody in Canada: A review of research and
policy responses. Canadian Criminology Forum, 10, 17-25.
Chammah, M. (2016, January 6). Stepping down from solitary confinement.
The Atlantic. Retrieved January 17, 2016, from http://www.theatlantic.com/
politics/archive/2016/01/solitary-confinement-reform/422565/.
Childress, S. (2014). The disturbing sounds of solitary confinement. Frontline.
Retrieved from http://www.pbs.org/wgbh/frontline/article/the-disturbing­
sounds-of-solitary-confinement/.
Cloyes, K. G., Lovell, D., Allen, D. G., & Rhodes, L. A. (2006). Assessment of
psychosocial impairment in a supermaximum security unit sample. Criminal
Justice and Behavior, 33(6), 760-781.

226 • National Institute of Justice | NIJ.gov

Commonwealth of Pennsylvania Department of Corrections. (2015). Access
to mental health care (Policy 13.8.1). Retrieved from http://www.cor.pa.gov/
Administration/Documents/DOC%20Policies/13.08.01%20Access%20to%20
Mental%20Health%20Care.pdf.
Correctional Association of New York. (2004). Mental health in the house of
corrections. Retrieved from http://www.correctionalassociation.org/wp-content/
uploads/2004/06/Mental-Health.pdf.
Ecclestone, C. E. J., Gendreau, P., & Knox, C. (1974). Solitary confinement of
prisoners: An assessment of its effects on offenders’ personal constructs and
andrenocortical activity. Canadian Journal of Behavioral Science, 6, 178-191.
Fazel, S., Cartwright, J., Norman-Nott, A., & Hawton, K. (2008). Suicide in prisoners:
A systematic review of risk factors. Journal of Clinical Psychiatry, 69, 1721–1731.
Frost, N. A. & Monteiro, C. E. (2016). Administrative segregation in prison.
Washington, DC: U.S. Department of Justice, National Institute of Justice.
Gawande, A. (2009, March 30). Hellhole. New Yorker, 36-45. Retrieved from
http://www.newyorker.com/magazine/2009/03/30/hellhole.
Gendreau, P., Freedman, N. L., Wilde, G. J. S., & Scott, G. D. (1972). Changes in
EEG alpha frequency and evoked response latency during solitary confinement.
Journal of Abnormal Psychology, 79, 54-59.
Gendreau, P., & Bonta, J. (1984). Solitary confinement is not cruel and unusual
punishment: People sometimes are! Canadian Journal of Criminology, 26, 467-478.
Gendreau, P., & Labrecque, R. M. (2015). The effects of administrative
segregation: A lesson in knowledge cumulation. In J. Wooldredge & P.
Smith (Eds.), Oxford handbook on prisons and imprisonment. New York, NY:
Oxford University Press. Retrieved from http://media.wix.com/ugd/7fc458_
edb83bc086d64e3bb8cb21f56e8ce98a.pdf.
Goldberger, L. (1966). Experimental isolation: An overview. American Journal of
Psychiatry, 122, 774-782.
Grassian, S. (1983). Psychopathological effects of solitary confinement. American
Journal of Psychiatry, 140, 1450-1454.
Grassian, S. (2006). Psychiatric effects of solitary confinement. Journal of Law &
Policy, 22, 325-383.
Grassian, S., & Friedman, N. (1986). Effects of sensory deprivation in psychiatric
seclusion and solitary confinement. International Journal of Law and Psychiatry,
8, 49-65.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 227

Guenther, L. (2012, August 12). The living death of solitary confinement. New
York Times. Retrieved from http://opinionator.blogs.nytimes.com/2012/08/26/
the-living-death-of-solitary-confinement/?_r=0.
Haney, C. (2003). Mental health issues in long-term solitary and “supermax”
confinement. Crime and Delinquency, 49(1), 124-156.
Haney, C., & Lynch, M. (1997). Regulating prisons of the future: A psychological
analysis of supermax and solitary confinement. New York University Review of
Law and Social Change, 23, 477-570.
Hodgins, S., & Cote, G. (1991). The mental health of penitentiary inmates in
isolation. Canadian Journal of Criminology, 33, 175-182.
Human Rights Watch. (2015, May). Callous and cruel: Use of force against inmates
with mental disabilities in U.S. jails and prisons. Retrieved January 16, 2016, from
https://www.hrw.org/sitesearch/Callous%20and%20cruel%3A%20Use%20of%20
force%20against%20inmates.
Jackson, M. (1983). Prisoners of isolation: Solitary confinement in Canada.
Toronto, ON: University of Toronto Press.
Kaba, F., Lewis, A., Glowa-Kollisch, S., Hadler, J., Lee, D., Alper, H., Selling, D.,
MacDonald, R., Solimo, A. A., & Venters, H. (2014). Solitary confinement and
risk of self-harm among jail inmates. American Journal of Public Health, 104,
442-447.
Kaba, F., Solimo, A., Graves, J., Glowa-Kollisch, S., Vise, A., MacDonald, R.,
Waters, A., ... Venters, H. (2015). Disparities in mental health referral and
diagnosis in the New York City jail mental health service. American Journal of
Public Health, 105(9), 1911-1916.
Korn, R. (1988). The effects of confinement in the high security unit at
Lexington. Social Justice, 15(1), 8-19.
Kupers, T. A., Dronet, T., Winter, M., Austin, J., Kelly, L., Cartier, W., Morris, T.
J., Hanion, S. F., Sparkman, E. L., Kumar, P., Vincent, L. C., Norris, J., Nagel, K.,
& McBride, J. (2009). Beyond supermax administrative segregation: Mississippi’s
experience rethinking prison classification and creating alternative mental health
programs. Criminal Justice and Behavior, 36, 1037-1050.
Lamb, H. R., & Weinberger, L. E. (1998). Persons with severe mental illness in
jails and prisons: A review. Psychiatric Services, 49, 483-492.
Labrecque, R. L. (2015). The effect of solitary confinement on institutional
misconduct: A longitudinal evaluation (Doctoral dissertation). Retrieved from
http://media.wix.com/ugd/7fc458_1e24bb04be834ad2ab0368b06448f7a0.pdf.

228 • National Institute of Justice | NIJ.gov

Lanes, E. C. (2011). Are the “worst of the worst” self-injurious prisoners more
likely to end up in long-term maximum-security administrative segregation?
International Journal of Offender Therapy and Comparative Criminology, 55(7),
1034-1050.
Liman Program & Association of State Correctional Administrators. (2015).
Time-in-cell: ASCA-Liman 2014 National Survey of Administrative Segregation in
Prison. New Haven, CT: Yale Law School.
Lovell, D. (2008). Patterns of disturbed behavior in a supermax prison.
Criminal Justice and Behavior, 35, 985-1004.
Lovell, D., Cloyes, K., Allen, D. G., & Rhodes, L. A. (2000). Who resides in
supermax prisons? A Washington state study. Federal Probation, 64(2), 33-38.
Lovell, D., Johnson, L. C., & Cain, K. C. (2007). Recidivism of supermax
prisoners in Washington state. Crime & Delinquency, 53(4), 633-56.
Martel, J. (1999). Solitude and cold storage: Women’s journeys of endurance in
segregation. Edmonton, AB, Canada: ACI Communication.
Metzner, J. L. (2002). Class action litigation in correctional psychiatry. Journal of
the American Academy of Psychiatry and the Law, 30, 19-29.
Metzner, J. L. (2015). Mental health considerations for segregated inmates.
Retrieved from www.ncchc.org.
Miller, H. (1994). Reexamining psychological distress in the current conditions
of segregation. Journal of Correctional Health Care, 1, 39-53.
Miller, H. A., & Young, G. R. (1997). Prison segregation: Administrative
detention remedy or mental health problem? Criminal Behaviour and Mental
Health, 7(1), 85-94.
Morgan, R. D., Gendreau, P., Smith, P., Gray, A. L., Labrecque, R. M., MacLean,
N., ... Mills, J. F. (2016, August 8). Quantitative syntheses of the effects of
administrative segregation on inmates’ well-being. Psychology, Public Policy, and
the Law. Advance online publication. Retrieved from http://dx.doi.org/10.1037/
law0000089.
Morris, R. G. (2016). Exploring the effect of exposure to short-term solitary
confinement among violent prison inmates. Journal of Quantitative Criminology,
32(1), 1-22. http://dx.doi.org/10.1007/s10940-015-9250-0.
Motiuk, L. L. & Blanchette, K. (1997). Case characteristics of segregated prisoners
in federal corrections. Ottawa, ON: Correctional Service of Canada.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 229

Motiuk, L., L. & Blanchette, K. (2001). Characteristics of administratively
segregated offenders in federal corrections. Canadian Journal of Criminology, 43,
131-143.
National Association of Social Workers. (2014). A social work response to solitary
confinement. Retrieved from http://www.socialworkblog.org/advocacy/2014/11/
a-social-work-response-to-solitary-confinement/.
National Commission on Correctional Health Care. (2015). Standards for mental
health services in correctional facilities. Chicago, IL: NCCHC.
National Commission on Correctional Health Care. (2016, April 10). Solitary
confinement (isolation). Retrieved from http://www.ncchc.org/solitary-confinement.
Patterson, R., & Hughes, K. (2008). Review of completed suicides in the
California Department of Corrections and Rehabilitation, 1999 to 2004.
Psychiatric Services, 59, 676–682.
Obama, B. (2016, January 25). Why we must rethink solitary confinement.
The Washington Post. Retrieved from https://www.washingtonpost.com.
O’Keefe, M. L. (2007). Administrative segregation for mentally ill inmates.
Journal of Offender Rehabilitation, 45, 149-165.
O’Keefe, M. L., Klebe, K. J., Metzner, J., Dvoskin, J., Fellner, J., & Stucker, A.
(2013). A longitudinal study of administrative segregation. Journal of the
American Academy of Psychiatry and the Law, 41, 49-60.
Reeves, R., & Tamburello, A. (2014). Single cells, segregated housing, and suicide
in the New Jersey Department of Corrections. Journal of the American Academy
of Psychiatry and the Law, 42, 484-488.
Roma, P., Pompili, M., Lester, D., Girardi, P., & Ferracuti, S. (2013). Incremental
conditions of isolation as a predictor of suicide in prisoners. Forensic Science
International, 233(1), e1-e2.
Scharff-Smith, P. (2006). The effects of solitary confinement on prison inmates:
A brief history and review of the literature. Crime and Justice, 34(1), 441-528.
Sestoft, D. M., Andersen, H. S., Lillebaek, T., & Gabrielsen, G. (1998). Impact
of solitary confinement on hospitalization among Danish prisoners in custody.
International Journal of Law & Psychiatry, 21, 99-108.
Shalev, S. (2008). A sourcebook on solitary confinement. London, England:
London School of Economics, Mannheim Centre for Criminology.

230 • National Institute of Justice | NIJ.gov

Smith, P., Labrecque, R., & Gendreau, P. (2015). An evaluation of the impact of
solitary confinement on offenders with mental illness. Paper presented at the
International Association of Law and Mental Health XXXIVth Congress, July
2015. Retrieved from http://www.ryanmlabrecque.com/#!presentations/c14lj.
Suedfeld, P., Ramirez, C., Deaton, J., Baker-Brown, G. (1982). Reactions and
attributes of prisoners in solitary confinement. Criminal Justice and Behavior,
9(3), 303-340.
Suedfeld, P., Roy, C. (1975). Using social isolation to control the behavior of
disruptive inmates. International Journal of Offender Therapy and Comparative
Criminology, 19, 90–99.
Thompson, J., & Rubenfeld, S. (2013). A profile of women in segregation. Research
report R-320. Ottawa, ON: Correctional Service of Canada.
United Nations General Assembly. (1990). Rules for the protection of juveniles
deprived of their liberty. Retrieved from http://www.un.org/documents/ga/res/45/
a45r113.htm.
United Nations General Assembly. (2011). Torture and other cruel, inhuman or
degrading treatment or punishment. Retrieved from http://solitaryconfinement.
org/uploads/SpecRapTortureAug2011.pdf.
United States Department of Justice. (2016). Report and recommendations
concerning the use of restrictive housing. Retrieved from https://www.justice.gov/
dag/file/815551/download.
Valera, P., & Kates-Benman, C.L. (2015, March 10). Exploring the use of
special housing units by men released from New York correctional facilities: A
small mixed-methods study. American Journal of Men’s Health. http://dx.doi.
org/10.1177/1557988315569880.
Walters, R. H., Callagan, R. H., & Newman, A., F. (1963). Effect of solitary
confinement on prisoners. American Journal of Psychiatry, 19(8): 771-773.
Ward, D. A., & Werlich, T. G. (2003). Alcatraz and Marion: Evaluating supermaximum custody. Punishment & Society, 5(1), 53-75.
Way, B. B., Miraglia, R., Sawyer, D., Beer, R., & Eddy, J. (2005). Factors related
to suicide in New York state prisons. International Journal of Law & Psychiatry,
28, 207-221.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 231

Way, B. B., Sawyer, D. A., Barboza, S., & Nash, R. (2007). Inmate suicide and
time spent in special disciplinary housing in New York State prison. Psychiatric
Services, 58(4), 558-560.
White, T. W., Schimmel, D. J., & Frickey, R. (2002). A comprehensive analysis of
suicide in federal prisons: A fifteen-year review. Journal of Correctional Health
Care, 9(3), 321-343.
Wormith, J. S., Tellier, M., & Gendreau, P. (1988). Characteristics of protective
custody offenders in a provincial correctional center. Canadian Journal of
Criminology, 30, 39-58.
Wynn, J. R., & Szatrowski, A. (2004). Hidden prisons: Twenty-three-hour
lockdown units in the New York state correctional facilities. Pace Law Review,
24(2), 497-525.
Zinger, I., Wichmann, C., & Andrews, D. A. (2001). The psychological effects of 60
days in administrative segregation. Canadian Journal of Criminology, 43(1), 47-83.
Zubek, J. P., Bayer, L., & Shephard, J. M. (1969). Relative effects of prolonged
social isolation and confinement: Behavioral and EEG changes. Journal of
Abnormal Psychology, 74, 625-631.

232 • National Institute of Justice | NIJ.gov

CHAPTER 7

Critical Research Gaps in
Understanding the Effects of
Prolonged Time in Restrictive
Housing on Inmates and the
Institutional Environment
Daniel P. Mears, Ph.D.
Florida State University College of Criminology and Criminal Justice

Introduction

T

his white paper reviews the state of research on the effects of prolonged
time in restrictive housing on inmates and the institutional environment
of prison systems. It then identifies prominent research gaps that warrant
attention in efforts to guide future policy. The summary assessment is that
there exists too little credible research to state with confidence the effectiveness
of restrictive housing. The practice may produce benefits, such as improved
inmate behavior and reduced recidivism. It also may produce harms, including
worsened inmate behavior and mental health and higher rates of recidivism.
Additional benefits and harms may exist, but extant research has not provided a
consistent scientific basis for establishing them.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 233

What is the context for evaluating the effects of restrictive housing? Prison
systems seek to incapacitate and rehabilitate individuals who have committed
crimes and to do so in a way that protects society, prison personnel, and inmates.
Achieving these goals is challenging. Those who go to prison do not want to
be there and may be prone to violence and violating rules. They also may act
violently or break rules in response to prison conditions. How prisons are run
may affect inmates as well. In addition, the prison system may implement rules,
programs, services, vocational training, educational training, and other measures
that may decrease or increase misconduct. When prison systems work well,
inmates, officers, and other staff are safe; programming, services, and treatment
are provided; and inmates returned to society are less likely to reoffend. When
these systems do not work well, inmates and officers are harmed; programming,
services, and treatment are minimal; and recidivism rates are high.
Corrections officials employ different strategies for promoting safety and order.
One strategy is restrictive housing, which typically involves confining inmates
who may pose a threat to prison safety and order alone in a single cell for
varying periods of time. Various terms have been used to capture this idea —
supermaximum (“supermax”) housing, administrative segregation, extended
isolation, restrictive housing, and so on. Inmates housed in this way may be
confined to their cells for up to 23 hours per day and have few privileges. This
housing is costly to build and operate, and yet may be critical to maintaining
safety and order. Alternatively, it may be ineffective, harmful, or less cost-efficient
than alternative strategies for managing inmates and prison systems.
The use of restrictive housing increased dramatically in recent decades. Its use
was justified in part by the idea that it would help manage disruptive and violent
inmates (King, 1999; Riveland, 1999; Briggs et al., 2003; Neal, 2003; Bruton, 2004;
Browne et al., 2011; Mears, 2013; Ross, 2013; Baumgartel et al., 2015; Richards,
2015). In this time, concerns have been raised about the need, effectiveness, or
appropriateness of this type of housing. These concerns have escalated in recent
years, centering on its potential harms and overuse. A recent national estimate
indicates that as many as one in five inmates in prison or jail spend time in
restrictive housing (Beck, 2015:1). Members of the U.S. Supreme Court and
Congress have called for reforms, as have the President and the Federal Bureau of
Prisons. In response, state departments of corrections have begun reviewing their
policies for placing inmates in restrictive housing (Durbin, 2012, 2013; Goode,
2012; Mears, 2013; U.S. Government Accountability Office, 2013; McGinnis
et al., 2014; Baker & Goode, 2015; Baumgartel et al., 2015; Obama, 2016; U.S.
Department of Justice Work Group, 2016). Central to discussions about this
type of prison housing are questions about the conditions under which it is
appropriate (leading to improved outcomes) and the conditions under which
it is inappropriate (potentially leading to adverse outcomes).

23 4 • National Institute of Justice | NIJ.gov

Table 1. The State of Research on Restrictive Housing’s Impact on Inmates, Prisons, and Society
Question

Answer

1. Definition?

Varied; no clear consensus.

2. Goals?

Varied; no clear consensus.

3. Need?

Unclear.

4. Theory?

Unclear and varied.

5. Implementation?

Largely unknown.

6. Impacts?

Largely unknown; adverse effects on
inmate mental health seem likely.

7. Cost-efficiency?

Unknown relative to various other
approaches for achieving system goals.

8. Critical research gaps?

Many.

The limited evidence of restrictive housing’s effectiveness stems from the fact
that few credible studies of impact exist. It also stems from more fundamental
issues — a lack of consensus about the definition of such housing, ambiguity
about its goals, a lack of coherent theory underlying its use, and limited
assessment of how it is used. Clarity and information about each of these
dimensions is essential for understanding and evaluating the impact of
restrictive housing (e.g., goals dictate which outcomes matter). Accordingly,
this paper discusses these topics before reviewing what is known about the
impacts of restrictive housing and, in turn, the cost-efficiency of such housing relative
to other strategies for managing prisons and inmates. It then identifies critical gaps
in the present research and addresses questions that remain unanswered. Table 1
summarizes the state of research across all of these dimensions.
To arrive at this assessment, the paper discusses the following topics:
1. Definitional challenges associated with discussions of and research on
restrictive housing and, by extension, with evaluations of its impacts.
2. The goals of restrictive housing and their importance for conducting and
interpreting research on it.
3. The need for restrictive housing and the relevance of need for assessing impact.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 235

4. The theory behind such housing and its relevance for anticipating and
identifying its impact on inmates and their families, as well as on officers,
prisons, and communities.
5. The implementation of restrictive housing and its relevance for creating
beneficial or harmful impacts.
6. The impacts of restrictive housing on inmates and their families, as well as on
officers, prisons, and communities.
7. The cost-efficiency of restrictive housing.
8. Critical research gaps and questions that ideally will be addressed to advance
science, promote accountability, and place prison system management
strategies on a more evidence-based foundation for achieving a variety of
prison system goals.
The first five topics to be discussed dictate the outcomes that are relevant for
evaluating the impact of restrictive housing. When research is limited or absent for
any of these factors, there is a direct ripple effect for interpreting impact evaluation
findings. For example, if a study shows that a prison system consistently employs
restrictive housing for “nuisance” inmates, it would raise questions about how to
interpret a finding that such housing reduces prison system violence. With the
final three topics, the paper turns to discussing impacts, cost-efficiency, gaps in the
research, and questions that remain unanswered.
The methodology for the literature review was guided by several considerations.
First, given the broad scope, the paper focused on identifying general findings,
patterns, and issues that have been identified in the empirical literature on
restrictive housing. To this end, published empirical research, as well as reviews
of research, on various aspects of restricted housing were reviewed (e.g., who is
sent to such housing, its effects on recidivism). Second, primary emphasis was
given to works published in peer-reviewed academic journals, though reports
published by public and private research organizations were also reviewed. Third,
works were consulted that may not have specifically used the term “restrictive
housing” but nonetheless focused on what amounts to such housing. Fourth, the
review primarily examined empirical research that emphasized the long-term
effects of restrictive housing on inmates.

Definitions
Discussions of restrictive housing are plagued by inconsistencies in defining
exactly what constitutes such housing. This issue is described here to establish the
foundation on which any account of restrictive housing descriptions, estimates,
uses, goals, or impacts rests.

236 • National Institute of Justice | NIJ.gov

What Is Restrictive Housing?
“Restrictive housing” does not refer to a single type of correctional housing
design. Broadly, it involves the isolation of some inmates in single-cell
confinement with little or no access to services, programming, privileges, or
other people, regardless of the particular goal of such confinement. Additionally,
a variety of terms are used to describe that situation. In some cases, one term
appears to describe different types of housing and goals, while in other cases,
different terms may describe similar housing and goals. In addition, there can
be considerable variation in the precise conditions or experiences that constitute
restrictive housing.
Accordingly, before discussing the state of empirical evidence on restrictive
housing, this paper reviews the variation in terminology, goals, and the “nuts and
bolts” of what counts as restrictive housing. The bottom-line assessment, though,
is that differences in terminology make it difficult to arrive at clear or definitive
statements about the state of empirical research on restrictive housing.

The Problem of Varying Terminology
One source of confusion in discussing restrictive housing is the varying
terminology used by federal and state governments to describe what appears to
be the same thing — extended isolation in single-cell confinement with limited
access to programming, services, treatment, visitation, or the like, which has been
a mainstay of prison systems for many decades (National Institute of Corrections,
1997; Henningsen et al., 1999; King, 1999; Naday et al., 2008; Mears, 2013; Ross,
2013; Frost & Monteiro, 2015; Richards, 2015). This variation in terminology
may indicate important differences in the goals or design of specific housing. It
may also mask similarities. For example, a study undertaken at the behest of the
National Institute of Corrections found that states used the following terms to
describe what appeared to be similar housing: “special housing unit, maxi-maxi,
maximum control facility, secure housing unit, intensive management unit, and
administrative maximum penitentiary” (Riveland, 1999:5). That report described
all such housing as “supermax housing.” In the past two decades, increased
attention has turned to supermaximum (supermax) security housing, in large
part because many states used this terminology (Mears, 2006).
More recently, however, attention has centered on the term “solitary
confinement,” in part because of calls by some legislators for reforms in the
use of inmate isolation (Sen. Dick Durbin 2012, 2013; Shames et al., 2015).
Despite this shift toward “solitary confinement” and “isolation” terminology,
many policymaker, advocacy, and scholarly accounts appear to refer to the same
underlying phenomenon of extended stays in single-cell confinement for up to 23
to 24 hours per day. At the same time, there continue to be widespread references
to supermax housing and to various other terms, such as “administrative

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 237

segregation” and “segregated housing,” that denote extended isolation in such
confinement for months or years. The terms segregation, solitary confinement,
and isolation all underscore that a central feature of this type of housing
experience is that inmates typically reside alone in a cell, rarely leaving it except
for an hour or two each day, depending on the inmate’s behavior.
Most recently, another term — restrictive housing — has come into use (Kane
et al., 2013; McGinnis et al., 2014; Baumgartel et al., 2015; Beck, 2015). It, too,
is defined as an extended stay in, typically, single-cell confinement with limited
access to programming, services, treatment, and so on. However, this term has
the potential virtue of avoiding some ideological connotations that other terms,
such as supermax housing, may convey. In addition, the Association of State
Correctional Administrators (ASCA) uses this term and has emphasized that it
covers a range of housing conditions captured by various prison classification
systems, though it states that the term should not be used to include protective
custody (Association of State Correctional Administrators, 2013:1). Other
reports use restrictive housing to refer to administrative segregation but not to
segregation that is used for punishment or protective custody (Baumgartel et al.,
2015:11).
This definitional ambiguity is amplified by the lack of clear operational
definitions and by state administrators and researchers who use the different
terms interchangeably. It is amplified, too, by the fact that inmates may
experience lengthy stays in single-cell housing regardless of whether they were
placed there for protective custody, punishment, or management goals.

The Problem of Varying Goals
Another source of confusion is that the goals associated with restrictive housing
(or analogous terms) vary. The following three-category typology, based on
different goals, is frequently used to distinguish different types of restrictive
housing (Mears & Watson, 2006; Shalev, 2009; McGinnis et al., 2014; Shames
et al., 2015).
The first goal is restrictive housing to protect an inmate, which is known as
protective custody housing. It provides temporary protection of an inmate.
Typically, inmates would not be housed for an extended stay in isolation merely
for their protection; however, inmates may well serve many months in such
housing for this purpose.
The second goal is restrictive housing to punish an inmate, which is known as
punishment or disciplinary custody. Here, again, lengthy stays in isolation should
be unusual, but lengthy stays may occur. Prison system rules may prohibit stays
of more than a set period of time (e.g., 30 days), but exceptions exist. Inmates,
for example, might violate rules or act violently while in punitive segregation,
thereby activating a new or extended stay. Alternatively, inmates’ status might

238 • National Institute of Justice | NIJ.gov

change; they may transition from punitive segregation to segregation for the
purpose of achieving some management goal. Once again, the frequency of such
extended stays is largely unknown.
The third goal is restrictive housing to isolate inmates for some management
reason. There is no specific naming convention for this sort of isolation beyond
suggesting that it serves managerial purposes. Restrictive housing may be used,
for example, to manage the “worst of the worst” inmates (Henningsen et al.,
1999; King, 1999; Riveland, 1999; Mears, 2006, 2008b, 2013; Mears & Watson,
2006; Shalev, 2009; Butler et al., 2012; Richards, 2015). The meaning of “manage”
frequently is not clear. The term might refer to any of a range of possible goals. One
national study of supermax housing, for example, asked corrections officials and
wardens for their views of the goals of such housing. The respondents reported that
the goals included (Mears & Castro, 2006; Mears & Watson, 2006) —
• Increasing prison safety systemwide.
• Increasing prison order systemwide.
• Reducing the likelihood of prison riots.
• Incapacitating violent inmates.
• Improving the behavior of inmates who experience supermax confinement.
• Reducing the influence of gangs in prisons.
• Increasing public safety.

The Problem of Varying Designs
Yet another source of confusion in discussions about restrictive housing is
variation in the nuts-and-bolts design of the housing (National Institute of
Corrections, 1997; Riveland, 1999; King, 1999; Toch, 2003; Pizarro & Stenius,
2004; Naday et al., 2008; Shalev, 2009; Butler et al., 2012; Ross, 2013). The
following questions illustrate this confusion. What building structures are
considered restrictive housing — a separate housing unit, a wing of a building, a
set of designated restrictive housing cells? Is technology of some type required?
Does such housing always entail single-cell confinement or can it include
housing two inmates in the same cell for extended periods? Does the duration
of the stay matter? Does the inmate need to spend 23 hours per day in the cell,
or would 16 or 22 hours per day suffice? Does an individual need to be held
in such housing for a minimum number of days? Does the precise goal matter
in identifying housing? For example, does confining an inmate in isolation
for 36 hours to help him or her to “cool down” count as restrictive housing? Is
it necessary that inmates have limited privileges, including limited access to

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 239

programming, services, treatment, visitation, and so on? Just how intense do the
limitations need to be to create conditions that amount to restrictive housing?
No agreement exists about such issues. The reality is that restrictive housing
refers to a highly heterogeneous set of conditions. In many cases, the
heterogeneity may mask a core commonality: extended isolation with little or
no access to programming, services, treatment, or visitation. In other cases, the
heterogeneity may signal important nuances that differentiate the experience of
restrictive housing from one facility or state to another.

The Relevance of Varying Definitions, Goals, and Design
These sources of confusion have compromised the ability of different parties
— policymakers, practitioners, advocates, researchers, and so on — to talk
meaningfully or accurately about restrictive housing. In some cases, these groups
may be discussing the same phenomenon, but in others they may be discussing
different phenomena. Conversely, it may appear that the same phenomenon is
under discussion when in reality different phenomena are being compared. The
situation is unfortunate because definitions, goals, and prison design matter.
First, definitions — If different housing types are being compared using similar
terminology, then the risk is that incorrect inferences will be drawn about the
effects of a given housing type. Similarly, if two types of housing share important
similarities, then differing terminology may create the impression that the two
types are fundamentally different from one another.
Second, goals — Goals matter greatly for evaluating the need, theory, use,
impact, and efficiency of restrictive housing. If the goal is to punish inmates, then
an assessment of the extent to which punishable behavior warrants restrictive
housing should be undertaken. An assessment of the extent to which the
punishment in fact constitutes punishment, on theoretical grounds, also should
be undertaken. In addition, an assessment of impact is needed. (Some inmates
may seek to be placed in restrictive housing, which highlights the fact that such
housing is not intrinsically punishment (Lovell et al., 2000; Mears & Watson,
2006; Mears, 2013). Not least, the extent to which restrictive housing results in
greater benefits than costs — ideally more so than some alternative strategy —
should be assessed. Similar assessments are needed for each goal of restrictive
housing.
Third, design — The conditions of restrictive housing also matter for evaluating
its impact. Different combinations of restrictive housing features or experiences
may result in varying effects on inmates or prisons. For example, a prison system
in which restrictive housing involves stays of four to six months with as much
programming, services, and treatment as can be provided may well result in
different effects on inmate mental health, misconduct, or recidivism compared to

24 0 • National Institute of Justice | NIJ.gov

a system where restrictive housing involves much longer stays and limited or no
programming, services, or treatment.

Restrictive Housing Defined
For this white paper, restrictive housing is defined as single-cell confinement for
relatively extended periods with limited or no access to programming, services,
treatment, visitation, and the like. Goals for its use are protective custody,
punishment, or any of a range of specific management objectives. This definition
recognizes that lengthy stays in restrictive housing may occur regardless of
whether inmates are placed there for protective custody, punishment, or a
managerial reason. It also recognizes that some restrictive housing may allow for
two inmates per cell. The primary commonality in discussions about extended
isolation is that inmates are isolated from others, so this paper defines restrictive
housing as single-cell confinement.
An advantage of this definition is that it accords with many sorts of restrictive
housing — administrative segregation, supermax prisons, extended isolation,
and so on — as well as with the 2015 Bureau of Justice Statistics (BJS) national
study of restrictive housing (see, generally, National Institute of Corrections,
1997; Beck, 2015; Riveland, 1999; King, 1999; Rhodes, 2004; Naday et al., 2008;
Shalev, 2009; Association of State Correctional Administrators, 2013; Mears,
2006, 2013; U.S. Government Accountability Office, 2013; Kane et al., 2013;
Ross, 2013; Gendreau & Labrecque, 2015; Labrecque, 2015a; Morris, 2015). In
addition, it recognizes that restrictive housing may begin for one purpose (e.g.,
management) but continue to be used for another (e.g., punishment). Even so, a
disadvantage of this definition is that it obscures the fact that extended isolation
may vary greatly with respect to factors such as goals, design, and operations. The
only alternative to consistently using “restrictive housing” or some other term is
to standardize a typology of housing that all prison systems might agree to use
and follow. Such a typology does not currently exist (Beck, 2015).
In the end, the general concern appears to center around extended isolation,
regardless of the particular goals or characteristics of facilities, and the need to
ensure that it is used only when necessary.

Restrictive Housing Inmate Estimates
Prior to the 2015 BJS study, estimates of the prevalence of inmates in restrictive
housing varied considerably and did so based on less credible sources of
information than the data obtained from the large survey sample of inmates
who were interviewed for the study. Before discussing the BJS estimates, several
previous sources of information warrant discussion.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 241

In 1996, the National Institute of Corrections examined the prevalence of
supermax housing. The Institute’s definition of supermax housing was as follows:
In this survey, ‘supermax’ housing is defined as a free-standing facility, or a
distinct unit within a facility that provides for the management and secure
control of inmates who have been officially designated as exhibiting violent or
serious and disruptive behavior while incarcerated. Such inmates have been
determined to be a threat to safety and security in traditional high-security
facilities, and their behavior can be controlled only by separation, restricted
movement, and limited direct access to staff and other inmates (National
Institute of Corrections, 1997:2).
The Institute excluded from consideration protective custody and disciplinary
segregation. Even so, it found that supermax housing is sometimes used for
“routine segregation purposes (e.g., discipline, protective custody, and program
segregation)” (National Institute of Corrections, 1997:3). The Institute provided
no national estimate, though its report indicated that “there are at least 57
supermax facilities/units nationwide . . . providing a total of more than 13,500
beds” (National Institute of Corrections, 1997:3).
In 1999, Roy King used the Institute’s information to estimate that in 1998 there
were 20,000 inmates — approximately 2 percent of all state and federal inmates
— serving time in supermax housing (King, 1999). Two-thirds of states had such
housing and others had plans for building it (King, 1999; Riveland, 1999).
In 2006, drawing on a national survey of prison wardens, an Urban Institute
study undertaken by Daniel P. Mears estimated that, in 2004, 44 states had
supermax housing. The definition used was similar to that used in the National
Institute of Corrections study:
For the purposes of this survey, a supermax is defined as a stand-alone unit or
part of another facility and is designated for violent or disruptive inmates. It
typically involves up to 23-hour-per-day, single-cell confinement for an indefinite
period of time. Inmates in supermax housing have minimal contact with staff
and other inmates (Mears, 2005:49).
Using this information and that of the King estimate, the Mears study
estimated that approximately 25,000 inmates served time in such housing, and
emphasized both that the number was a rough approximation and that it likely
underestimated the true prevalence of supermax incarceration (Mears, 2005:40).
In 2008, Alexandra Naday and colleagues estimated that the number of inmates
in supermax housing nationally ranged from a low of 5,000 to a high of
100,000, with the most frequently cited estimate being around 20,000 (Naday
et al., 2008:77). In recent years, the 2004 estimate of 25,000 from the Urban

242 • National Institute of Justice | NIJ.gov

Institute study has prevailed and, for example, surfaced in the Davis v. Ayala
2015 U.S. Supreme Court decision.1 Naday and colleagues examined the basis
for the counts of supermax housing provided in these sources and argued “that
disagreements about definitions, changing policies and court decisions, reporting
and recording errors, and different counting procedures have led to a lack of
reliable and valid data on supermax” housing (Naday et al., 2008:69). They
noted, for example, that the National Institute of Corrections study expressly
excluded any segregation housing used for protective custody or disciplinary
segregation (Naday et al., 2008:72). They also observed that the count estimates
provided by the American Correctional Association and the Criminal Justice
Institute stemmed from a definition that focused on the conditions of supermax
housing and largely ignored the various uses for solitary confinement, thus
creating a discrepancy between this definition and that of the National Institute
of Corrections (Naday et al., 2008:72). The authors then identified marked
inconsistencies in the security-level status of many facilities throughout the
country, noting that various studies used operational capacity rather than actual
utilization rates, thereby creating different estimates. The authors concluded that
such variation, as well as inconsistent definitions across states and in studies,
made it difficult to arrive at a valid estimate of the number of inmates who
experienced supermax housing.
In 2014, Yale Law School, in conjunction with ASCA, conducted a survey of
state and federal corrections administrators about the use of administrative
segregation (Baumgartel et al., 2015). The results suggested that 66,000 prisoners
in the 34 jurisdictions that provided counts were “in some form of restricted
housing — whether termed ‘administrative segregation,’ ‘disciplinary segregation,’
or ‘protective custody’” (Baumgartel et al., 2015:ii). The report then noted, “If
that number is illustrative of the whole, some 80,000 to 100,000 people were,
in 2014, in segregation.” The numbers did not include individuals in restrictive
housing in jails or juvenile custodial facilities (Baumgartel et al., 2015:ii). This
estimate was based on self-reports from corrections administrators, so its
accuracy is unknown (Baumgartel et al., 2015:12).
The BJS study was published in 2015. Based on a national survey sample of
91,177 inmates in state and federal prisons and jails, the study provides, arguably,
the most accurate current estimate of restrictive housing prevalence nationally.
It bears emphasizing, however, that prior studies used varying definitions and
data sources. As a result, anyone comparing the estimates should be aware of

1

Davis v. Ayala, 135 S. Ct. 2187, decided on June 18, 2015, concerned a case involving peremptory challenges
that were alleged by Ayala to have been race-based; the Court reversed the judgement of the Court of Appeals
for the Ninth Circuit. The legal aspects of the case did not involve restrictive housing. However, Justice Kennedy,
in a concurring opinion, took note of the fact that the respondent, Ayala, had spent the bulk of his term of
incarceration, since receiving a death penalty sentence in 1989, in solitary confinement. Justice Kennedy then
noted that 25,000 inmates were estimated to be serving their sentence in solitary confinement and criticized the
use of such confinement, arguing that it was overused and caused harms to inmates.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 24 3

those differences. The National Institute of Corrections study and the Mears/
Urban Institute study, for example, used supermax terminology and thus focused
primarily on the subset of inmates placed in restrictive housing for managerial
purposes. Accordingly, the studies underestimated prevalence for all forms of
restrictive housing. Conversely, the Yale/ASCA study focused on restrictive
housing, defined as segregation that might include managerial purposes
but also protective custody and punishment. The BJS survey asked inmates
directly whether they were, or had spent time, in disciplinary or administrative
segregation or solitary confinement (Beck, 2015:16).
According to the BJS study, the following prevalence estimates can be generated
from its National Inmate Survey:
• An estimated 4.4 percent of state and federal inmates and 2.7 percent of jail
inmates were in restrictive housing on an average day in 2011-2012.
• An estimated 20 percent of prisoners and 18 percent of jail inmates spent
time in such housing in the prior year or since arriving at the facility where
they were surveyed.
• An estimated 10 percent of prisoners and 5 percent of jail inmates spent 30
days or longer in some form of restrictive housing (Beck, 2015:1).
BJS did not include counts of inmates. If the percentages above were applied
to the year-end 2013 prison and jail population counts nationally, they would
suggest the following:2
• An estimated 90,000 inmates (70,000 prisoners and 20,000 jail inmates)
reside in restrictive housing on an average day.
• An estimated 447,000 inmates (315,000 prisoners and 132,000 jail inmates)
spent time in restrictive housing in the past year or since arriving at their
current facility.
• An estimated 195,000 inmates (158,000 prisoners and 37,000 jail inmates)
spent 30 days or longer in restrictive housing.
Several observations about these estimates warrant discussion. First, the
percentages suggest that restrictive housing is used far more frequently than past
estimates indicate. Given that prison populations have greatly increased in recent
decades, restrictive housing is used for many more inmates than in prior years,
and thereby may have greater effects on prisoners than previously thought.
Second, the fact that in one 10 inmates reports having spent at least 30 days in
restrictive housing suggests that prior studies substantially undercounted inmates
2

Glaze and Kaeble (2015:2) reported that, at year-end 2013, 1,574,700 individuals were incarcerated in state or
federal prisons and that 731,200 individuals were incarcerated in jails.

24 4 • National Institute of Justice | NIJ.gov

in supermax housing. Inmates might well serve 30 days or more for protective
custody or disciplinary confinement; however, this confinement may well be
restricted primarily to situations in which an inmate is placed in isolation for a
broader managerial purpose, which would fit more with the goal of supermax
housing. If so, the estimate is, again, substantially greater than what has been
suggested in prior studies of supermax incarceration.
Third, restrictive housing is common in jails. Accordingly, national discussions
about restrictive housing should focus on both prisons and jails.
Fourth, if the percentages identified in 2011-2012 hold for current prison and
jail populations, then prior counts of inmates in restrictive housing — including
those held for protective custody, disciplinary segregation, and managerial
segregation — are substantially underestimated. Upper-end estimates have
indicated that 100,000 inmates are in solitary confinement. The estimates based
on applying the BJS percentages to current (year-end 2013) prison and jail
populations suggest that 195,000 inmates spent 30 days or more in restrictive
housing, and that 447,000 inmates spent some time in such housing.
Fifth, restrictive housing exists in many other countries, but the United States
appears to employ it more than other countries. However, little is known
empirically about country-by-country differences in the design, uses, or effects of
such housing (Ross, 2013; Richards, 2015).
Finally, it is unlikely that agreement about whether such housing is ever
appropriate on moral grounds will occur (Mears & Castro, 2006; Mears &
Watson, 2006; Mears et al., 2013). Regardless, it may be possible to garner
agreement about its appropriateness based on whether its benefits exceed its
costs and do so to a markedly greater degree than other approaches to achieving
prison and jail goals.

Goals
Evaluation of a policy’s impacts begins with clarity about its goals. There is,
in fact, little clarity about the precise goals of restrictive housing. Instead, a
range of goals for its use can be and have been articulated, which has greatly
impeded progress in evaluating the impacts of restrictive housing. The lack of
consistent, precise goals across studies is one reason that so little is known about
its effectiveness. The different goals, and their implications for research on and
discussions about restrictive housing, are discussed below.

Diverse Goals
Accounts of restrictive housing consistently identify that it can serve three goals:
to protect some inmates, to punish others, or to achieve management goals, such

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 24 5

as incapacitating out-of-control inmates or creating more systemwide safety and
order. Although protection and punishment appear to be straightforward goals,
they are not. For example, it is not clear that isolation provides better protection
than some other strategy or that it achieves some desired level of retribution. In
addition, the management goals that might be achieved through segregation vary.
The following goals associated with restrictive housing were identified by drawing
on prior studies, particularly the Urban Institute study of supermax housing.3

Goal 1 — Increased systemwide prison safety, including fewer riots, murders,
and assaults.
This requires improved conduct among inmates living in or released from
restrictive housing and among general population inmates.

Goal 2 — Increased systemwide prison order.
Although safety and order are often viewed as a single concept, the two words
represent distinct phenomena.4 Whereas safety involves preventing violence or
injury, order centers on the extent to which inmates adhere to rules, procedures,
policies, and everyday routines. Disorder in a prison system might contribute
to violence, but it need not do so. Conversely, an orderly prison system might
experience violence. In any case, prison systems view order as a priority, one that
they hope improves safety. Greater order may stem from inmates who experience
restrictive housing, from all other inmates, or both.

Goal 3 — Increased control of prisoners.
Many accounts observe that restrictive housing serves to control certain
prisoners. The meaning of control varies. For example, control can mean
incapacitating an inmate who poses an immediate and pressing threat to safety,
reducing an inmate’s ability to instigate others, or providing an inmate with a
setting in which to cool down before potentially harming themselves or others
(Burt, 1981; DiIulio, 1987; Logan, 1993; Sparks et al., 1996).

3

Analysis indicates that restrictive housing serves many different goals in other countries as well (Ross, 2013;
Richards, 2015).

4

Full discussion of the results is provided in Mears (2005, 2006), Mears and Castro (2006), and Mears and
Watson (2006). The identified goals are not necessarily prioritized by each state, and some states may
prioritize one or more of them. They are ones that the Urban Institute study identified from interviews with state
legislators and corrections officials and staff, a survey of state prison wardens, and from articles and reports on
supermax housing (e.g., National Institute of Corrections, 1997; Stickrath & Bucholtz, 2003; Pizarro & Stenius,
2004; Shalev, 2009; Richards, 2015). Some states may prioritize one or more or all of the goals. In the Urban
Institute study, supermax housing was defined in such a way as to include protective custody, punishment, and
management goals. Analysis indicates that restrictive housing serves many different goals in other countries as
well (Ross, 2013; Richards, 2015).

24 6 • National Institute of Justice | NIJ.gov

Goal 4 — Improved behavior of violent and disruptive prisoners (Stickrath &
Bucholtz, 2003).
Improved behavior among this population should contribute to greater prison
safety. With restrictive housing, however, the emphasis is on safety that results
from rehabilitation or some other change in the inmate, as opposed to safety that
results solely from incapacitating the inmate. Corrections officials in the Urban
Institute study who emphasized this goal also emphasized the goal of successfully
reintegrating inmates from restrictive housing back into the general prison
population (Atherton, 2001). From this perspective, success includes not only
incapacitating the inmate from engaging in or inciting violence or misconduct
but also promoting successful reentry into general population facilities.

Goal 5 — Reduced gang influence.
Many accounts point to restrictive housing as a way to reduce gang influence in
prisons (Mears & Watson, 2006; see also Ward & Werlich, 2003). This outcome
can be viewed as a goal in and of itself or as a means to achieve other goals, such
as improved systemwide prison safety or order.

Goal 6 — Protection of certain inmates.
Some inmates may be at risk of victimization and may warrant special housing
to protect them. In such cases, such housing would not necessarily need to be
restrictive (e.g., 23 hours per day in solitary confinement) unless a particular
need for certain restrictions existed. Stays in such housing typically would not
be extended, though it is possible that inmates placed in restrictive housing for
protection may reside there for extended periods. Such circumstances may stem
from the need for further protection, misconduct committed while in restrictive
housing, or some other reason.

Goal 7 — Punishment.
Restrictive housing may be used to punish inmates. Punishment can be viewed
as a goal in and of itself — in such cases, it constitutes retribution. How much
retribution is appropriate within prisons for various infractions or crimes is
not well-established. Since no clear consensus exists, prison systems institute
sanctions, within legal limits, that become established practice. Punishment can
also implicate several related goals. It can be viewed as a means to other ends,
such as public safety. For example, it can serve to deter an inmate from engaging
in problem behavior again. It can also serve as a general deterrent that reduces
the likelihood that other inmates will engage in violent or disruptive behavior.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 247

Goal 8 — Increased public safety.
Restrictive housing can be viewed as a strategy for increasing public safety.
A legislator in one study commented, for example, that a low escape rate
might serve to show a supermax prison’s effectiveness (National Institute of
Corrections, 1997; Austin et al., 1998; Grann, 2004; Pizarro & Stenius, 2004;
Reiter, 2012). Recidivism is, in general, a metric that has been used to measure
the effectiveness of corrections.

Goal 9 — Improved correctional system efficiency.
There are several strategies for achieving various prison system goals. Restrictive
housing can be viewed as a cost-effective strategy for achieving these goals
(Mears & Watson, 2006). For example, instead of dispersing violent or disruptive
inmates throughout the prison system, concentrating them in one place creates
potential economies of scale, including reduced transportation and staff training
costs (Hershberger, 1998).

Evaluating Restrictive Housing — Need, Theory, Use, Impact,
and Efficiency
Identification of relevant goals is essential for evaluating the need for restrictive
housing, the theory that guides its use, whether its use is appropriate, its impacts,
and its efficiency. These dimensions — need, theory, use, impact, and efficiency
— collectively constitute what Peter Rossi and colleagues call “evaluation
hierarchy,” which provides a foundation on which to systematically assess a
policy (Mears & Reisig, 2006; Mears & Watson, 2006).
Some examples might illustrate how evaluation hierarchy might work. If the
goal is to punish certain inmates using restrictive housing, then information is
needed about how many inmates engage in punishable behavior that warrants
placement in such housing. Research is also needed on the theory underlying
restrictive housing as punishment, the extent to which restrictive housing is used
appropriately, and whether it in fact achieves a desired level of retribution.
By contrast, if the goal is to reduce prison riots, a parallel set of questions arises
that center not on punishment but on riots. For example, to what extent do riots
occur and to what extent are they caused by individual inmates? How many such
inmates exist? By what logic will restrictive housing reduce their ability to cause
riots? Are the inmates most at risk of inducing riots placed in restrictive housing,
or are nuisance inmates placed in such housing instead? Does restrictive housing
reduce riots? If so, does it do so more cost-efficiently than some other alternative?
These questions — those focused on need, theory, use, impact, and efficiency —
are all specific to the particular goals associated with restrictive housing (Rossi

24 8 • National Institute of Justice | NIJ.gov

et al., 2004; Mears 2010). For example, if restrictive housing is used to improve
systemwide prison safety and order, then evaluations that address each of these
dimensions must be undertaken (Mears & Reisig, 2006). How much violence
and disorder exists? What causes violence and disorder? Are a select few inmates
the primary cause? How exactly will brief, intermediate, or long-term stays in
restrictive housing improve safety and order? Are the inmates who contribute
most to violence and disorder in fact in restrictive housing? How much
systemwide safety and order result from the use of restrictive housing? What is
the cost-benefit ratio? Is it greater than what might arise from other strategies for
improving systemwide safety and order?

A Balanced Assessment of the Impact of Restrictive Housing
A balanced assessment of the impact of restrictive housing requires evaluation of
the extent to which it contributes to its different goals. It also requires subjective
assessments about which particular goals are most important. States may vary,
for example, in the extent to which they use restrictive housing for one goal
rather than another (Mears & Castro, 2006; Mears & Watson 2006). Accordingly,
impact evaluations should assess how well restrictive housing achieves various
goals and weight these goals against a state’s stated goals for the housing.
In addition, even if restricted housing seeks to achieve only one goal, unintended
impacts, whether positive or negative, should be considered in arriving at a
balanced assessment of impact (Mears & Watson, 2006). For example, no state
uses restrictive housing with the intended goal of worsening inmate mental
health or recidivism outcomes, yet such housing may do so.

Need
Restrictive housing may be needed and may serve a critical role in achieving
prison system goals. Alternatively, it may not be needed, and some other strategy
may be more appropriate for achieving these goals or addressing specific
problems. Accordingly, clarity about the need for restrictive housing is critical for
understanding the extent to which it achieves various goals. The state of evidence
to date provides little credible empirical basis for knowing whether or how much
restrictive housing has been or is needed.

The Need for Restrictive Housing Depends on the Goal
Is restrictive housing needed? The answer depends on the goal. If restrictive
housing serves to control especially dangerous, violent, or disruptive inmates,
then there may be a need if such inmates exist. They assuredly do, and so need
would be based on assessing how many such inmates warrant restrictive housing.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 249

However, the situation is complicated by other goals that exist for restrictive
housing. For each goal, a prison system will want to understand the prevalence
of a particular problem that restrictive housing is intended to address. For
example, if restrictive housing serves to punish inmates, how many individuals
commit the kinds of acts that warrant placement in restrictive housing? How
long can inmates be expected to reside in restrictive housing? By answering such
questions, prison system officials might begin to determine how many restrictive
housing beds it needs. Whatever the final answer may be, it will differ from the
number arrived at if the focus is on a different goal, such as protecting inmates or
reducing gang influence in prisons.

The Need for Restrictive Housing Depends on the Causes of
Particular Problems
Whether restrictive housing is needed depends on the causes of the problems
it is intended to address — perhaps an inmate cannot be protected through any
other approach, perhaps an inmate committed an act that warrants punishment,
or perhaps an inmate engaged in activities that somehow created problems
that posed a risk to the prison system. In each instance, many facilities might
determine that restrictive housing is needed. Typically, their focus would then
shift to how many inmates fit these different profiles to determine how much
restrictive housing should be built. However, this straightforward approach
would be inappropriate for several reasons. Two examples illustrate why.
First, the need may not be for restrictive housing; rather, it may be for efforts that
target the causes of the inmate behaviors. Poor staff professionalism or limited
rehabilitative programming at a given facility, for example, might result in some
inmates not receiving the protection they may need. These conditions also might
contribute to other problems such as violence and infractions. Accordingly,
in such situations, the need is not for restrictive housing. Instead, the facility
may need to institute efforts that improve staff professionalism and increase
rehabilitative programming opportunities (Sykes, 1958; DiIulio, 1987; Cooke,
1989; Sparks et al., 1996; Kassel, 1998; Reisig, 1998; Bottoms, 1999; Toch, 2003;
Rhodes, 2004; French & Gendreau, 2006; Mears & Reisig, 2006; Mears, 2008a,
2013; Useem & Piehl, 2008; Browne et al., 2011; Cullen et al., 2014).
Second, the need may not be for restrictive housing; rather, it may be for first
employing a continuum of alternative and less costly strategies. Restrictive housing
has been described by corrections officials as a “last resort” option (Mears,
2013) — the equivalent of an emergency room in a hospital. Long before
using this option, other, less costly options — such as promoting greater staff
professionalism, ensuring fair and consistent enforcement of rules, improving
officer and inmate culture, providing targeted programming that addresses
each inmate’s individual risks and needs, relying on evidence-based behavioral

250 • National Institute of Justice | NIJ.gov

interventions, and so on — should be considered (Sykes, 1958; DiIulio, 1987;
French & Gendreau, 1996; Sparks et al., 1996; Reisig, 1998; Bottoms, 1999;
Riveland, 1999; Gendreau & Keyes, 2001; Briggs et al., 2003; Mears & Watson,
2006; Shalev, 2009). In many instances, there may not be a need for restrictive
housing but instead for more and better implementation of other options. If such
options have been exhausted and behavior still falls short, then there may be a
need for restrictive housing as another tool for addressing a particular problem.
In short, empirical assessment of the need for restrictive housing requires going
well beyond identifying a certain number of inmates for protection, punishment,
or some management goal. It requires explicit description of the precise problems
that restrictive housing is intended to address and the goals that it is intended to
achieve. It also requires empirical evaluation of the causes of these problems and
the extent to which those causes can be addressed through less costly options. Not
least, it requires evaluating the extent to which such options have been pursued in
sufficient doses, or with sufficient fidelity, to reduce a particular problem. If they
have not, then attention presumably should focus on improving the options, not
in taking recourse in a more expensive one. With such analyses in place, the focus
logically would turn to examining the need for restrictive housing.
Although there is little empirical evidence in the published literature that
restrictive housing is needed, it appears likely that prisons do need some such
housing. What is lacking, however, are empirically based evaluations that
document the need for restrictive housing relative to the need for a range of
other approaches that address the problems that restrictive housing is supposed
to address. Instead, states by and large appear to have focused on forecasting
based on current use. For example, if 500 inmates were in restrictive housing in
a given year, a state might assume that the use was needed and would continue
to incarcerate these inmates in restrictive housing. The state then might forecast
a need to increase the number of restrictive housing beds based on an expected
increase in the general inmate population. This approach largely sidesteps the
issues of whether the actual use was needed in the first place or what the current
need may be.5

The Need for and Impact of Restrictive Housing Depends on the
Point of Comparison
Restrictive housing, especially when used for management purposes, may be
viewed from at least two vantage points. The first is that such housing is rarely
5

An illustration of the potential problems with this approach can be found in a National Public Radio (2012)
interview with Walter Dickey, who was the director of the Wisconsin Division of Corrections from 1983-1987 and
subsequently the Federal Monitor for the supermax prison at Boscobel, Wisconsin: “I think one of the things that’s
happened, at least in a lot of states, Wisconsin’s one of them, is I think we grossly exaggerated the need for the
supermax prison and overbuilt it, and I think, not surprisingly, when you’ve got empty cells in a crowded prison
system, you tend to fill them up.”

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 251

used. This characterized most state prison systems in the United States in the
1970s and early 1980s. When restrictive housing was rarely used, the question
for many states would have been, “To what extent is restrictive housing needed
in a context where it rarely is used?” Given the marked growth in restrictive
housing in subsequent decades, states clearly viewed restrictive housing as greatly
needed. However, there is little evidence that states systematically and empirically
assessed the need for this growth.
The second vantage point is that such housing is used relatively frequently.
This characterizes most states in contemporary America. The question for
many such states is, “To what extent is the current level of restrictive housing
needed?” Without a needs evaluation of the various problems that such housing
is intended to address, as well as the options available to address them, it will
remain difficult to answer that question.
Distinguishing between these scenarios is important, primarily to highlight
that current use should not necessarily be viewed as appropriate or needed.
Indeed, it may well be that past use consistently well exceeded the need or,
in some states, fallen short of the need. In both scenarios, research would
document the need empirically.

Theory
Theory, whether explicitly articulated or not, guides any policy or program.6
Restrictive housing is no different (Mears & Watson, 2006; Mears, 2013.) It
involves a causal logic: By changing some set of conditions or factors, restrictive
housing creates improvement in outcome. Any policy or program that lacks
a clear causal logic built on credible theory and research risks failure. A welldesigned policy or program that builds on credible theory is more likely to
provide clear guidance about implementation and to be effective. This axiom
applies to all social policy arenas, including criminal justice and corrections
(Rossi et al., 2004; Mears, 2010; Latessa et al., 2014). As discussed below, the
theory guiding restrictive housing is unclear; ambiguity in the theory complicates
evaluation of implementation and impact. It is a central reason why so little is
known about the impacts of restrictive housing or what contributes, in extant
studies, to any identified impact.

Ambiguity in the Theoretical Foundations of Restrictive Housing
Several ambiguities lie in the theoretical foundations of restrictive housing. The
diverse goals associated with such housing create ambiguity because a single
6

Theory guides research as well. Ideally, researchers and policymakers explicitly articulate the theory that guides
their efforts. Doing so identifies potentially problematic assumptions as well as important information about how
exactly a given phenomenon, or policy, contributes to a given outcome (Rossi et al., 2004; Mears, 2010).

252 • National Institute of Justice | NIJ.gov

treatment — restrictive housing — is unlikely to achieve each goal through the
same causal pathways. Another source of ambiguity is the lack of clarity about
how restrictive housing might achieve any particular goal. This paper describes
prominent ambiguities in the theoretical or causal logic of restrictive housing.
First, with rare exception, the causal logic for restrictive housing has not been
formally articulated. Exceptions exist, such as a theoretical analysis of supermax
housing and how it might affect prison system safety and order (Mears & Reisig,
2006). However, states have not undertaken such analyses for particular housing
goals that they seek to achieve through restrictive housing (e.g., increasing prison
system safety or order, increasing control of prisoners, improving behavior
of violent or disruptive inmates, reducing gang influence, protecting certain
inmates, punishing certain inmates, increasing public safety, and improving
correctional system efficiency) (Mears & Reisig, 2006).
Second, the theory by which restrictive housing might achieve certain goals rests on
questionable assumptions. Several examples illustrate this issue.
Increasing systemwide safety and order. Restrictive housing has been viewed
as a primary strategy for increasing systemwide safety and order. One theory is
that inmates in restrictive housing, as well as those in the general population,
will be deterred from misconduct. However, any such effect depends on
restrictive housing meeting certain conditions required for deterrence to occur,
including objective and perceived certainty, celerity, and severity of punishment
(Paternoster, 2010). For example, inmates in restrictive housing must view it
as punishment; the fact that inmates sometimes seek to be placed in restrictive
housing suggests that it is not always viewed as punishment (Suedfeld, 1974;
Suedfeld et al., 1982; Lovell et al., 2000; Singer, 2003; Mears & Watson, 2006;
Richards, 2015). At the same time, inmates in the general prison population
may take stock of precisely what behaviors must occur to be placed in restrictive
housing and engage in only those that stop short of what is likely to result in such
placement (Pizarro & Stenius, 2004). In addition, perceived prison system misuse
or abuse of restrictive housing may engender ill will among inmates directly
or indirectly exposed to it. From the perspective of defiance theory, restrictive
housing may contribute to more rather than less violence and disorder (Sherman,
1993; see also Ward & Werlich, 2003, regarding the “rage hypothesis”).
Decreasing prison riots. In a number of states, restrictive housing increased
in response to prison riots; in turn, its use was justified on the grounds that
it could prevent riots from occurring. Riots in fact may stem from many
causes, only one of which is a set of violent or disruptive inmates (Martin &
Zimmerman, 1990; Boin & Rattray, 2004; Mears, 2008a). Officer professionalism
and interactions with inmates, consistent rule enforcement, provision of various
types of programming, and so on, all may contribute to reducing riots, while
their absence may contribute to increasing riots (Useem & Kimball 1991; Sparks
et al., 1996; Reisig, 1998; Bottoms, 1999; Gendreau & Keyes, 2001; Mears, 2008a;

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 253

Useem & Piehl, 2008; Morris et al., 2012). Restrictive housing does little to
address these root causes of riots. Instead, its use proceeds from the assumption
— one not supported by research — that a certain group of inmates constitutes
the primary if not exclusive cause of riots.
Punishment. When the goal of restrictive housing is retribution, the assumption
is often that inmates experience placement in such housing as punishment.
However, it is unclear precisely which theory of punishment suggests that
isolation constitutes a sanction that individuals experience as punishment or
that agrees with societal views about appropriate levels of punishment. This
observation underscores a more general issue: Punishment can consist of any
of a wide range of possibilities in prisons (e.g., restricted privileges, transfer to a
less desirable work assignment, fines, transfer to another facility). The manner in
which restrictive housing fits within this range of possibilities, or which theory
guides these possibilities, typically has not been articulated by prison systems.
More generally, systems tend not to address the following questions: How
does restrictive housing result in a desired level of retribution? What level of
retribution is desired? How much time in restrictive housing achieves that level
of retribution? A theory might shed light on such questions. The absence of such
a theory implies that restrictive housing does not achieve in any obvious way
intended levels of retribution for society or correctional systems.
The theoretical logic for how restrictive housing might achieve these and any of
its other intended goals remains uncertain. Also uncertain is the extent to which
the different assumptions underlying the logic are correct. For example, to what
extent does incarcerating security threat group (gang) leaders prevent them
from communicating with other gang members? If they are able to communicate
despite segregation, then a central theoretical argument for restrictive housing
would be undermined (Grann, 2004; Preer, 2004; Mears & Watson, 2006).
Conversely, if restrictive housing does inhibit communication, and if no
members “step up” to replace a gang leader placed in such housing, then it may
be reasonable to expect gang violence to decline. In a related vein, placing an
inmate in protective custody might reduce opportunities for victimization, but
doing so depends on at least two factors: credible information about the risk to
inmates, and other inmates not viewing protective custody as a negative signal
about that inmate’s character. In addition, protective custody itself does not target the
prison conditions that might contribute to victimization of certain inmates.
Third, credible theoretical arguments that restrictive housing worsens a range
of outcomes can be and have been articulated. A theoretical argument that
restrictive housing worsens outcomes is not necessarily accurate, any more than
the argument that restrictive housing improves outcomes. An ideal scenario for
any policy, however, is when the theoretical arguments for it rest on credible
theory and research, while the arguments against it do not. That scenario does
not characterize restrictive housing. For example, restrictive housing does little
to address known causes of inmate misconduct, recidivism, riots, or the like.

25 4 • National Institute of Justice | NIJ.gov

It potentially diverts resources and attention away from approaches that may
be more effective and may create conditions that worsen inmate behavior. It
reduces opportunities for inmates in restrictive housing to receive effective drug
or mental health treatment and to maintain ties with family or friends. Use
of restrictive housing also largely precludes provision of cognitive-behavioral
counseling and other interventions that can improve inmate behavior and
reduce recidivism. In addition, the sensory deprivation and isolation from others
that occurs with restrictive housing may contribute to mental illness. In these
instances and in others, scholars have traced theoretical mechanisms through
which restrictive housing may create adverse outcomes (Sykes, 1958; McCorkle
et al., 1995; Bottoms, 1999; Kurki & Morris, 2001; Haney, 2003; Rhodes, 2004;
Irwin, 2005; King, 2005; Cloyes et al., 2006; Mears & Watson, 2006; Smith, 2006,
2008; Kupers, 2008; Kupers et al., 2009; Gawande, 2009; Shalev, 2009; Haney et
al., 2015).
Fourth, restrictive housing does not address a variety of potential causes of inmate
violence, prison disorder, gang influence, victimization of certain inmates, and
so on. Focusing in particular on systemwide safety and order, Richard Sparks
and colleagues noted that “what [special prison units] cannot do is magically
to unlock the problem of order for a prison system as a whole” (Sparks et
al., 1996:313). Some accounts depict state prison wardens as enthusiastically
embracing restrictive housing because of the control that it affords over some
inmates.7 Others, however, suggest that wardens’ views about the appropriateness
and effectiveness of restrictive housing vary (Mears, 2006; Mears & Castro, 2006;
Mears & Watson, 2006). Support for and perceptions of restrictive housing as a
solution to prison problems8 may stem from the fact that wardens often confront
and must manage extremely dangerous individuals. Yet, restrictive housing
in and of itself does not directly address the causes of a particular individual’s
behavior or, more generally, the causes of systemwide operations, safety, and
order. Accordingly, on theoretical grounds, it is unclear whether restrictive
housing can achieve many of the goals associated with it (King, 1999, 2007;
Mears & Reisig, 2006).

Theory and Its Relevance for Evaluating Implementation and Impact
Advances in the theory underlying restrictive housing are directly relevant to
evaluating its implementation and impact. For example, if restrictive housing
is expected to achieve systemwide safety and order through specific deterrence, it
should be used in such a way that this goal is achieved. That might include using
restrictive housing only for those inmates who in fact may be deterred from bad
behavior by placement in it. Similarly, if restrictive housing is expected to reduce
7

Ward and Werlich (2003:59) reported that state prison wardens who visited the federal supermax facility at
Marion, Illinois, indicated that they “had died and gone to heaven” because of the extensive control that the staff
had over inmates.

8

Lorna Rhodes (2004:36) has noted, for example, that “the dream of the perfect prison has deep historical roots.”

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 255

gang influence, then it is important that only the most influential gang members are
placed in it. Theory identifies, too, the intermediate outcomes that may be relevant to
evaluating impact. In these and other instances, appropriate implementation requires
clarity about the goals of restrictive housing, the types of inmates for whom it is
appropriate, the “dose” of restrictive housing that may be most efficacious, and the
treatment and services that may achieve the intended goals.
Credible theory is also relevant for increasing intended benefits and reducing
unintended harms. For example, general deterrence does not require specific
lengths of stay. Accordingly, it may be that placing more inmates in restrictive
housing for short durations may be more effective than segregating fewer
inmates for longer durations. Crime theories suggest that severing inmates’ social
ties and failing to address their criminogenic needs may contribute to more
offending rather than less (Latessa et al., 2014; Mears & Cochran, 2015). From
this perspective, greater gains in reducing misconduct or recidivism may arise
either from not using restrictive housing or from combining its use with a variety
of rehabilitative interventions.
Similar illustrations exist for the other goals associated with restrictive housing.
In each instance, theory is needed that can guide evaluation of implementation
and impact. Such theory can be used to identify aspects of restrictive housing that
might be modified to maximize its beneficial impacts and minimize its harms.

Implementation
A necessary condition for restrictive housing to achieve its intended goals is that
it house appropriate inmates and be operated according to design. Typically,
full and quality implementation should be established prior to evaluating
policies and programs and estimating their impact. Otherwise, failure to achieve
intended goals may be due to poor implementation rather than poor program
design (Rossi et al., 2004; Mears, 2010). To date, however, there are no systematic
empirical accounts, by year, that document the extent to which the housing
is used according to design (King, 1999; Kurki & Morris, 2001; Shalev, 2009;
Mears & Bales, 2010; Browne et al., 2011; Katel, 2012; Mushlin, 2012; Reiter,
2012; Mears, 2013; U.S. Government Accountability Office, 2013; McGinnis et
al., 2014). It is therefore not possible at present to know how (if at all) restrictive
housing has contributed to the benefits or harms that have been ascribed to it.

Appropriate Implementation Depends on Goals
Determining the appropriate inmate population and operations of restrictive
housing depends entirely on the goals associated with it. The ambiguity about the
precise goals of restrictive housing — or the weighting of multiple goals — makes
such determinations difficult. In addition, many states lack clear restrictive

256 • National Institute of Justice | NIJ.gov

housing designs or blueprints that detail the theory or causal logic that guides
its use. Such blueprints would include information about the following: specific
goals, types of inmates who warrant placement in restrictive housing, the length
of time in restrictive housing necessary to achieve the goals, and the precise
operations, services, and treatment that are collectively expected to achieve these
same goals. Ambiguity about the goals, theory, or design of restrictive housing
creates ambiguity about assessing implementation. By extension, ambiguity in
assessing the extent to which restrictive housing is used appropriately — that is,
as intended — creates ambiguity in determining how to interpret its estimated or
assumed impacts (Mears, 2006, 2008b, 2013).

Appropriate Implementation of Protocols, Rules, and Procedures
The federal government and states have protocols, rules, and procedures for
restrictive housing (National Institute of Corrections, 1997; Austin et al., 1998;
Riveland, 1999; Neal, 2003; Bruton, 2004; Collins, 2004; Mears, 2006; Shalev,
2009; Butler et al., 2012; Association of State Correctional Administrators, 2013;
U.S. Government Accountability Office, 2013; McGinnis et al., 2014; Richards,
2015). A review by Kenneth McGinnis and colleagues, for example, provided
an extensive description of the Federal Bureau of Prisons’ special housing units
(McGinnis et al., 2014). They found substantial gaps in documentation assessing
whether the Bureau’s housing is used appropriately and with fidelity to its design.
The review also found that although the “general conditions of confinement in
restricted housing units are consistent with national regulations and standards,
the Bureau does not have adequate non-punitive protective custody housing
units that have equivalent levels of programs and privileges as general population
inmates.” They also found that “backlogs in inmates awaiting transfer to the
next program level negate the intent of the program design and decrease the
motivation to change behavior.” In addition, the team found that “mental health
services in restrictive housing require improvement in three specific areas:
(1) proper mental health diagnoses, (2) more effective treatment, and (3)
providing sufficient psychiatric staffing.” The team concluded that “the lack of
time parameters for completion of disciplinary hearings results in substantial
variation among facilities in the amount of time served in segregation for similar
offenses, and can result in disproportionately long sanctions” (McGinnis et al.,
2014). It is likely that comparable reviews of state prison systems would identify
similar gaps between intended and actual operations of restrictive housing
(Riveland, 1999; Briggs et al., 2003; Pizarro & Stenius, 2004; Mears, 2006, 2013;
Mears & Watson, 2006; Reiter, 2012; Labrecque, 2015b; Richards, 2015).

Appropriate Dose of Restrictive Housing
The proper amount, or dose, of any intervention or policy requires an explicit
articulation of the theory and research that justify a given level of treatment.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 257

Too little intervention may result in no improvement, and too much may be
unnecessary and harmful. What dose of restrictive housing is required to
protect or punish an inmate? What dose is required to deter both inmates who
experience restrictive housing (specific deterrence) and general population
inmates (general deterrence)? Finally, what dose is required to improve
systemwide safety and order, reduce gang influence, or improve correctional
system efficiency or public safety?
To date, few studies have documented the precise lengths of stay that inmates in
restrictive housing experience, the frequency of their placement in such housing,
or the percentage of all prison system inmates placed in restrictive housing at some
point or another. Those few studies suggest that terms of confinement in restrictive
housing can range from days, weeks, or months to 10 or more years; that inmates
can cycle into and out of restrictive housing repeatedly; and that prison systems
devote widely varying percentages of bed space to restrictive housing (BarakGlantz, 1983; National Institute of Corrections, 1997; King, 1999, 2007; Mears &
Bales, 2010; Reiter, 2012; U.S. Government Accountability Office, 2013; Beck, 2015;
Shames et al., 2015). However, no clearly articulated theory exists that establishes
the precise durations needed in restrictive housing, or the number of inmates who
need to be placed in it, to achieve its various intended goals.

Appropriate Inmates for Restrictive Housing
Whether an inmate is an appropriate candidate for restrictive housing depends
first on the goals and design of the housing. For example, if it is used for
protective custody, then inmates should be placed in restrictive housing when
they meet a predefined set of criteria that indicate they need protection that
cannot be provided through less costly or intrusive strategies. If punishment is
the goal, then only those inmates who engage in acts that warrant punishment
should be placed in restrictive housing. Not least, if restrictive housing serves
any of a range of more general managerial goals, then these goals — and the
types of inmates whose placement in restrictive housing could achieve these
goals — must be explicitly articulated . As discussed above, the varying goals
and the ambiguity in the theory or design of restrictive housing complicate
any assessment of its implementation. Some studies have described random
(representative) samples or non-random (nonrepresentative) samples of inmates in
restrictive housing, as well as reports about the characteristics of individuals in such
housing (Suedfeld, 1974; Suedfeld et al., 1982; Haney, 2003; Mears & Castro, 2006;
Mears & Watson, 2006; Shalev, 2009; Mears & Bales, 2010; Reiter, 2012; O’Keefe et
al., 2013; McGinnis et al., 2014; Baumgartel et al., 2014; Beck, 2015; Helmus, 2015;
Labrecque, 2015; Richards, 2015). These studies do not, however, establish clearly
whether appropriate inmates have been placed into restrictive housing.

258 • National Institute of Justice | NIJ.gov

Perhaps the only clear area of agreement in the literature is that nuisance inmates
and the seriously mentally ill should not reside in restrictive housing. Yet,
nuisance inmates clearly can and do get placed in restrictive housing (Riveland,
1999; Lovell et al., 2000; Kurki & Morris, 2001; Pizarro & Stenius, 2004; Mears,
2006, 2013; O’Keefe, 2008; Shalev, 2009; Browne et al., 2011; Richards, 2015), as
do seemingly large numbers of inmates with serious mental disorders (Kurki &
Morris, 2001; Haney, 2003; Cloyes et al., 2006; Mears, 2006; Smith, 2006, 2008;
Kupers, 2008; Lovell, 2008; O’Keefe, 2008; Kupers et al., 2009; Shalev, 2009;
O’Keefe et al., 2011; Beck, 2015; Richards, 2015). The precise prevalence of
nuisance inmates or the seriously mentally ill in restrictive housing — by state
and over time — is unknown. However, a large number of studies and reviews
have found that many inmates who are not appropriate for restrictive housing
reside in it and may do so for extended periods of time, suggesting that facilities
may not be using restrictive housing as an option of last resort (Riveland, 1999;
Lovell, et al., 2000; DeMaio, 2001; Kurki & Morris, 2001; Haney, 2003; Pizarro &
Stenius, 2004; Cloyes et al., 2006; Mears, 2006, 2013; Smith, 2006; Kupers, 2008;
Lovell, 2008; Kupers et al., 2009; Shalev. 2009; Mears & Bales, 2010; Browne et al.,
2011; O’Keefe et al., 2011; Baumgartel et al., 2015; Beck, 2015; Labrecque, 2015a;
Richards, 2015; Shames et al., 2015).

Implementation and Impact
Appropriate implementation can be viewed as an important goal in its own
right. From this perspective, state and federal prison systems ideally can provide
empirical research that documents fidelity to the protocols, rules, and procedures
for restrictive housing. Full and appropriate implementation is important, too,
for determining whether restrictive housing can or is likely to produce intended
outcomes. Accordingly, the absence of systematic empirical research on the
implementation of restrictive housing means, by extension, that it will remain
difficult to determine whether restrictive housing can be credited with achieving
the goals set forth for it.

Impacts
Research on restrictive housing impacts can be summarized briefly. First,
extant research does not provide a credible foundation on which to assert
with confidence the impacts of restrictive housing in general or in specific
prison systems. Second, few empirical studies of restrictive housing impacts
have been undertaken. Third, extant studies examine only a small number of
relevant outcomes. Fourth, these studies typically have not relied on strong
methodological designs — including the use of relevant comparison groups or
conditions — which makes it difficult to know whether restrictive housing causes
observed outcomes. Fifth, even if many credible evaluations of impacts across a

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 259

range of outcomes existed, questions would remain about how to interpret them
because the need for and the theory and implementation of restrictive housing
have not been well established empirically. For example, if an evaluation found
that restrictive housing failed to improve systemwide safety and order, that failure
might be due to faulty design (e.g., restrictive housing simply does not work) or
poor implementation (e.g., the inmates most appropriate for such housing were
not placed in it).
Collectively, these limitations underscore the considerable need for caution in
drawing conclusions about the impacts of restrictive housing. The discussion
below identifies potential impacts, intended or not, of restrictive housing (Mears
& Watson, 2006). Because impact assessments hinge entirely on identifying the
appropriate counterfactual — that is, what would have happened had restrictive
housing not been used — this issue is discussed first. The discussion then
turns to the potential impact of restrictive housing on prison safety and order,
inmates during and after release (including re-entry to society), retribution, gang
influence, society, and prison system operations. The discussion also describes
potential impacts of not using restrictive housing when it may be needed and,
conversely, of using it when it may not be needed.

Impacts and Counterfactuals
A valid evaluation of the impact of restrictive housing requires information
about what a prison system would have done instead of using restrictive housing.
Unfortunately, it frequently is not clear what the counterfactual condition entails.
Would a prison system have continued to employ its typical array of strategies for
addressing a particular problem? Invested more in one particular strategy, such
as strategically dispersing inmates throughout the prison system? Built a new
maximum security prison? Increased treatment and rehabilitative programming?
Improved officer training? Promoted cultural change among officers and inmates?
In each instance, what the prison system would have done likely depends on
which problem it sought to solve. There might have been a greater need for
punishing inmates, protecting them, or reducing gang violence or prison riots.
Each need entails different potential sets of responses.
To further complicate the situation, researchers typically do not know what
would have happened had restrictive housing not been an option. Identifying
the relevant counterfactual requires guidance from corrections officials about
what they would have done. Consider a situation in which inmates released from
restrictive housing are compared with inmates from a prior time period who
seem to be similar in all respects but could not be placed in such housing. The
assumption here is that the prison system would have continued in a business­
as-usual mode had it not built restrictive housing. That assumption might be
correct. However, officials might have invested in other strategies to address a

260 • National Institute of Justice | NIJ.gov

particular perceived problem if they could not rely on restrictive housing. Here,
then, using “matched” inmates from a prior time period would be inappropriate.
The lack of clarity about the most appropriate or relevant counterfactual to use
when estimating restrictive housing impacts undermines almost all extant empirical
studies of the effects of restrictive housing. These studies still can and do offer
insight into the impacts of restrictive housing, but the insight ultimately is about
potential impacts as compared to a largely unknown counterfactual. This issue
also confronts any attempt to estimate the impact of reducing restrictive housing.
What is the impact of reducing the use of restrictive housing? The answer
depends greatly on what the prison system otherwise would do.

Impacts on Prison System Safety and Order
Few studies have empirically examined the effects of restrictive housing on
systemwide safety and order. A study by Howard Bidna in the 1970s found little
evidence that a lockdown of many of California’s maximum-security prisons
affected stabbings in high-security units but some evidence that it reduced
stabbings in other units (Bidna, 1975). Research by Ben Crouch and James
Marquart suggested that Texas’ use of supermax housing may have reduced
prison system homicides in the 1980s (Crouch & Marquart, 1989). Neither study
included comparison sites or examined a diverse range of measures of violence
or disorder. In what remains the strongest evaluation of supermax prison effects
to date, Chad Briggs and colleagues compared three experimental sites (Arizona,
Illinois, and Minnesota) and one comparison site (Utah) and then sought to
determine whether systemwide levels of inmate-on-inmate and inmate-on-staff
assaults declined more in the experimental sites (Briggs et al., 2003; see also
Sundt et al., 2008). Little evidence of a consistent or appreciable beneficial impact
of supermaxes surfaced; most of the analyses identified null effects, some pointed
to modest reductions in violence, and one pointed to an increase. Beyond such
studies, there exist primarily anecdotal accounts (including interviews with non­
random samples of staff or inmates) that suggest a range of possible benefits and
possible harms (Kurki & Watson, 2001; Briggs et al., 2003; Haney, 2003; Bruton,
2004; Rhodes, 2004; Mears, 2006; Mears & Watson, 2006; Shalev, 2009; Mears,
2013; Richards, 2015; Valera & Kates-Benman, 2015).
In short, there remain almost no strong evaluations of the systemwide impacts
of restrictive housing on inmate or staff safety or on the orderly operations of
prison systems. Indeed, there are virtually no methodologically rigorous studies
that examine the effects of restrictive housing on order alone; that is, the extent
to which restrictive housing improves the amount and quality of inmate rule
compliance and the day-to-day operations of prison systems.
Some qualitative research accounts suggest that restrictive housing may be
helpful in managing a crisis, such as a prison riot (Crouch & Marquart, 1989;

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 261

Ward & Werlich, 2003; Mears, 2008a; Austin & Irwin, 2012). However, there
remains little methodologically rigorous empirical research to support that claim
or to suggest that restrictive housing prevents riots.
Restrictive housing may seem to incapacitate inmates from contributing to
violence or disorder. However, inmates can and do affect both while in such
housing (Austin et al., 1998; Preer, 2004; Mears & Watson, 2006; Mears,
2013). On theoretical grounds, restrictive housing might contribute to
improved systemwide safety and order through incapacitation, deterrence,
and normalization of prison environments, but it might as easily worsen these
outcomes. For example, use of restrictive housing might divert resources from
more effective strategies and better prison management, and it may antagonize
inmates in restrictive housing and in other prison facilities (Briggs et al., 2003;
Mears & Reisig, 2006). Such possibilities remain largely unexamined empirically.

Impacts on Inmates While in Restrictive Housing
Considerably more research has focused on the impact of restrictive housing on
individuals while they reside in restrictive housing rather than after they leave it.
This research has tended to focus almost exclusively on mental health. A range of
outcomes is relevant to assessing the impact of restrictive housing on inmates.

Protection — Less Victimization
A goal of restrictive housing is to protect inmates from victimization. Any
such benefit depends on placing individuals most at risk of victimization in the
housing. It also depends on the duration of the placement. A temporary stay in
restrictive housing might prevent victimization during that stay, but it would do
nothing to protect the inmate upon release. Little systematic empirical research
has been undertaken to document the prevalence of restrictive housing for
the purpose of protecting an inmate, the extent to which only highly at-risk
individuals experience protective custody housing, the extent to which they are
victimized upon release, and whether the rate of victimization declined after such
custody. It may be safe to assume that certain inmates are victimized less often
while in restrictive housing. However, it is possible, if not likely, that protective
custody placement may increase victimization once the inmate is released back
into general population facilities. It is also possible that some inmates are placed
in restrictive housing under the auspices of seeking to protect them when in
reality the primary goal is to remove nuisance inmates from these facilities
(Lovell et al., 2000; Richards, 2015).

Misconduct
Inmates can and do engage in misconduct while in restrictive housing; indeed,
some accounts suggest that housing design may directly induce violent behavior

262 • National Institute of Justice | NIJ.gov

(Austin et al., 1998; Bruton, 2004; Preer, 2004; Rhodes, 2004; Mears & Watson,
2006; King et al., 2008; Richards, 2015). The precise prevalence of misconduct
among inmates in restrictive housing is not known. It also is not known how
much the type or rate of misconduct decreases when an inmate is placed in
restrictive housing.

Mental Health
Since the 1970s, a large number of studies have suggested that restrictive housing
may harm inmates’ mental health. In fact, mental illness is substantially more
prevalent in the inmate population than in society at large (Mears, 2004; Mears &
Cochran, 2012, 2015; Prins, 2014). Accordingly, inmates with mental illness can
and do get placed in restrictive housing, despite legal challenges to this practice
(Haney, 2003; Smith, 2006; Kupers et al., 2009; Beck, 2015; Haney et al., 2015).
Whether restrictive housing causes mental illness or exacerbates existing mental
illness is less clear. On the one hand, numerous studies document that inmates
in restrictive housing have a mental illness and report that restrictive housing
appears to contribute to the illness. This view is supported by the argument
that humans fundamentally are social beings and that depriving them of social
contact for extended periods contributes to mental illness, self-harm, and
possibly suicide (Grassian, 1983; Haney, 2003; Smith, 2006; Dye, 2010; Kaba et
al., 2014; Lanes, 2015; Shames et al., 2015). On the other hand, few studies exist
that include appropriate, matched comparison groups to document the effect
of restrictive housing on inmates with mental illness relative to what otherwise
would happen to these inmates (Smith, 2006; Mears, 2008b, 2013; Gendreau &
Labrecque, 2015). An exception is a study by Maureen O’Keefe and colleagues
that found little evidence that restrictive housing caused mental illness (O’Keefe
et al., 2011, 2013). This finding was echoed by a meta-analysis of prior published
work (Gendreau & Labrecque, 2015).
The effect of restrictive housing on an inmate’s mental health likely varies
depending on a variety of factors. These factors include the extent to which
inmates actively seek placement in restrictive housing, the duration of exposure
to it, the ability of certain individuals to cope with isolation, and the conditions
of confinement, such as the availability of treatment and programming and
the quality of inmate and staff relationships (Lovell et al., 2000; Kurki &
Morris, 2001; Mears & Watson, 2006; O’Keefe et al., 2011, 2013; Mears, 2013;
O’Donnell, 2014; Gendreau & Labrecque, 2015; Morris, 2015; Richards, 2015;
Valera & Kates-Benman, 2015). The effect likely varies, too, depending on the
counterfactual condition, such as transfer to or placement in a maximumsecurity prison or mental health facility. Such possibilities have not been
subjected to empirical analysis.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 26 3

Physical Health
No systematic empirical research has been undertaken to document changes
in the physical health of individuals placed in restrictive housing. However,
confinement to a small cell likely contributes to minimal active movement or
exercise and, by extension, associated harms (Booth et al., 2012).

Participation in Rehabilitative Programming
Studies show that inmates receive little or no rehabilitative programming, and
services and treatment more generally, while in restrictive housing, and such
programming is or may be of low dose and quality (Kurki & Morris, 2001; Haney,
2003; Mears & Watson, 2006; O’Keefe et al., 2011; Gendreau & Labrecque, 2015;
Haney et al., 2015; Richards, 2015). However, the precise magnitude of difference
between programming, or its effects, in restrictive housing and the general prison
population is unknown and has not been documented for individual states or
over time.

Ties to Family, Friends, and Communities
Inmates typically fear the loss of ties to family, friends, and their home
communities (Adams, 1992). The fear is justified — few inmates receive visitors
or sustain contact with their social networks outside the prison walls (Bales &
Mears, 2008; Cochran & Mears, 2013). This separation may weaken ties to family
and contribute to inmate misconduct, recidivism, and poor re-entry outcomes
(Mears & Bales, 2008; Cochran, 2012; Mears et al., 2012; Cochran & Mears,
2013; Cochran et al., 2014; Siennick et al., 2013). Inmates in restrictive housing
typically cannot or do not receive visits or telephone calls. However, the precise
effect of restrictive housing, and its varying duration, on social ties to others
remains largely unstudied.

Risk of Misconduct and Reoffending
While inmates reside in restrictive housing, they may change in ways that
contribute to the likelihood of a specific deterrent effect upon release from
the housing. At the same time, they are unlikely to receive rehabilitative
programming, services, or treatment, which means that restrictive housing
does little to address criminogenic factors that contribute to misconduct and
recidivism. To date, however, no studies have systematically examined the change
in an inmate’s propensity to engage in misconduct from the time of entry into
restrictive housing to release from it.

26 4 • National Institute of Justice | NIJ.gov

Impacts on Inmates During Re-entry Into General Population Prisons
Restrictive housing may affect inmates while they are confined to it, but it also may
affect them after they return to the general prison population. Here, again, a range
of outcomes may be relevant to evaluating the impacts of restrictive housing.

Protection — Less Victimization
There are no systematic or rigorous research studies that have documented the
effect of restrictive housing on victimization of inmates after they leave it. Some
research indicates that inmates may seek to be placed in restrictive housing
for their own safety (Lovell et al., 2000; Mears & Watson, 2006; Gendreau &
Labrecque, 2015). Indeed, it is also possible that placement in restrictive housing
may signal to other inmates that a particular inmate is vulnerable or a problem.
One inmate’s account indicates, for example, that “convicts look down on ‘checking
in [to restrictive housing]’” for protection or to detox, avoid paying a gambling
debt, or to “take a break” (Ferranti, 2015:55). It is also possible that placement
in restrictive housing may give an inmate a reputation as someone who is tough
and warrants challenging, or it may reduce the inmate’s ability to cope with other
people, thereby increasing the risk of victimization (Haney, 2003; Mears & Watson,
2006). However, no studies yet have empirically evaluated possibilities.

Misconduct
A stay in restrictive housing may deter inmates from future misconduct.
Conversely, it may antagonize them — what has been referred to as the “rage
hypothesis” (Ward & Werlich, 2003) — or increase their risk of misconduct.
Few studies have tested this idea. One study of inmates, which used a matching
design, found that short-term stays in solitary confinement as punishment did
not appear to increase infractions (Morris, 2015). The study has its limitations;
it focused on one large southern state, did not examine protective custody or
administrative management segregation, and excluded gang members and
inmates sentenced to capital punishment or life without parole. Accordingly, its
generalizability to other states or other uses of restrictive housing is unknown.
A separate study examined adult inmates in Ohio who were serving at least one
year in prison and were placed in solitary confinement for punishment. This
study, too, found no robust evidence that restrictive housing contributed to
subsequent misconduct (Labrecque, 2015).9 There was, however, some evidence
that, among mentally ill inmates and gang members, solitary confinement might
be associated with subsequent nonviolent or drug misconduct (Labrecque,
2015:113). Both studies stand out because they constitute the only studies to

9

The motivation for the Morris (2015) paper stemmed in part from the fact that no published empirical study had
examined the impact of restrictive housing, among the individual inmates exposed to it, on their subsequent inprison behavior. Labrecque’s (2015) study was undertaken after the Morris (2015) study was published.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 265

date to employ strong research designs to examine restrictive housing effects on
misconduct. Their generalizability to other populations (e.g., inmates in restrictive
housing for non-punitive reasons or for extended periods) or states is unknown.

Mental Health
The inference from extant research is that restrictive housing may adversely
affect inmate mental health prior to and after release from restrictive housing.
However, recent work and a meta-analysis raise questions about such claims
(O’Keefe et al., 2011; Gendreau & Labrecque, 2015). Research on the effects of
restrictive housing on inmate mental health has focused primarily on the period
of time in which the inmates reside in the housing. Many studies proceed from
the assumption that restrictive housing creates near-instantaneous effects on
mental health (O’Keefe et al., 2011; Gendreau & Labrecque, 2015). The effect
of restrictive housing on subsequent in-prison mental health outcomes has not
been systematically evaluated using “apples to apples” comparisons between
inmates exposed to restrictive housing and similar inmates who were not.
Some researchers have suggested that, though restrictive housing may have
harmful effects on inmate mental health, longer exposure to restrictive housing
is required (Kupers, 2008; O’Keefe et al., 2013). Research has not examined the
range of restrictive housing stays and conditions across different populations and
prison systems using matched comparison groups to assess such possibilities.

Physical Health
Studies have not systematically evaluated the effects of restrictive housing on
inmates’ physical health during or after release from it. Some commentaries and
qualitative accounts involving nonrepresentative samples suggest that placement
in restrictive housing might lead to a greater risk of victimization after release.
For example, an inmate may have acquired or reinforced an affiliation as a gang
member while in restrictive housing, or alternatively, that inmate may have
denounced an affiliation and so exposed himself or herself as a potential snitch
(Hunt et al., 1993; Kassel, 1998; Kurki & Morris, 2001; Briggs et al., 2003; Mears
& Watson, 2006; King et al., 2008; Richards, 2015).

Participation in Rehabilitative Programming
Little is known about the extent to which inmates released from restrictive
housing engage in or receive rehabilitative programming and whether receipt
of programming varies from what it otherwise would be if inmates had not
been placed there. To the extent that restrictive housing adversely affects mental
health, there would seem to be a need for more such programming. To the
extent that it contributes to increased misconduct, inmates may be more likely
to be returned to restrictive housing, thereby reducing their opportunities for
rehabilitative intervention even more. This possibility, however, remains largely
unexamined empirically.
26 6 • National Institute of Justice | NIJ.gov

Ties to Family, Friends, and Communities
Restrictive housing — particularly lengthy stays — may weaken or sever inmates’
ties to other inmates and to family members, friends, and others in their home
communities. To the extent that social ties are helpful in navigating and coping
with prison life, and that restrictive housing stays degrade these ties (Adams,
1992; Bottoms, 1999; Cochran, 2012; Mears & Cochran, 2015), inmates may
be more likely to experience adverse outcomes when they return to general
population facilities. This possibility, too, has not been examined empirically.

Risk of Reoffending
An inmate’s risk of recidivism might increase or decrease as a result of time in
restrictive housing, and this change might be greater among inmates who have
greater exposure — whether through total duration or frequency of placement —
to restrictive housing. Although several studies have examined restrictive housing
and recidivism, none has estimated changes in the risk of reoffending at the point
of release from restrictive housing to general population facilities or to society.

Impacts on Inmates During Re-entry Into Society
Recidivism
Few studies exist that use methodologically strong research designs, such as
matching analyses, to examine restrictive housing impacts on recidivism.10 One
exception is a study of supermax inmates in Washington state. David Lovell
and colleagues found that supermax incarceration was not associated with
recidivism; however, inmates released directly from supermax incarceration
were more likely to recidivate compared to inmates with supermax stays that
entailed first returning to a general inmate population facility before release
(Lovell et al., 2007). A related study of Washington state inmates that employed
a weaker research design found that restrictive housing was associated with a
greater likelihood of violent recidivism (Lovell & Johnson, 2004). Another study
examined supermax incarceration among Florida inmates. Using propensity
score matching, the study found a positive effect of supermax incarceration on
violent recidivism but no effect on general recidivism (Mears & Bales, 2009). It
also found no evidence that time in supermax housing or direct-release from it
to society exerted a greater effect on reoffending. Finally, a study by Daniel Butler
and colleagues found no significant effect of supermax housing on recidivism
(Butler et al., forthcoming, as reported in Steiner & Beard, 2015).

10

The motivation for the Morris (2015) paper stemmed in part from the fact that no published empirical study had
examined the impact of restrictive housing, among the individual inmates exposed to it, on their subsequent inprison behavior. Labrecque’s (2015) study was undertaken after the Morris (2015) study was published.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 267

Collectively, these few studies suggest that restrictive housing may increase
violent recidivism. Whether the total duration in or timing of release from
restrictive housing matters is unclear. Indeed, the main conclusion mirrors
that for other outcomes — too few methodologically strong studies have been
conducted to state with confidence the effect of restrictive housing. Extant
studies not only are few in number, but they also do not systematically evaluate
the impact of various dimensions of restrictive housing, such as the effects of
using restrictive housing for protective custody or punishment or the potential
for effects to vary according to the characteristics of inmates or the conditions in
restrictive housing.

Employment
Prison may adversely affect the employment prospects of individuals (Mears
& Cochran, 2015). Empirical research has not systematically examined how
different types of prison experiences may differentially influence employment
outcomes upon release. Short-term stays in restrictive housing would appear,
on the face of it, to exert little appreciable effect on such outcomes. Longerterm stays, however, might do so by limiting participation in vocational and
educational programming as well as in re-entry planning. Restrictive housing
also might affect employment outcomes by adversely affecting inmates’ mental
health. Regardless, to date only one study has examined restrictive housing’s
effect on employment; it found no evidence that such housing harmed
employment after an inmate is released into society (Butler et al., forthcoming, as
reported in Steiner & Beard, 2015). It is presently unknown what would be found
in state-by-state studies of varying doses of restrictive housing and of different
approaches to the design and operation of restrictive housing.

Mental Health
Research that examines mental health outcomes among inmates during their
stays in restrictive housing has been hampered by methodological limitations,
and few studies have examined mental health outcomes among inmates after
they leave restrictive housing and return to general population facilities. No
methodologically rigorous studies have examined the long-term effects of
restrictive housing on inmate mental health during incarceration and after re­
entry into society.

Physical Health
No empirical research has used strong study designs to systematically examine
the impacts of restrictive housing on the physical health outcomes of inmates
when they return to society. The accounts of gang members suggest the
possibility, one not evaluated empirically, that placement in restrictive housing
may increase the risk of victimization when inmates return to society (Hunt

268 • National Institute of Justice | NIJ.gov

et al., 1993; Kassel, 1998; Kurki & Morris, 2001; Briggs et al., 2003; Mears &
Watson, 2006; King et al., 2008; Richards, 2015).

Ties to Family, Friends, and Communities
Here, again, there is no empirical research that draws on strong methodological
research designs to estimate the effect of restrictive housing on ties to family,
friends, and communities.

Impacts on Punishment (Retribution)
One of the central justifications for, or goals of, restrictive housing is to punish
inmates (King, 1999; Riveland, 1999; Lovell et al., 2000; Kurki & Morris, 2001;
Neal, 2003; Shalev, 2009; Browne et al., 2011; Mears, 2013; Beck, 2015; Morris,
2015). Such punishment can be viewed as instrumental in promoting specific or
general deterrence. However, it also clearly has been and is viewed as a goal in
and of itself. Whether restrictive housing achieves a desired level of retribution
has not, to date, been empirically evaluated by researchers. To do so would
require an empirically based, explicitly articulated calculus of a given set of
sanctions, including restrictive housing. This type of “science of punishment” —
or, more specifically, science of retribution — has not been developed (Mears et
al., 2015:706). If applied to restrictive housing, any such science would need to
examine the amount of perceived punishment among those sent to restrictive
housing as well as the amount of perceived punishment among other inmates
and the prison administration. In so doing, it would need to take into account
the fact that some inmates may seek placement in restrictive housing, while
others may be greatly harmed by it. Without such information, the extent to
which restrictive housing achieves the goal of retribution will remain unclear —
both conceptually and empirically.

Impacts on Gang Influence
Some prison systems have used restrictive housing as a means by which to control
gangs and reduce their influence on prison operations and violence (Ralph &
Marquart, 1991; National Institute of Corrections, 1997; Ward & Werlich, 2003;
Pizarro & Stenius, 2004; Mears & Castro, 2006; Mears & Reisig, 2006; Shalev, 2009;
Mears & Watson, 2006; Naday et al., 2008; Sundt et al., 2008; Butler et al., 2012;
Reiter, 2012; Mears, 2013; McGinnis et al., 2014; Richards, 2015). To date, there
are no methodologically rigorous evaluation of the impact of restrictive housing
on gang influence. Such studies would need to examine rates of misconduct before
and after the use of restrictive housing or in relation to changes in how much it is
used, and they should address potential confounding from other strategies that
prison systems simultaneously employed to address gang activity. The latter issue is
critical because when prison systems face a crisis of any kind — such as riots or a

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 269

dramatic increase in violence or disorder — they will employ a range of strategies.
That approach is understandable but makes it difficult to isolate the unique effects
of restrictive housing.
Existing accounts point to ways in which restrictive housing may decrease or
increase gang influence and violence. Beneficial effects may arise when restrictive
housing allows prison officials to inhibit gang leadership. Harmful effects may
arise when the leadership is replaced, new gangs emerge to fill a void, reliance on
restrictive housing distracts officials from focusing on potential root causes of
gang problems or prison violence, and so on. However, extant research has not
addressed these methodological issues and, therefore, provide little credible basis
on which to claim that restrictive housing affects gang activity in either the short
term or long term.11

Impacts on Society
Restrictive housing might benefit society by improving re-entry outcomes
and by enabling the prison system to operate more effectively and efficiently.
For example, if staff can better perform their roles and obligations and
rehabilitative programming can be better implemented, an improvement might
occur in aggregate inmate behavior and recidivism, as well as housing, mental
health, and employment outcomes (Mears & Reisig, 2006). Such benefits might
arise from improvements among inmates sent to restrictive housing or general
population inmates.
Conversely, restrictive housing may harm society by worsening outcomes for
segregated housing and general population inmates and officers.12 It might
distract officials and divert funding from the causes of the problems that
restrictive housing seeks to address such as violence, disorder, and gangs.

11

Thoughtful attempts to study restrictive housing impacts on gang activity exist and paint a mixed portrait. Ward
and Werlich’s (2003) account suggests that prison wardens view restrictive housing as effective in controlling
gangs (see also Mears & Castro, 2006). Paige Ralph and James Marquart (1991) examined whether use
of Texas supermax facilities to house gang leaders affected gang violence; they found little evidence that it
did so (as compared to Austin et al., 1998). Geoffrey Hunt and colleagues (1993) interviewed ex-prisoners
who reported ways in which placing gang members in restrictive housing might increase prison violence.
One mechanism is that doing so may create a “vacuum” of power in the inmate culture that allows new and
potentially more dangerous prison gangs to emerge. The study did not directly examine systemwide effects
of restrictive housing on gang-related activity. Other empirical accounts that rely on non-random samples or
qualitative research methodologies exist and identified mixed findings (see, generally, Kassel, 1998; Bottoms,
1999; Briggs et al., 2003; Mears & Watson, 2006; Pizarro & Narag, 2008; Austin & Irwin, 2012; O’Donnell,
2014; Richards, 2015).

12

“[Solitary confinement] is not solely a corrections issue; the overwhelming majority of people incarcerated will be
released, and the impact of long periods of isolation on their health, employability, and future life chances will be
felt in the families and communities to which they return. It is important to understand the health impacts of the
widespread use of segregation at the population level, in addition to assessing the effect of time spent in solitary
confinement on individual health outcomes” (Cloud et al., 2015:21; see also Mears, 2013).

270 • National Institute of Justice | NIJ.gov

It also might engender a culture of abuse among officers and of defiance among
inmates. In addition, it might harm inmates’ families by eliminating or greatly
reducing contact between families and inmates. Finally, use of restrictive housing
might harm society by releasing inmates to communities who are less equipped
to succeed than they otherwise might be.
These potential benefits and harms are speculative. No systematic, methodologically
rigorous evaluations have examined their existence or magnitude. What is known
is that prison-based behavioral programs can reduce inmate misconduct and that
this benefit can be sustained during their transition into the community (French &
Gendreau, 2006). Accordingly, the benefits or harms of restrictive housing may have
effects on inmates that also extend into the community.

Impacts of Not Using Restrictive Housing When Needed
When restrictive housing may be needed but is not used, the clear risk of
harm arises. For example, if such housing serves as the only viable option for
preventing or halting a prison riot, the failure to use it — or to have it available
to use — may enable a riot to happen or lead to longer riots. Presently, there is
no clear empirical basis for establishing when or how much restrictive housing
is needed to achieve various goals. Accordingly, there is no clear basis on which
to estimate the harms that may arise from having no — or enough — restrictive
housing, or from failing to use it at all or using it in an insufficient dose to
achieve a particular goal.
This situation applies to virtually all of the goals associated with restrictive
housing. For example, if the goal is to punish inmates, there may be a
consequence of not using restrictive housing, but that consequence is not clear.
Many options exist for punishing inmates. In such a context, the only basis on
which to estimate the impacts of not having restrictive housing is to determine
precisely which punishments are available and which are appropriate for various
acts of misconduct. Prison systems may designate and follow a hierarchy
of sanctions, but that practice merely affirms a policy; it does not establish
empirically the extent to which restrictive housing is needed or, in turn, the
impacts of not employing the housing.
Similarly, if the goal is to protect certain inmates, there may be a consequence of not
having or using restrictive housing. However, the consequence cannot be established
empirically without first knowing what factors contributed to the need to protect
inmates and what other approaches can be adopted to address the problem.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 271

Impacts of Using Restrictive Housing When It May Not Be Needed
When restrictive housing is used but not needed, a range of potential harms may
arise, including the diversion of resources from more cost-effective approaches
to managing prison systems and inmates, creation of a culture of abuse among
officers, and emergence of a culture of inmate hostility and defiance. Other
possibilities include greater harm to officers, higher rates of mental illness among
inmates, more inmate misconduct, and worse re-entry outcomes for all inmates,
especially for those who have experienced restrictive housing. There is no
empirical research that directly identifies the extent of such harms or the extent
to which restrictive housing is not used as an option of last resort (Mears, 2013).
Studies have documented potential harms of restrictive housing, but they have
not systematically and simultaneously evaluated the range of benefits and harms
associated with its use. Accordingly, those studies primarily draw attention to a
particular benefit or harm without demonstrating the net benefit or harm relative
to a particular counterfactual (Mears & Watson, 2006).

Cost-Efficiency
Impacts alone are irrelevant if they are small, are offset by substantial financial
costs, or could be achieved more effectively through alternatives. To date, no
research documents the effect of the cost-efficiency of restrictive housing on
the overall efficiency of jails and prison systems (Lawrence & Mears, 2004;
Pizarro & Stenius, 2004; Mears, 2006, 2008b, 2013; Shalev, 2009). Two important
conceptual issues, discussed below, bear on evaluating cost-efficiency.

Restrictive Housing as Presumptive Policy or as Alternative
In the early 1980s, prior to the expansion of restrictive housing, a central policy
question was “What is the effect of increasing the use of restrictive housing?” That
question remains relevant in contemporary times. However, the widespread use of
restrictive housing also raises a related question: “What is the effect of decreasing
the use of restrictive housing?” The first question examines the effects of increasing
restrictive housing in a context where such housing is little used; the second
examines the effects of decreasing restrictive housing in a context where it is widely
used. Two implications flow from these opposing circumstances.
First, in both instances, research on the effectiveness of restrictive housing is
lacking. Second, the most relevant research question is not necessarily whether
restrictive housing should be retained until evidence emerges that more effective
and cost-efficient alternatives exist. That question makes restrictive housing
appear to constitute the presumptive policy. The equally relevant question
begins with presuming that restrictive housing constitutes an alternative to other

272 • National Institute of Justice | NIJ.gov

policies for protecting, punishing, and managing inmates, and for achieving
prison system goals.

The Cost-Efficiency of Restrictive Housing and Other Approaches
Prison systems seek to achieve a range of goals (e.g., protecting, punishing, and
managing inmates, and creating a safe and orderly environment) that not only
minimizes harms to inmates and staff but also prepares inmates for re-entry into
society. Prison systems can employ a wide range of approaches to achieve these
goals. Approaches include —
• Better inmate classification.
• Dispersing certain inmates strategically throughout the prison system.
• Increasing and improving staff training, education, and professionalism.
• Improving inmate culture.
• Improving officer culture.
• Ensuring fair and consistent enforcement of rules.
• Implementing and using incentives to motivate inmates to comply with prison rules.
• Providing evidence-based cognitive-behavioral treatment and rehabilitative,
vocational, and educational programming.
• Structuring the prison environment and activities in ways that reduce
opportunities for misconduct and promote rule conformity and participation in
programming (DiIulio, 1987; Logan, 1993; Sparks et al., 1996; Gendreau et al.,
1997; Reisig, 1998; Bottoms, 1999; Gendreau & Keyes, 2001; Kurki & Morris,
2001; Briggs et al., 2003; Irwin, 2005; French & Gendreau, 2006; Mears, 2008a,
2008b, 2013; Pizarro & Narag, 2008; Sundt et al., 2008; Useem & Piehl, 2008;
Cullen et al., 2014; Browne et al., 2015; Wooldredge & Steiner, 2015).
How cost-efficient is each approach, or combination of approaches, to achieving
prison system goals? That question remains largely undetermined. Instead,
research documents that some approaches effectively improve certain outcomes
in certain contexts. No rigorous comparative empirical analyses of the effects of
the different approaches exist. Yet, the fact that such a diversity of approaches
exists highlights the problem of assuming that restrictive housing is needed or
that it is more effective and cost-effective than other approaches at achieving
various prison system goals.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 273

Critical Research Gaps and Questions
An abundance of research gaps and questions on restrictive housing have been
identified by researchers.13 Addressing them presents many opportunities to
advance science and policy. For example, knowledge about the conditions that
create safe and orderly prisons would contribute to theory and scholarship
aimed at understanding how organizations operate. It also would contribute to
corrections officials’ efforts to improve safety and order in their prison systems.
Below are some of the most critical empirical research gaps and questions that
bear directly or indirectly on efforts to understand the impacts of restrictive
housing on inmates and the institutional environment of prison systems. The
gaps and questions can be organized broadly into five categories: need, theory,
implementation, impacts, and cost-efficiency. They include, however, additional
areas that relate to restrictive housing impacts and efforts to improve the science
and policy on this housing.

Restrictive Housing Classification
Restrictive housing discussions are severely hampered by inconsistent definitions
of the housing and its design and goals. Definitions and classifications are not
correct or incorrect, but rather more or less useful. Which definitions and
classifications are most useful in ensuring that appropriate inmates are sent to
restrictive housing to achieve particular goals?

Clarification of the Goals and Need for Restrictive Housing
The goals relate directly to the need for restrictive housing. Clear goals have
ripple effects along many dimensions, including identifying appropriate inmates
for restrictive housing, the design of the housing, and its impacts. Thus, what are
the precise goals of restrictive housing? Which goals should be weighted more
heavily than others? For each goal, how many inmates fit the profile of those who
require restrictive housing?
13

A non-exhaustive listing includes the following: Cooke (1989); Ward (1995); King (1999, 2005, 2007); Rivel&
(1999); Lovell & colleagues (2000); Kurki & Morris (2001); Briggs & colleagues (2003); Henningsen & colleagues
(2003); Haney (2003); Neal (2003); Toch (2003); Ward & Werlich (2003); Pizarro & Stenius (2004); Mears
(2005, 2006, 2008a, 2008b, 2013); Cloyes et al. (2006); Mears & Castro (2006); Mears & Reisig (2006); Mears
& Watson (2006); Smith (2006, 2008); Lovell & colleagues (2007); King & colleagues (2008); Kupers (2008);
Naday & colleagues (2008); Pizarro & Narag (2008); Sundt & colleagues (2008); Kupers & colleagues (2009);
Mears & Bales (2009, 2010); Shalev (2009); Browne & colleagues (2011); Butler & colleagues (2012); Reiter
(2012); Mears & colleagues (2013); McGinnis & colleagues (2014); O’Donnell (2014); Pizarro & colleagues
(2014); Baumgartel & colleagues (2015); Cloud & colleagues (2015); Frost & Monteiro (2015); Gendreau &
Labrecque (2015); Labrecque (2015a); Morris (2015); Richards (2015); Valera & Kates-Benman (2015); U.S.
Department of Justice Work Group (2016).

274 • National Institute of Justice | NIJ.gov

Effective Use of Steps to Reduce the Need for Restrictive Housing
Inmates do not act in isolation. Their behavior typically derives from their
individual characteristics and the settings in which they reside. Ignoring the
effects of these settings — including the composition of the inmate population
and prison conditions, programming, staffing, and management — contributes
to inmate misconduct. This in turn creates an apparent need to seek recourse
with incapacitating measures, such as restrictive housing. For any given goal
associated with restrictive housing, to what extent have other approaches for
achieving the goals been pursued? Have they been implemented well? If they have
not been pursued or implemented well, then there is likely a need for improvements
that target these approaches, rather than a need for restrictive housing.

The Theory of Restrictive Housing
The theory underlying the use of restrictive housing for any of a variety of goals
remains poorly developed, which creates missed opportunities to use, design,
and modify restrictive housing to maximize benefits and minimize harms.
What, then, are the most credible theoretical grounds on which to anticipate
that restrictive housing — or a particular design (e.g., duration, deprivations,
programming) — will improve any given outcome (e.g., reduced gang influence,
reduced violence among the most violent inmates, improved systemwide safety
and order)?

Adherence to Protocols, Rules, and Procedures
Numerous protocols, rules, and procedures guide restrictive housing operations,
in part to protect against lawsuits (Collins, 2004; Naday et al., 2008; Baumgartel
et al., 2015). To what extent do states and the federal government — and various
prison facilities — administer restrictive housing in ways that fulfill these
different operational requirements and that comport with the Constitution and
court rulings (King et al., 2008; Reiter, 2012; Mears, 2013)?

Amount and Quality of Services, Treatment, and Privileges in
Restrictive Housing
Prison systems typically purport to offer services, treatment, and some privileges
to inmates in restrictive housing. To what extent do these systems actually do so?
What is, and what affects, the amount and quality of such amenities? What is the
amount and quality of mental health counseling and treatment?

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 275

Admissions to and Releases From Restrictive Housing
Restrictive housing operates largely in the equivalent of a black box (Mears, 2005,
2006, 2013; Butler et al., 2012). Little systematic or comprehensive empirical
analysis exists — by state or federal prison system and over time — in the
actual use of restrictive housing. Some exceptions exist. For example, a study of
Washington state inmates in restrictive housing found that they varied greatly in
their characteristics. Some had extensive histories of violence and others did not.
Approximately one-third had a serious mental illness. Many younger inmates
appeared to behave in ways that would result in their placement in intensive
management housing as a way to protect themselves. Some inmates spent little
time in restrictive housing while others spent most of their total prison stay in it
(Lovell et al., 2000).
A study of Florida supermax inmates found that 55 percent had experienced
three or more stays in supermax housing (Mears & Bales, 2010). For 44 percent
of supermax inmates, their time in restrictive housing constituted less than 15
percent of their total term of incarceration. Fourteen percent of the supermax
inmates spent more than half of their incarceration in supermax housing. One
inmate in four (28 percent) had been in supermax housing within three months
of their release to society. The study also found that younger inmates and black
inmates were more likely to be placed in supermax housing; this difference
stemmed primarily from these groups engaging in more of the behaviors that
lead to such confinement. (The study did not examine whether these inmates
were differentially managed and treated by the prison system, which could
contribute to such behavioral differences.)
A study of California supermax inmates identified similar variations and patterns
(Reiter, 2012). These included widely variable durations — ranging from months
to 10 years or more — spent in supermax housing. Other findings included
a greater likelihood that Hispanic inmates resided in restrictive housing,
approximately one-third or more of inmates being released directly from
supermax housing to the streets, and considerable variability in patterns across
supermax facilities.
In 2015, BJS released its national study of restrictive housing, which provided
a one-time snapshot of restrictive housing in 2011-2012. The report provided
more representative and extensive details about restrictive housing inmates than
has ever been published (Beck, 2015). As with the 2014 Yale/ASCA survey of
corrections administrators, the study documented considerable heterogeneity
in the inmate population in restrictive housing and the use of restrictive
housing across facilities (Baumgartel et al., 2015). Inmates who spent time in
restrictive housing were more likely to be young, lesbian, gay, or bisexual; to have
committed a violent offense; and to have a mental illness. Other studies point to
additional factors associated with restrictive housing placement, such as gang
membership, infractions, and prior stays in segregation (Motiuk & Blanchette

276 • National Institute of Justice | NIJ.gov

2001; O’Keefe et al., 2011; McGinnis et al., 2014; Helmus, 2015). Some prisons
rarely place inmates in restrictive housing, while others do so frequently (King,
1999; Mears, 2006; Baumgartel et al., 2015).
These studies have advanced understanding about the use of restrictive housing
— including who is admitted to and released from it — but they provide only
one-time snapshots and leave many questions unaddressed. For states and the
federal government, critical questions remain to be answered. In every instance,
data are needed by year to establish changes in patterns and to explain what
accounts for variations in the use of restrictive housing over time. For example,
to what extent are policies for restrictive housing fully implemented? What are
the demographic, social, criminal, and mental and physical health characteristics
of individuals in restrictive housing? What characteristics or behaviors lead
to such confinement? What other strategies or options are pursued with these
inmates? How frequently and how long do individuals spend time there? To what
extent are admission and release processes fair? Which facilities use restrictive
housing more than others? Which inmates in restrictive housing experience
more benefits and which experience higher rates of self-injury or harm? What
factors explain variation across these different dimensions? For example,
if minorities are more likely to be in restrictive housing, what explains the
difference? How have patterns of restrictive housing use changed over time, and
what factors explain such variation?

Impacts of Restrictive Housing
For state or federal prison systems, what are the impacts of restrictive housing
for a range of outcomes? Few credible studies of impact exist, and those that do
focus on one state, point in time, or outcome. What is needed are studies that
employ rigorous research methodologies, including appropriate comparison
groups or conditions identified through matching or related procedures, that
estimate restrictive housing impacts on a range of outcomes. These include
systemwide prison safety and order, gang influence, inmate protection, perceived
and objective severity of punishment, and inmate outcomes during and after
restrictive housing (e.g., misconduct, self-injury, mental and physical health,
family ties). Also needed are reviews of inmate outcomes upon release from
restrictive housing to society (e.g., recidivism, employment, reunification with
family), as well as restrictive housing’s effects on the inmates’ families and on
public health and safety. Other outcomes to be examined include prison system
operations and staff (e.g., effects on available programming or on staff), the
effects of not using restrictive housing when it is needed, and the effects of using
restrictive housing when it is not needed.
In each instance, information is needed on the magnitude of the impacts and the
features, including the dose and timing, of restrictive housing that create them.
For example, does release directly from restrictive housing affect recidivism?

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 277

If so, is the effect greater than that on inmates who are first transitioned from
restrictive housing to general population prison facilities before release into the
community (Lovell et al., 2007; Mears & Bales, 2009; Reiter, 2012)? Research is
also needed on how inmates perceive the experience of restrictive housing. What
aspects of the experience — such as duration, staff-inmate interactions, treatment
and services, re-entry preparation — might be changed to maximize benefits and
minimize harms (Cooke, 1989; Crouch & Marquart, 1989; Bottoms, 1999; Kurki
& Morris, 2001; Mears, 2008a, 2008b, 2013; Gendreau & Labrecque, 2015)?

Conditions Under Which Restrictive Housing Achieves Intended Goals
Impact evaluations provide an estimate of the effect of a given policy on one
or more outcomes. They do not necessarily provide insight into why exactly
the impacts arose. The more general research question, then, is: Under what
conditions does restrictive housing achieve its various intended goals? For
example, are there certain durations of exposure to restrictive housing that
must occur for inmates to be protected or for systemwide safety and order to
be increased? What percentage of inmates must be placed in restrictive housing
to halt a riot or to prevent one? Under what conditions does litigation improve
restrictive housing use and its impacts?

Conditions Under Which Unintended Harms Can Be Minimized
Restrictive housing may contribute to any of a range of unintended harms.
For example, it may worsen inmate mental health and possibly contribute to
recidivism and systemwide violence and disorder. A related question, then, is:
Under what conditions can all appreciable unintended harms be minimized
while maximizing any potential benefits?

Impacts of Duration, Frequency, and Recency of Restrictive Housing
The effect of time served in restrictive housing has gone largely unexamined.
Time served constitutes a critical issue because it directly relates to cost. Holding
inmates longer than necessary wastes scarce resources. If lengthier stays create
more benefits, then longer stays may be warranted. Conversely, if they cause
more harm, then two negative conditions occur — scarce resources are wasted
and their expenditure causes more harm than benefit.
No clear theoretical basis exists for establishing a precise amount of time in
restrictive housing that must occur to create particular outcomes. Accordingly, this
lack of theoretical basis, along with the apparent heterogeneity in time served, gives
rise to a series of questions. How long must individuals reside in restrictive housing
to achieve a particular benefit or to produce harm? What are the minimum

278 • National Institute of Justice | NIJ.gov

durations necessary to achieve benefits? What are the effects of repeated restrictive
housing placement? What are the effects on re-entry of the recency of placement
in restrictive housing? From a more macro-level perspective, what prison-specific
or systemwide numbers or percentages of inmates must be placed in restrictive
housing to achieve particular impacts? What are the effects of dramatic increases or
decreases in the use of restrictive housing?

Cost-Efficiency of Restrictive Housing Compared to Other Approaches
Restrictive housing stands as but one of myriad approaches that prison systems
can use to achieve various goals (DiIulio, 1987; Sparks et al., 1996; Reisig, 1998;
Bottoms, 1999; Gendreau & Keyes, 2001; Toch, 2003; French & Gendreau, 2006;
King, 2007; Mears, 2008a, 2008b, 2013; Useem & Piehl, 2008; Shalev, 2009;
Browne et al., 2011; Association of State Correctional Administrators, 2013; Ross,
2013; Baumgartel et al., 2014; Browne et al., 2015; Frakes, 2015; Shames et al.,
2015). How cost-efficient is restrictive housing relative to the other approaches?
Ideally, prison systems would have credible answers to this question prior to
investing in any particular approach, including restrictive housing. When they
already have these approaches and the housing, the relevant comparison can
be challenging. What in fact is a prison system willing or able to implement?
Can it build a new maximum-security facility? Hire more staff? Invest in more
rehabilitative programming? Answers to such questions may be dictated by
political considerations or management philosophies. In such instances, costefficiency analyses should focus on comparisons to approaches or changes that
are most likely to be implemented.

Special Populations, Prison Contexts, and Restrictive Housing
Programs and policies may have different effects on certain groups or in certain
contexts (Gendreau et al., 1997; Mears, 2010, 2013; Browne et al., 2015; Morris,
2015). The same possibility holds for restrictive housing, which may have more
beneficial or harmful effects for certain groups (e.g., female inmates, inmates
with mental illness, very young or older inmates). It also may have variable
effects in different prison systems. For example, in a prison system where
inmates view correctional administrators as wielding authority in a legitimate
manner, restrictive housing may serve as a useful tool for improving safety and
order. In a system where inmates view administrators as lacking in legitimacy,
restrictive housing may serve primarily to antagonize inmates and reinforce their
unwillingness to comply with rules. In short, what are the uses and effects of
restrictive housing for different populations and prison system contexts?

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 279

Juveniles and Restrictive Housing
The extent to which the juvenile justice system uses restrictive housing or
analogous housing is largely unknown. A study in California estimated that
“between 10-12 percent of wards were housed in units in which they were
confined to their rooms for 23 hours a day” (Krisberg, 2003:51; see, generally,
Richards, 2015). The prevalence of restrictive housing since the time of that study
or in other states, and the impact of restrictive housing on young people, their
families, or the juvenile justice system, remains largely unknown.

Jails and Restrictive Housing
The above-identified research gaps apply even more to the use of restrictive
housing in jails, about which almost no empirical research exists. One national
study found that the jail inmate population is as likely as the prison population
to be placed in restrictive housing and that inmates in jail segregation or prison
segregation share many characteristics (Beck, 2015). A study of Rikers Island,
one of the largest jails in the country, found that restrictive housing was widely
used and that its use often appeared to be inappropriate and harmful (Haney
et al., 2015). In general, though, restrictive housing in jails has not been well
studied, and it warrants attention because of the large numbers of individuals in
jails and the unique challenges that jails face.

Views and Opinions From Corrections Officers and Administrators,
Policymakers, and the Public
Restrictive housing is used in part because of the belief that the public supports
it and that correctional system officers and administrators find it necessary and
effective. Its use, too, stems from policymaker support. However, few studies have
systematically examined the views and opinions of these groups. Extant studies
suggest that uniform support for restrictive housing should not be assumed.
Corrections administrators and wardens, for example, can and do disagree about
the need for restrictive housing, its goals or effectiveness, and the conditions
under which it is needed and helpful (Wells et al., 2002; Mears & Castro, 2006;
Mears & Watson, 2006; Ferdik & McKee, 2015). Legislators, too, have varied
understandings and opinions about it (Mears, 2005, 2006; Pizarro & Narag, 2008;
Goode, 2012; Baumgartel et al., 2015). Similarly, public opinion varies. In a public
opinion study of Floridians, 80 percent of adults expressed support for supermax
housing (Mears et al., 2013). However, this support declined to 60 percent when
respondents were asked whether they supported its use if no public safety benefit
would occur. Approximately 30 percent of respondents viewed restrictive housing
as inhumane, but most (70 percent) disagreed. Views about supermax housing
varied among groups; for example, whites, men, political conservatives, and

280 • National Institute of Justice | NIJ.gov

individuals who strongly adhered to a philosophy of punishment as retribution
were more likely to support the use of supermax incarceration.
The salience of examining the opinions of the public, prison officers and
administrators, and policymakers stems from several considerations. Since
restrictive housing constitutes an extreme form of incarceration that many view
as inhumane, public opinion should perhaps be considered (Mears et al., 2013).
Prison officers and wardens work “on the ground” and may have unique insights
about particular inmates, the factors that contribute to particular problems, and
how best to address them. Not least, policymakers may correctly or incorrectly
understand public views and the problems that prison systems face; taking
stock of the public’s or practitioner’s views thus provides a platform on which to
ensure that policymakers more fully appreciate and understand the diverse set of
considerations that attend the use of restrictive housing.
In short, how do the public, policymakers, and corrections officers and
administrators view restrictive housing? What are their opinions about the causes
of particular correctional system problems and how best to address them? What
explains variation among these groups in their support for restrictive housing,
their views about the conditions under which it may be used appropriately, and
their opinion about its effectiveness in achieving various goals?

Ethical Concerns and How They Might Be Addressed
Restrictive housing has been criticized by many different groups and
organizations, domestic and international, for being inhumane and for being
operated in a procedurally unjust manner and in ways that harm inmates
(Grassian, 1983; Kurki & Morris, 2001; Haney, 2003; Mears & Watson, 2006;
Smith, 2006; King, 2007; King et al., 2008; Shalev, 2009; Katel, 2012; Mears et
al., 2013; Amnesty International, 2014; O’Donnell, 2014; Baumgartel et al.,
2015; Cloud et al., 2015; Haney et al., 2015; Richards, 2015; Valera & KatesBenman, 2015). Although it has withstood legal scrutiny for several decades,
lawsuits continue to be filed that challenge its use on constitutional grounds and
as violating human rights (Collins, 2004; Amnesty International, 2014; Haney
et al., 2015). Research is needed that systematically takes stock of the empirical
grounds on which the ethical concerns exist. Some individuals may view restrictive
housing as fundamentally unacceptable; conversely, others may view it as a moral
imperative when inmates engage in certain activities. It is possible, however, that
such views derive from misunderstandings about the need for restrictive housing;
the theory, uses, impacts, and cost-efficiency of restrictive housing; and the variety
of alternative sanctioning and management strategies available to corrections
officials. What views, then, do various groups — the public, corrections officials,
policymakers, advocacy organizations, and so on — hold about restrictive housing?
Are their assumptions about it empirically supported? What steps can state or
federal jails or prison systems take to address objective and perceived concerns

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 281

about the potential misuse, abuse, or harms of restrictive housing? How can they
address disparities in its use among racial and ethnic minorities or other social
and demographic groups (Mears & Bales, 2010; Reiter, 2012; Schlanger, 2013)?

How to Improve Research on Restrictive Housing
Conducting research on restrictive housing is not easy (Ward, 1995; Ward &
Werlich, 2003; Mears, 2006; Baumgartel et al., 2014; Agha, 2015). A significant
barrier to undertaking empirical studies and monitoring restrictive housing is
the limited research infrastructure in jail and prison systems. Typically, research
divisions in these settings are small and over-extended. They lack the time or
resources required to regularly monitor restrictive housing or evaluate its uses,
impacts, and efficiency; to conduct financial audits and case studies; or to survey
officers, wardens, or inmates about their views of and experiences with restrictive
housing. When outside researchers attempt to access jail or prison system data,
they frequently confront closed doors. Even when data are made available,
researchers typically confront a difficult situation — no codebook exists and the
data are complicated to understand and use.
These and other barriers create a situation in which studies of restrictive housing
necessarily either do not occur or occur rarely. In the latter instance, the studies
typically have limited usefulness because they apply to only one prison or state at
one point in time and illuminate only one particular issue. This situation can be
remedied in part through federal funding for research. However, the only long­
term viable solution for regular monitoring and assessment of the uses, impacts,
and cost-efficiency of restrictive housing is to significantly increase the research
capacity of states or state correctional systems.
The situation stands out in part because of longstanding calls for greater
government accountability and evidence-based policy. Neither can occur
without empirical research that continuously monitors and evaluates jail and
prison operations. There are examples of states using the results of empirical
research to modify their approaches to restrictive housing (Goode, 2012; Agha,
2015; Browne, 2015; Frakes, 2015), and, in general, corrections officials appear
to embrace research that assists them in better understanding their facilities
and how to improve their operations and effectiveness. These examples and the
receptiveness to (useful) research, along with calls for government accountability
and evidence-based corrections, lead to the question: Why have states not
invested more in the kinds of research and research infrastructure necessary
to create and support accountability and evidence-based practice in the use of
restrictive housing?

282 • National Institute of Justice | NIJ.gov

A “Top 10” List of Critical Research Questions That Should Be and Can Be Addressed
Few, if any, of the research gaps and questions discussed in this paper can be justified in an era
in which government accountability and evidence-based practice are promoted. Few, too, can
be investigated without a substantial infusion of funds to create the research infrastructure
necessary for monitoring and evaluating prison system operations. Even so, some research
questions can be prioritized on the basis that (1) they involve a critical concern, (2) they can
be evaluated, and (3) study results may be actionable in the near term. Any prioritization of
research questions necessarily involves subjective judgment. Even so, the importance of these
questions is reflected in the broader literature on restrictive housing.
Here, then, are 10 critical research questions that meet the above criteria and that hold
considerable potential for improving accountability and creating more effective and efficient
prison systems.
1. To what extent do states employ effective strategies for managing their prison systems and
limiting the use of restrictive housing to situations in which it is most needed?
2. What factors determine which inmates are placed in restrictive housing? To what extent is
restrictive housing placement affected by variation in officer-, warden-, or facility-specific use
of such housing? To what extent is such variation explained by inmate behavior?
3. What are the most important causes of prison violence and disorder? Compared to these
causes, what is the relative contribution of certain inmates to violence and disorder?
4. To what extent do restrictive housing placements result from a propensity among some
inmates to act violently or from poor administrative management practices or operations?
5. How effective and cost-efficient is restrictive housing relative to other approaches to managing
prisons and inmates?
6. How effective and cost-efficient is restrictive housing relative to other approaches to
punishing inmates?
7. What frequency and duration of restrictive housing create the most benefits and harms for
inmates? Which aspects of or experiences in restrictive housing contribute to these outcomes?
8. What are the characteristics of inmates who most benefit from restrictive housing? Which
inmates are most harmed by it? Why do these differences exist?
9. What are the short-term effects of restrictive housing on inmates while they are incarcerated?
What are the longer-term effects during re-entry? What is the cost-efficiency of these effects
relative to other prison management strategies or inmate sanctions?
10. To what extent do policymakers, the public, and prison administrators, staff, and officers
support the use of restrictive housing? Which factors influence their support?

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 283

Conclusion
Although research on restrictive housing has increased in recent decades, the
overarching finding of this paper is that too little credible empirical research
exists to state with confidence the need for such housing, its effects on inmates
or prison systems, or its cost-efficiency. A balanced evaluation of the impacts of
restrictive housing requires carefully examining the need for such housing, its
implementation, its effects on inmates and other groups, its effects on the prison
system at large, and its cost-efficiency.
Many critical research gaps exist along precisely these dimensions. Indeed, there
is little about restrictive housing that has been consistently evaluated or wellevaluated using rigorous research methodologies, such as quasi-experimental
designs that identify appropriate comparison groups or conditions. If the gaps
remain unaddressed, jail and prison systems risk wasting their resources and
missing opportunities to improve inmate, staff, and public safety. If appropriately
addressed, restrictive housing policies — and jail and prison management
policies, more generally — have the potential to rest on an evidence-based
foundation. Doing so would help correctional systems to be more accountable,
effective, and efficient.

References
Adams, K. (1992). Adjusting to prison life. Crime and Justice, 16, 275-359.
Agha, A. (2015). Partnering with corrections systems to assess segregation/isolation in
confinement. Washington, DC: American Society of Criminology Annual Meeting.
Amnesty International. (2014). Entombed: Isolation in the U.S. federal prison
system. London, England: Amnesty International.
Association of State Correctional Administrators. (2013). Restrictive housing status
policy guidelines. Hagerstown, MD: Association of State Correctional Administrators.
Atherton, E. (2001). Incapacitation with a purpose. Corrections Today 63(4), 100-103.
Austin, J., & Irwin, J. (2012). It’s about time: America’s imprisonment binge. (4th ed.)
Belmont, CA: Wadsworth.
Austin, J., Repko, S., Harris, R., McGinnis, K., & Plant, S. (1998). Evaluation of
the Texas Department of Criminal Justice administrative segregation population.
Washington, DC: National Council on Crime and Delinquency.
Baker, P., & Goode, E. (2015). Critics of solitary confinement are buoyed as
Obama embraces their cause. New York Times, July 22, p. A16.

28 4 • National Institute of Justice | NIJ.gov

Bales, W. D., & Mears, D. P. (2008). Inmate social ties and the transition to
society: Does visitation reduce recidivism? Journal of Research in Crime and
Delinquency, 45, 287-321.
Barak-Glantz, I. L. (1983). Who’s in the hole? Criminal Justice Review 8, 29-37.
Baumgartel, S., Guilmette, C. Kalb, J., Li, D., Nuni, J., Porter, D., Resnik, J., Camp,
C., & Camp, G. 2015. Time-in-cell: The ASCA-Liman 2014 National Survey of
Administrative Segregation in Prison. New Haven, CT: Yale Law School Liman
Program and the Association of State Correctional Administrators.
Beck, A. J. (2015). Use of restrictive housing in U.S. prisons and jails, 2011-12.
Washington, DC: Bureau of Justice Statistics.
Bidna, H. (1975). Effects of increased security on prison violence. Journal of
Criminal Violence, 3, 33-46.
Boin, A., & Rattray, W. A. R. (2004). Understanding prison riots: Towards a
threshold theory. Punishment and Society, 6, 47-65.
Booth, F. W., Roberts, C. K., & Laye, M. J. (2012). Lack of exercise is a major
cause of chronic diseases. Comprehensive Physiology, 2, 1143-1211.
Bottoms, A. E. (1999). Interpersonal violence and social order in prisons. Crime
and Justice, 26, 205-282.
Briggs, C. S., Sundt, J. L., & Castellano, T. C. (2003). The effect of supermaximum
security prisons on aggregate levels of institutional violence. Criminology, 41,
1341-1376.
Browne, A. (2015). A systems approach to reducing the use of segregation in
confinement. Washington, DC: American Society of Criminology Annual Meeting.
Browne, A., Cambier, A., & Agha, S. (2011). Prisons within prisons: The use of
segregation in the United States. Federal Sentencing Reporter 24, 46-49.
Browne, A. Hastings, A., Kall, K., & diZerega, M. (2015). Keeping vulnerable
populations safe under PREA: Alternative strategies to the use of segregation in
prisons and jails. New York, NY: National PREA Resource Center.
Bruton, J. H. (2004). The big house: Life inside a supermax security prison.
Stillwater, MN: Voyageur Press.
Burt, M. (1981). Measuring prison results: Ways to monitor and evaluate
corrections performance. Washington, DC: National Institute of Justice.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 285

Butler, H., Griffin, D. O. H., III, & Johnson, W. W. (2012). What makes you the
‘worst of the worst’? An examination of state policies defining supermaximum
confinement. Criminal Justice Policy Review 24, 676-694.
Butler, H. D., Steiner, B., Makarios, M., & Travis, L. (Forthcoming.) Assessing the
effects of exposure to supermax confinement on offender post-release behaviors.
The Prison Journal.
Cloud, D. H., Drucker, E., Browne, A., & Parsons, J. (2015). Public health and solitary
confinement in the United States. American Journal of Public Health, 105, 18-26.
Cloyes, K. G., Lovell, D., Allen, D. G., & Rhodes, L. A. (2006). Assessment
of psychosocial impairment in a supermaximum security unit sample.
Criminal Justice and Behavior, 33, 760-781.
Cochran, J. C. (2014). Breaches in the wall: Imprisonment, social support, and
recidivism. Journal of Research in Crime and Delinquency, 51, 200-228.
Cochran, J. C. (2012). The ties that bind or the ties that break: Examining the
relationships between visitation and prisoner misconduct. Journal of Criminal
Justice, 40, 433-440.
Cochran, J. C., & Mears, D. P. (2013). Social isolation and inmate behavior: A
conceptual framework for theorizing prison visitation and guiding and assessing
research. Journal of Criminal Justice, 41, 252-261.
Collins, W. C. (2004). Supermax prisons and the Constitution: Liability concerns in
the extended control unit. Washington, DC: National Institute of Corrections.
Cooke, D. J. (1989). Containing violent prisoners: An analysis of the Barlinnie
Special Unit. British Journal of Criminology, 29, 129-143.
Crouch, B. M., & Marquart, J. W. (1989). An appeal to justice: Litigated reform of
Texas prisons. Austin, TX: University of Texas Press.
Cullen, F. T., Jonson, C. L., & Stohr, M. K. (Eds.) (2014). The American prison:
Imagining a different future. Thousand Oaks, CA: Sage.
DeMaio, J. R. (2001). If you build it, they will come: The threat of overclassification
in Wisconsin’s supermax prison. Wisconsin Law Review 2001, 207-248.
DiIulio, J. J., Jr. (1987). Governing prisons. New York, NY: Free Press.
Durbin, D. (2013). Durbin statement on Federal Bureau of Prisons assessment of
its solitary confinement practices. Washington, DC: United States Senate.

286 • National Institute of Justice | NIJ.gov

Durbin, D. (2012). Reassessing solitary confinement: The human rights, fiscal
and public safety consequences. Opening Statement of Senator Dick Durbin
for the hearing before the United States Senate Committee on the Judiciary
Subcommittee on the Constitution, Civil Rights and Human Rights, June 19.
Dye, M. H. (2010). Deprivation, importation, and prison suicide: Combined
effects of institutional conditions and inmate composition. Journal of Criminal
Justice, 38, 796-806.
Ferdik, F. V., & McKee, K. T. (2015). The relationship between correctional officer
safety and wellness and administrative segregation. White paper. Washington,
DC: National Institute of Justice.
Ferranti, S. (2015). The realities of special housing units in the Federal Bureau
of Prisons. In S. C. Richards (Ed.), The Marion experiment: Long-term solitary
confinement and the supermax movement (pp. 35-58). Carbondale, IL: Southern
Illinois University Press.
Frakes, S. (2015). Reforming segregation by redefining prisons. Washington, DC:
American Society of Criminology Annual Meeting.
French, S. A., & Gendreau, P. (2006). Reducing prison misconducts: What works!
Criminal Justice and Behavior 33, 185-218.
Frost, N., and Carlos Monteiro, C. (2015). Administrative segregation in U.S.
prisons. White paper. Washington, DC: National Institute of Justice.
Gawande, A. (2009). Hellhole: The United States holds tens of thousands of
inmates in long-term solitary confinement. Is this torture? The New Yorker,
March 30, 36-45.
Gendreau, P., Goggin, C. E., & Law, M. A. (1997). Predicting prison misconducts.
Criminal Justice and Behavior 24, 414-431.
Gendreau, P., & Keyes, D. (2001). Making prisons safer and more humane
environments. Canadian Journal of Criminology 43, 123-130.
Gendreau, P., & Labrecque, R. M. (2015). The effects of administrative
segregation: A lesson in knowledge cumulation. In J. Wooldredge & P. Smith
(Eds.), Oxford handbook on prisons and imprisonment. New York, NY: Oxford
University Press.
Glaze, L. E., & Kaeble, D. (2014). Correctional populations in the United States, 2013.
Washington, DC: Bureau of Justice Statistics.
Goode, E. (2012). Prisons rethink isolation, saving money, lives and sanity.
The New York Times, March 11, p. A1.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 287

Grann, D. (2004). The brand: How the Aryan Brotherhood became the most
murderous prison gang in America. The New Yorker, February 16, LXXX, 157-171.
Grassian, S. (1983). Psychopathological effects of solitary confinement. American
Journal of Psychiatry, 140, 1450-1454.
Haney, C. (2003). Mental health issues in long-term solitary and ‘supermax’
confinement. Crime and Delinquency, 49, 124-156.
Haney, C., Weill, J., Bakhshay, S., & Lockett, T. (2015). Examining jail isolation:
What we don’t know can be profoundly harmful. The Prison Journal, 96(1), 126­
152. http://dx.doi.org/10.1177/0032885515605491.
Helmus, L. M. (2015). Developing and validating a risk assessment scale to
predict inmate placements in administrative segregation in the Correctional
Service of Canada. White paper. Washington, DC: National Institute of Justice.
Henningsen, R. J., Johnson, W. W., & Wells, T. 1999. Supermax prisons: Panacea
or desperation? Corrections Management Quarterly 3, 53-59.
Hershberger, G. L. (1998). To the max: Supermax facilities provide prison
administrators with more security options. Corrections Today 60, 54-57.
Hunt, G., Riegel, S., Morales, T., & Waldorf, D. (1993). Changes in prison culture:
Prison gangs and the case of the ‘Pepsi Generation.’ Social Problems 40, 398-409.
Irwin, J. (2005). The warehouse prison: Disposal of the new dangerous class.
Los Angeles, CA: Roxbury.
Kaba, F., Lewis, A., Glowa-Kollisch, S., Hadler, J., Lee, D., Alper, H., Selling,
D., MacDonald, R., Solimo, A., Parsons, A., & Venters, H. (2014). Solitary
confinement and risk of self-harm among jail inmates. American Journal of
Public Health, 104, 442-447.
Kane, M., Pierce, B., & Haynes, M. (2013). Restrictive housing FAQ. Boston, MA:
The Crime and Justice Institute.
Kassel, P. (1998). The gang crackdown in Massachusetts’ prisons: Arbitrary and
harsh treatment can only make matters worse. New England Journal on Criminal
and Civil Confinement, 24, 37-63.
Katel, P. (2012). Solitary confinement: Is long-term isolation of prisoners
inhumane? Congressional Quarterly 22, 767-788.
King, K., Steiner, B., & Breach, S. R. (2008). Violence in the supermax: A selffulfilling prophecy. The Prison Journal, 88, 144-168.

288 • National Institute of Justice | NIJ.gov

King, R. D. (2007). Imprisonment: Some international comparisons and the
need to revisit panopticism. In Y. Jewkes (Ed.), Handbook of prisons (pp. 95-102).
Portland, OR: Willan Publishing.
King, R. D. (2005). Effects of supermax custody. In A. Liebling & S. Maruna (Eds.),
The effects of imprisonment (pp. 118-145). Portland, OR: Willan Publishing.
King, R. D. (1999). The rise and rise of supermax: An American solution in
search of a problem? Punishment and Society, 1, 163-186.
Krisberg, B. (2003). General corrections review of the California youth authority.
Oakland, CA: National Council on Crime and Delinquency.
Kupers, T. A. (2008). What to do with the survivors? Coping with the long-term
effects of isolated confinement. Criminal Justice and Behavior 35, 1005-1016.
Kupers, T. A., Dronet, T., Winter, M., Austin, J., Kelly, L., Cartier, W., Morris, T.
J., Hanlon, S. F., Sparkman, E. L., Kumar, P., Vincent, L. C., Norris, J., Nagel, K.,
& McBride, J. (2009). Beyond supermax administrative segregation: Mississippi’s
experience rethinking prison classification and creating alternative mental health
programs. Criminal Justice and Behavior, 36, 1037-1050.
Kurki, L., & Morris, N. (2001). Supermax prisons. Crime and Justice: A Review of
Research 28, 385-424.
Labrecque, R. M. (2015a). The effect of solitary confinement on institutional
misconduct: A longitudinal evaluation. Washington, DC: U.S. Department of Justice.
Labrecque, R. M. (2015b). Who ends up in administrative segregation? A metaanalytic review. White paper. Washington, DC: National Institute of Justice.
Lanes, E. C. (2011). Are the ‘worst of the worst’ self-injurious prisoners more
likely to end up in long-term maximum-security administrative segregation?
International Journal of Offender Therapy and Comparative Criminology, 55,
1034-1050.
Latessa, E. J., Listwan, S. J., & Koetzle, D. (2014). What works (and doesn’t) in
reducing recidivism. Waltham, MA: Anderson Publishing.
Lawrence, S., & Mears, D. P. (2004). Benefit-cost analysis of supermax prisons:
Critical steps and considerations. Washington, DC: The Urban Institute.
Neal, D. (Ed.). (2003). Supermax prisons: Beyond the rock. Lanham, MD:
American Correctional Association.
Logan, C. H. (1993). Criminal justice performance measures for prisons. In
J. J. DiIulio, Jr., and James Q. Wilson (Eds.), Performance measures for the
criminal justice system (pp. 15-59). Washington, DC: Bureau of Justice Statistics.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 289

Lovell, D. (2008). Patterns of disturbed behavior in a supermax population.
Criminal Justice and Behavior, 35, 985-1004.
Lovell, D., Cloyes, K., Allen, D., & Rhodes, L. (2000). Who lives in super-maximum
custody? Federal Probation, 64, 33-38.
Lovell, D., Johnson, L. C., & Cain, K. C. (2007). Recidivism of supermax
prisoners in Washington State. Crime and Delinquency, 53, 633-656.
Martin, R., & Zimmerman, S. (1990). A typology of the causes of prison riots
and an analytical extension to the 1986 West Virginia riot. Justice Quarterly 7,
711-737.
McCorkle, R. C., Miethe, T. D., & Drass, K. A. (1995). The roots of prison
violence: A test of the deprivation, management, and not-so-total institution
models. Crime and Delinquency, 41, 317-331.
McGinnis, K., Austin, J., Becker, K., Fields, L., Lane, M., Maloney, M., ... Felix,
T. (2014). Federal Bureau of Prisons: Special housing unit review and assessment.
Arlington, VA: CNA Analysis and Solutions.
Mears, D. P. (2004). Mental health needs and services in the criminal justice system.
Houston Journal of Health Law and Policy, 4, 255-284.
Mears, D. P. (2005). A critical look at supermax prisons. Corrections
Compendium 30, 6-7, 45-49.
Mears, D. P. (2006). Evaluating the effectiveness of supermax prisons. Washington, DC:
The Urban Institute.
Mears, D. P. (2008a). Accountability, efficiency, and effectiveness in corrections:
Shining a light on the black box of prison systems. Criminology and Public Policy, 7,
143-152.
Mears, D. P. (2008b). An assessment of supermax prisons using an evaluation
research framework. The Prison Journal, 88, 43-68.
Mears, D. P. (2009). Supermax incarceration and recidivism. Criminology, 47, 801-836.
Mears, D. P. (2010). American criminal justice policy: An evaluation approach to
increasing accountability and effectiveness. New York, NY: Cambridge University Press.
Mears, D. P. (2013). Supermax prisons: The policy and the evidence. Criminology
and Public Policy 12, 681-719.
Mears, D. P., & Bales, W. D. (2010). Supermax housing: Placement, duration, and time
to reentry. Journal of Criminal Justice, 38, 545-554.

29 0 • National Institute of Justice | NIJ.gov

Mears, D. P., & Castro, J. L. (2006). Wardens’ views on the wisdom of supermax prisons.
Crime and Delinquency 52, 398-431.
Mears, D. P., & Cochran, J. C. (2012). U.S. prisoner reentry health care policy in
international perspective: Service gaps and the moral and public health implications.
The Prison Journal, 92, 175-202.
Mears, D. P., & Cochran, J. C. (2015). Prisoner reentry in the era of mass incarceration.
Thousand Oaks, CA: Sage Publications.
Mears, D. P., Cochran, J. C., & Cullen, F. T. (2015). Incarceration heterogeneity and its
implications for assessing the effectiveness of imprisonment on recidivism. Criminal
Justice Policy Review 26, 691-712.
Mears, D. P., Cochran, J. C., Greenman, S. J., Bhati, A. S., & Greenwald, M. A. (2011).
Evidence on the effectiveness of juvenile court sanctions. Journal of Criminal Justice, 39,
509-520.
Mears, D. P., Cochran, J. C., Siennick, S. E., & Bales, W. D. (2012). Prison visitation and
recidivism. Justice Quarterly, 29, 888-918.
Mears, D. P., Mancini, C., Beaver, K., & Gertz, M. (2013). Housing for the ‘worst of
the worst’ inmates: Public support for supermax prisons. Crime and Delinquency 59,
587-615.
Mears, D. P., & Reisig, M. D. (2006.) The theory and practice of supermax prisons.
Punishment and Society, 8, 33-57.
Mears, D. P., & Watson, J. (2006). Towards a fair and balanced assessment of supermax
prisons. Justice Quarterly, 23, 232-270.
Morris, R. G. (2015). Exploring the effect of exposure to short-term solitary
confinement among violent prison inmates. Journal of Quantitative Criminology, 32(1).
http://dx.doi.org/10.1007/s10940-015-9250-0.
Morris, R. G., Carriaga, M. L., Diamond, B., Piquero, N. L., & Piquero, A. R.
(2012). Does prison strain lead to prison misbehavior? An application of General
Strain Theory to inmate misconduct. Journal of Criminal Justice, 40, 194-201.
Motiuk, L. L., & Blanchette, K. (2001). Characteristics of administratively segregated
offenders in federal corrections. Canadian Journal of Criminology, 43, 131-143.
Mushlin, M. B. (2012). Unlocking the courthouse door: Removing the barrier of
the PLRA’s physical injury requirement to permit meaningful judicial oversight of
abuses in supermax prisons and isolation units. Federal Sentencing Reporter, 24,
268-275.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 291

Naday, A., Freilich, J. D., & Mellow, J. (2008). The elusive data on supermax
confinement. The Prison Journal, 88, 69-93.
National Institute of Corrections. (1997). Supermax housing: A survey of current
practice. Washington, DC: National Institute of Corrections.
National Public Radio. (2012). The grim realities of life in supermax prisons.
Talk of the Nation, hosted by Neal Conan, June 21. Washington, DC: National
Public Radio.
Neal, D. (Ed.). (2003). Supermax prisons: Beyond the rock. Lanham, MD:
American Correctional Association.
Obama, B. (2016). Why we must rethink solitary confinement. The Washington
Post, January 25, p. A15.
O’Donnell, I. (2014). Prisoners, solitude, and time. New York, NY: Oxford
University Press.
O’Keefe, M. L. (2008). Administrative segregation from within: A corrections
perspective. The Prison Journal, 88, 123-143.
O’Keefe, M. L., Klebe, K. J., Metzner, J., Dvoskin, J., Fellner, J., & Stucker, A.
(2013). A longitudinal study of administrative segregation. Journal of the
American Academy of Psychiatry and the Law, 41, 49-60.
O’Keefe, M. L., Klebe, K. J., Stucker, A., Sturm, K., & Leggett, W. (2011). One
year longitudinal study of the psychological effects of administrative segregation.
White paper. Washington, DC: U.S. Department of Justice.
Paternoster, R. (2010). How much do we really know about criminal deterrence?
Journal of Criminal Law and Criminology, 100, 765-824.
Pizarro, J. M., & Narag, R. E. (2008). Supermax prisons: What we know, what we
do not know, and where we are going. The Prison Journal, 88, 23-42.
Pizarro, J. M., & Stenius, V. M. K. (2004). Supermax prisons: Their rise, current
practices, and effect on inmates. The Prison Journal, 84, 248-264.
Pizarro, J. M., Zgoba, K. M., & Haugebrook, S. (2014). Supermax and recidivism:
An examination of the recidivism covariates among a sample of supermax
ex-inmates. The Prison Journal, 94, 180-197.
Preer, R. (2004). The solitary men. The Boston Globe Magazine, April 4, pp. 30-37.
Prins, S. J. (2014). Prevalence of mental illnesses in U.S. state prisons: A
systematic review. Psychiatric Services, 65, 862-872.

292 • National Institute of Justice | NIJ.gov

Ralph, P. H., & Marquart, J. W. (1991). Gang violence in Texas prisons.
The Prison Journal, 71, 38-49.
Reisig, M. D. (1998). Rates of disorder in higher-custody state prisons: A
comparative analysis of managerial practices. Crime and Delinquency 44, 229-244.
Reiter, K. A. (2012). Parole, snitch, or die: California’s supermax prisons and
prisoners, 1997-2007. Punishment and Society 14, 530-563.
Rhodes, L. A. (2004). Total confinement: Madness and reason in the maximum
security prison. Los Angeles, CA: University of California Press.
Richards, S. C. (Ed.). (2015). The Marion Experiment: Long-term solitary
confinement and the supermax movement. Carbondale, IL: Southern Illinois
University Press.
Riveland, C. (1999). Supermax prisons: Overview and general considerations.
Washington, DC: National Institute of Corrections.
Ross, J. I. (Ed.). (2013). The globalization of supermax prisons. New Brunswick,
NJ: Rutgers University Press.
Rossi, P. H., Freeman, H. E., & Lipsey, M. W. (2004). Evaluation: A systematic
approach. 7th ed. Thousand Oaks, CA: Sage.
Schlanger, M. (2013). Prison segregation: Symposium introduction and
preliminary data on racial disparities. Michigan Journal of Race and Law, 18, 1-12.
Shalev, S. (2009). Supermax: Controlling risk through solitary confinement.
London, England: Willan Publishing.
Shames, A., Wilcox, J., & Subramanian, R. (2015). Solitary confinement: Common
misconceptions and emerging safe alternatives. New York, NY: Vera Institute of Justice.
Sherman, L. W. (1993). Defiance, deterrence and irrelevance: A theory of the
criminal sanction. Journal of Research in Crime and Delinquency 30, 445-473.
Siennick, S. E., Mears, D. P., & Bales, W. D. (2013). Here and gone: Anticipation
and separation effects of prison visits on inmate infractions. Journal of Research
in Crime and Delinquency 50, 417-444.
Singer, P. (2003). Warden: Inmates volunteer to stay at ‘supermax’ prison.”
Associated Press, March 3.
Smith, P. S. (2008). ‘Degenerate criminals’: Mental health and psychiatric studies
of Danish prisoners in solitary confinement, 1870-1920. Criminal Justice and
Behavior, 35, 1048-1064.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 293

Smith, P. S. (2006). The effects of solitary confinement on prison inmates: A brief
history and review of the literature. Crime and Justice, 34, 441-528.
Sparks, R., Bottoms, A. E., & Hay, W. (1996). Prisons and the problem of order.
Oxford, England: Oxford University Press.
Steiner, B., & Beard, J. (2015). The relationship between institutional violence and
administrative segregation. White paper. Washington, DC: National Institute of Justice.
Stickrath, T. J., & Bucholtz, G. A. (2003). Supermax prisons: Why? In D. Neal
(Ed.), Supermax prisons: Beyond the rock (pp. 1-14). Lanham, MD: American
Correctional Association.
Suedfeld, P. (1974). Solitary confinement in the correctional setting: Goals,
problems, and suggestions. Corrective Social Psychiatry, 141, 10-20.
Suedfeld, P., Ramirez, C., Deaton, J., & Baker-Brown, G. (1982). Reactions and
attributes of prisoners in solitary confinement. Criminal Justice and Behavior 9,
303-340.
Sundt, J. L., Castellano, T. C., & Briggs, C. S. (2008). The sociopolitical context of
prison violence and its control: A case study of supermax and its effect in Illinois.
The Prison Journal, 88, 94-122.
Sykes, G. M. (1958). The society of captives. Princeton, NJ: Princeton University Press.
Toch, H. (2003). The contemporary relevance of early experiments with
supermax reform. The Prison Journal, 83, 221-228.
U.S. Department of Justice Work Group. (2016). U.S. Department of Justice report
and recommendations concerning the use of restrictive housing. Washington, DC:
U.S. Department of Justice Work Group on Restrictive Housing.
U.S. Government Accountability Office. (2013). Improvements needed in Bureau
of Prisons’ monitoring and evaluation of segregated housing. Washington, DC: U.S.
Government Accountability Office.
Useem, B., & Kimball, P. (1991). States of siege. New York, NY: Oxford University Press.
Useem, B., & Piehl, A. M. (2008). Prison state: The challenge of mass incarceration.
New York, NY: Cambridge University Press.
Valera, P., & Kates-Benman, C. L. (2015). Exploring the use of special housing
units by men released from New York correctional facilities: A small mixedmethods study. American Journal of Men’s Health, March 10. http://dx.doi.
org/10.1177/1557988315569880.

29 4 • National Institute of Justice | NIJ.gov

Ward, D. A. (1995). A corrections dilemma: How to evaluate super-max regimes.
Corrections Today, 57, 104-108.
Ward, D. A., & Werlich, T. G. (2003). Alcatraz and Marion: Evaluating supermaximum custody. Punishment and Society, 5, 53-75.
Wells, T. L., Johnson, W. W., & Henningsen, R. J. (2002). Attitudes of prison
wardens toward administrative segregation and supermax prisons. In L. F.
Alarid & P. F. Cromwell (Eds.), Correctional perspectives (pp. 171-180).
Los Angeles, CA: Roxbury.
Wooldredge, J., & Steiner, B. (2015). A macro-level perspective on prison inmate
deviance. Punishment and Society, 17, 230-257.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 295

CHAPTER 8

The Effect of Administrative
Segregation on Prison Order
and Organizational Culture
Jody Sundt, Ph.D.
Indiana University — Purdue University Indianapolis

Introduction

O

rder is the central challenge of the well-managed prison, and the use of
separation, in varying degrees of restrictiveness, features prominently
in efforts to control the institution and its inhabitants. Indeed, how to
classify, isolate, organize, and discipline prisoners to the “habits of order” (de
Beaumont & de Tocqueville, 1833) has preoccupied correctional officials for
centuries (Rothman, 1971). Early penitentiaries placed great faith in the ability
of social isolation — enforced by austere architecture and authoritarian regimes
— to impose a physical order even in the absence of an inmate’s moral reform.
Experiments with solitary confinement during the 1800s soon revealed, however,
the inhumanity and ineffectiveness of isolation, and the practice was largely
abandoned (Toch, 2003). Today, we again confront questions about the appropriate
role of segregation and search for the lines of demarcation between its legitimate
use as a management strategy, its overuse, and its potential to cause harm.
Managing the risks that prisoners pose and maintaining order are the primary
purposes of classification decisions to place individuals in prison facilities

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 297

that differ by security level and organizational regime. “The very nature of our
prisons,” observes Cohen (2008), “means we must have some means by which
to separate prisoners on the basis of those who are at risk from those who create
those risks” (p. 1017). Thus, prisoners who present little risk to others and are
unlikely to attempt escape are placed in minimum-security prisons; high-risk
prisoners are placed in maximum-security prisons. Administrative segregation
is a security classification for managing prisoners who are considered too
dangerous or too disruptive to be housed among the general inmate population.
We can distinguish administrative segregation from other types of separation and
isolation by its purpose: to control individuals who may pose a current or future
threat (Metcalf, Morgan, Oliker-Friedland, Resnik, Spiegel, Tae, … Holbrook,
2013). In contrast, the purpose of disciplinary (or punitive) segregation is to
punish inmates who engage in misconduct; protective custody is used to isolate
inmates for their own safety.1 A supermax facility is a stand-alone prison or a
unit within a prison built or retrofitted specifically for the purpose of segregation
(King, 1999; Riveland, 1999b).
Inmates held in segregation are typically confined to a cell for 23 to 24 hours per
day, often behind solid doors. Segregation cells are austere. Inmates are allowed
one hour of exercise five times per week, typically alone in small pens with
metal fencing. Access to such personal items as family photographs and reading
material is often restricted or denied. When inmates must be moved, they are
shackled, chained, and escorted by two or more correctional officers. Meals are
delivered within the cell, as are religious and educational services, sometimes
via closed-circuit television. Visits and phone calls are sharply limited and
closely monitored, or disallowed completely (see Cohen, 2008; U.S. Government
Accountability Office, 2013; Kurki & Morris, 2001; Metcalf et al., 2013; Reiter,
2012; Toch, 2001).
The primary objective of administrative segregation is to improve prison order
and safety by removing dangerous inmates from the general population and more
effectively managing them in isolation (Metcalf et al., 2013; also see Mears &
Watson, 2006). For this reason, administrative segregation and supermax units are
sometimes referred to as “prisons within prisons.” They are intended to serve a dual
purpose: to incapacitate inmates and to deter them from future misconduct.
The number of inmates housed outside of the general prisoner population in
some type of segregated housing has increased precipitously in the past decade.
A recent study conducted by the Yale University Liman Program and the
1

In practice, these distinctions may lose their meaning. Inmates may seek safety in administrative and punitive
segregation, without the stigma of protective custody. Inmates in distress who harm themselves or attempt
suicide may be punished for their actions, which violate prison rules. They also may be put in administrative
segregation because they are disruptive and may harm themselves. Rumors that an inmate “snitched” may result
in placement in protective custody. Regardless of why inmates are in segregated housing, the living conditions
are similar, such as keeping inmates in a cell for 23 hours per day.

298 • National Institute of Justice | NIJ.gov

Association of State Correctional Administrators (ASCA) (2015) estimates that
between 80,000 and 100,000 inmates were held in restricted housing in 2014.
These numbers — coupled with longstanding concerns about the legal and
ethical dimensions of the practice — have contributed to an emerging consensus
that segregated housing is overused (Liman Program & ASCA, 2015; Mohr &
Raemisch, 2015).
Despite the sense that there are too many inmates housed in isolation units,
observers disagree about which inmates require segregation, when segregation
is appropriate, whether it is an effective tool, or whether viable alternatives exist.
This paper, commissioned by the National Institute of Justice, Office of Justice
Programs, United States Department of Justice, assesses the research on the
effectiveness of administrative segregation as a management strategy, including
its effect on prison order and organizational culture. The paper also outlines a
broad research agenda to help fill knowledge gaps and to learn how best to use
this controversial strategy to maintain secure, humane correctional institutions
that serve the public safety goals of our nation. These issues are relevant to
anyone interested in better understanding the organizational effects
of administrative segregation.
Corrections officials have argued that administrative segregation improves prison
safety. The review of the research discussed below demonstrates, however, that its
effects are inconsistent: sometimes improving order (Sundt, Castellano, & Briggs,
2008), sometimes making it worse (Briggs, Sundt, & Castellano, 2003; Labrecque,
2015; Wooldredge & Steiner, 2015), but mostly having no effect (Briggs et al., 2003;
Huebner, 2003; Sundt et al., 2008; also see Labrecque, 2015; Morris, 2015).
Scholars also question how reliance on administrative segregation affects the
organizational culture of prisons and shapes correctional employees’ roles and
work experience. Haney (2008), for example, warns of the deleterious effects
that a “culture of harm” has on prison staff and their ability to work effectively
and humanely with inmates in segregation. Correctional employees who work
in segregation may be exposed to high levels of stress and trauma, which may
contribute to destructive attitudes, the loss of professional skills, excessive use
of force, and burnout. Some correctional leaders question whether the focus
on isolating disruptive inmates detracts from the prisons’ public safety mission
(Mohr & Raemisch, 2015). More broadly, scholars caution that coercive strategies
that rely on the use or threat of force may erode the legitimacy of prison
management and lead to more, not less, prison disorder (Colvin, Cullen, &
Vander Ven, 2002; Liebling, 2004; Sparks & Bottoms, 1995).
The next section examines the concepts that link the use of administrative
segregation to prison order and reviews the research on the relationship between
systemwide prison order and segregation. The latter portion of the review
explores the relationship between administrative segregation and organizational

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 29 9

culture. Finally, the paper considers why so little research exists on this topic and
concludes by recommending research priorities.

Administrative Segregation and Systemwide Order
Administrative segregation may affect prison order through three mechanisms
(Mears & Reisig, 2006). First, it may incapacitate inmates by removing them
from the general prison population, thereby reducing their opportunity to
engage in serious misconduct. Second, administrative segregation may deter
inmates from serious misconduct because of its promise for swift, certain, and
severe punishment. Third, segregation may normalize facilities by removing
troublesome inmates from the general prison population.

Administrative Segregation as Incapacitation
Incapacitation is the primary means by which administrative segregation is
expected to improve systemwide prison safety. Severe restrictions on inmate
movements and social interactions, the use of technology to control and surveil
inmates, and stringent limitations on inmate property all reduce opportunities
for inmates to assault others and engage in serious misconduct. Removing
disruptive inmates from the general population — limiting their ability to
interact with others and to access contraband and information — is expected to
reduce the incidence of serious and violent misconduct.
Some evidence (Barak-Glantz, 1983; Bennett, 1976; Flanagan, 1983; Porporino,
1986; Toch, 1997; Toch & Adams, 1986), and much anecdotal information,
support the theory that a small number of inmates is responsible for the majority
of prison violence, lending credence to the claim that a policy of selective
incapacitation may lower overall rates of prison violence. Bennett (1976) found,
for example, that just 2 percent of the inmates held at San Quentin in 1960 were
responsible for all the violent incidents that year. In a series of more rigorous
studies, Toch (1997) and colleagues (Toch & Adams, 1986) demonstrated
that “disturbed disruptive” inmates often struggle to adapt to prison and are
responsible for a disproportionate amount of disruption and violence.
Using segregation less strategically — placing enough inmates in segregation
for a sustained period in a process analogous to collective incapacitation — may
reduce systemwide disorder. During the mid-1980s, for example, Texas placed all
known and suspected gang members in administrative segregation, regardless of
whether they were involved in an incident of serious misconduct. Although the
Texas policy was not rigorously evaluated, Ralph and Marquart (1991) observed
declines in the number of inmate murders and assaults (with and without
weapons) following this change. Similarly, Irwin and Austin (1997) and Crouch

30 0 • National Institute of Justice | NIJ.gov

and Marquart (1989) credited the decline in inmate violence in California and
Texas to these states’ extensive use of segregation (but see Useem & Piehl, 2006).
In contrast, a number of studies indicate that rather than eliminating
opportunities to engage in misconduct, administrative segregation may
exacerbate it (Toch & Kupers, 2007). Rhodes’ (2004) ethnography of segregation
units in Washington state, for example, documented that “through practices
that yield more trouble the tighter their hold, the prison tends to secrete the
very thing it most tries to eliminate.” Extreme control measures may result
in extreme reactions and acts of resistance. Flooding and setting cells on fire;
breaking furniture and cell fixtures; throwing and smearing blood, semen,
urine, and feces; riots; hunger strikes; self-mutilation; and suicide have all been
documented in segregated housing units, from the earliest penitentiaries (Rubin,
2015) to contemporary prisons (Edge & Jones, 2014; Human Rights Watch, 1997;
King, Steiner, & Breach, 2008; Kupers, Dronet, Winter, Austin, Kelly, Cartier,
… Vincent, 2009; Reiter, 2012; Rhodes, 2004). As Gawande (2009) explained,
inmates who experience prolonged isolation “begin to see themselves primarily
as combatants in the world, people whose identity is rooted in thwarting prison
control” (paragraph 38). Thus, inmates may engage in acts of resistance to
maintain a sense of purpose and identity in situations of severe social isolation
(Rhodes, 2004; Toch & Adams, 1986).
Data from ethnographies, historical documents, and interviews about the high
levels of violence and disorder found in segregation units are consistent with
more quantitative reports. Bidna (1975) found rates of stabbings in California’s
secure housing units higher than in the general prison population, a difference
that was statistically significant in the 1973-1974 study period, but not in the
1972-1973 study period. Bidna attributed the increased rate of assaults in secure
housing units between 1972 and 1974 to a statewide crackdown on prison
violence that included, among other policy changes, locking down the state’s
four maximum-security prisons. Similarly, a 1986 California task force report
noted that rates of violence in “special housing (lockup) units” were particularly
high. Rates of violence at Folsom State Prison’s special housing unit were more
than twice the rate for the mainline unit. Rates of violence in San Quentin State
Prison’s lockup unit were 60 percent higher than in its mainline unit (California
Department of Corrections, 1986).
Further research points to a similar pattern. In a study of the Canadian prison
system, Porporino (1986) found that, between 1980 and 1984, close to onethird of all self-directed violence and one-third of all property damage occurred
in administrative segregation, even though it held only 5 percent of the total
inmate population. Rates of assaultive behavior and general disruption were
also disproportionately high in segregated housing. Similarly, more than half
of all serious assaults against staff in Texas occurred in segregated housing and
other close-control environments (Sorensen, Cunningham, Vigen, & Woods,

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 301

2011). The Sorensen team also found that a large proportion of violent assaults at
medium-security prisons occurred within segregation cells.
Commenting on this pattern, Porporino (1984) concluded that the
“concentration of violent incidents in higher security correctional settings
suggest a simple, though often overlooked,” fact.
Efforts to maintain order and control through more restrictive security can
attain only limited success in curbing the incidence of prison violence. In the
extreme, such measures may increase the motivation to engage in violence or
prod the ingenuity of inmates and result in more extreme violence (p. 218).
Thus, rather than reducing systemwide violence through incapacitation,
segregation may simply change the location and form of the disorder and
violence (Bidna, 1975; Sundt, Castellano, & Briggs, 2004) or amplify serious
misconduct (Toch & Kupers, 2007).

Administrative Segregation as Deterrence
The significant deprivations associated with administrative segregation
may also deter inmate misconduct. Inmates may be generally deterred by
the threat of administrative segregation. Commenting on the opening of a
supermax prison, for example, an Illinois prison official argued, “The majority
of inmates will detest this place. … How much they detest it is going to be the
key to how successful it is” (Hallinan, 1995). Similarly, those who are placed
in administrative segregation may be specifically deterred by the experience,
persuaded to never again warrant a return to segregation.
Speculation about the specific deterrent effects of segregation emphasizes the
severity of conditions, extreme deprivation of social contact and basic amenities,
and length of stay. For example, the U.S. Supreme Court upheld a Pennsylvania
policy that denied inmates in long-term administrative segregation access to
newspapers, magazines, and personal photographs, holding that the policy was
reasonably related to the goal of motivating better behavior (see Cohen, 2008).
Research on the specific deterrent effects of short-term punitive segregation
and solitary confinement does not support the idea that isolation motivates
good behavior. In two early studies, Barak-Glantz (1983) found no relationship
between initial placement in punitive segregation and subsequent placements
there, and Suedfeld (1974) concluded that punitive isolation was not related to
“productive change” in inmates. Two recent, methodologically rigorous studies
by Morris (2015) and Labrecque (2015) confirm these conclusions.
Using a matched sample of 1,834 inmates from a large southern state, Morris
(2015) found that punitive segregation had no effect on the probability, timing,
or trajectory of violent misconduct. Similarly, Labrecque (2015) determined that
302 • National Institute of Justice | NIJ.gov

neither the experience of punitive segregation nor the length of time spent there
affected subsequent involvement in violent misconduct, nonviolent misconduct,
or drug use among a sample of Ohio inmates.
Notably, Labrecque (2015) detected some effect heterogeneity. Gang-affiliated
inmates who experienced punitive segregation were 10 percent more likely to
engage in violent misconduct and 14 percent more likely to engage in nonviolent
misconduct. Inmates with mental illnesses were 23 percent more likely to engage
in nonviolent misconduct and 24 percent more likely to engage in drug use
following a term in punitive segregation. In contrast, inmates convicted of a drug
offense were 28 percent less likely to commit an act of nonviolent misconduct
after returning to the general prison population than were inmates convicted of a
nonviolent offense.
Both Morris (2015) and Labrecque (2015) restricted their analyses to misconduct
following punitive segregation, which shares some conditions of confinement
with administrative segregation but likely differs in other important ways.
Specifically, the length of stay in administrative segregation is much greater
and inmates may interpret the experience differently. Nevertheless, Morris’s
(2015) and Labrecque’s (2015) findings are consistent with a well-established
body of research demonstrating that the severity of punishment has little effect
on whether someone will reoffend. Instead, deterrence effects are mediated by
perceptions and other factors such as stakes in conformity (Pratt, Cullen, Blevins,
Daigle, & Madensen, 2006). To date, there is no empirical evidence that links
deprivation or the restrictive conditions of confinement to improved inmate
behavior. There are, however, some hints that the effects of confinement may vary
among subgroups of inmates.
Little is known about how consistently punitive segregation or other types of
restricted housing are used. Doubts about the certainty of punishment are fed by
complaints that segregation is used arbitrarily and that the criteria for entry and
exit from segregation are vague or nonexistent (Metcalf et al., 2013; Toch, 2007).
Correctional officers’ wide discretion in reporting and responding to misconduct
(Bottoms, 1999) may undermine the certainty of punishment. In the only study
to examine the effect of certainty of punitive segregation, Huebner (2003)
found that the proportion of inmates in a prison facility who were punished
with solitary confinement for a rule violation was not related to the frequency
of assaults on inmates or on staff. Clearly, much more research is needed to
understand the risk of punishment for prison misconduct and how the certainty
of sanctions influences inmate behavior.

Administrative Segregation as Normalization
Finally, administrative segregation may normalize the general prison
population by incapacitating the “bad apples” who instigate misconduct among

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 303

other inmates: the normalization hypothesis. The practice may also free up
organizational resources and staff attention to focus on the routine maintenance
of order and service provision. Officials from the Federal Bureau of Prisons
maintain, for example, that segregated housing can reduce the use of facility
lockdowns (locking general population inmates in their cells as a security
measure), which are costly and require staff to perform custodial duties rather
than other tasks (GAO, 2013, p. 33).
Lockdown days may be considered a proxy of disorder — prisons are locked
down when there is a threat to safety or security. It is also reasonable to assume
that the fewer days that inmates in general population spend locked in their cells,
the more likely they are to engage in programs and access services. Sundt and
colleagues (2008) examined lockdown use in the Illinois prison system, testing
whether the opening of a supermax prison had a normalizing effect on the state’s
other prisons. Between 1996 and 1998, Illinois’ 26 prisons were locked down an
average of 55 times per month. The results of an interrupted time-series analysis
indicated that opening the supermax prison resulted in 29 fewer lockdown days
per month, a decline of 52 percent, which lends strong support for the potential
normalizing effect of administrative segregation.
Additional research is needed to determine whether Sundt and colleagues’
(2008) findings can be generalized and replicated. In addition, their method did
not allow them to identify the mechanism that linked the use of administrative
segregation to fewer lockdown days. More direct tests of the predictions derived
from the normalization hypothesis are needed to reach a conclusion about the
effect of segregation on systemwide prison operations.

The Total Effect of Administrative Segregation on Systemwide
Prison Violence
Three studies have directly tested the effect of administrative segregation on
systemwide levels of prison violence: Briggs and colleagues (2003), Sundt and
colleagues (2008), and Wooldredge and Steiner (2015). Analogous to research
on imprisonment and crime (Travis, Western, & Redburn, 2014), these studies
tested the total effect of administrative segregation on prison violence and
disorder. This strategy has the advantage of capturing simultaneously all
incapacitating, normalizing, and deterrent effects, and the disadvantage of being
unable to identify the specific mechanisms that link administrative segregation to
prison safety and order.
Briggs and colleagues (2003) evaluated the effect of opening four supermax
prisons on systemwide levels of prison violence against inmates and staff in
three states — Arizona, Minnesota, and Illinois — using a quasi-experimental,
interrupted time-series design. The prisons in these states differed in some
important respects. Arizona and Illinois built stand-alone facilities for the

30 4 • National Institute of Justice | NIJ.gov

specific purpose of administrative segregation, and the supermax prisons in
both states are notoriously punitive and austere (see Reiter, 2012; Kurki &
Morris, 2001). Minnesota, however, retrofitted a prison to create its supermax
facility. Inmates at Minnesota’s Oak Park Heights facility were provided more
opportunities to participate in programs and were less socially and physically
isolated than were supermax inmates in Arizona and Illinois. The facilities also
differed in operating capacity, utilization rates, and the proportion of supermax
beds statewide. Arizona had the greatest supermax confinement capacity and the
highest certainty of placement, while Illinois had the lowest. Minnesota used its
supermax prison at a rate slightly higher than Illinois, and its inmates had more
opportunity to participate in programs.
The opening of Arizona’s two supermax prisons had no effect on statewide levels
of inmate safety. Its first supermax facility had no effect on overall staff safety;
however, a temporary (but significant) increase in assaults against correctional
staff resulting in injury occurred in the month after Arizona opened its second
supermax prison. Using Utah as a control to rule out regional effects, the analysis
confirmed that the spike in serious assaults against staff was unique to Arizona.
The analysis of Minnesota also found no relationship between the opening of
its supermax facility and statewide assaults against inmates or staff. In Illinois,
the statewide level of assault against inmates did not change after the supermax
facility opened; however, its opening did correspond with a gradual (but sustained)
improvement in statewide levels of staff safety. The Briggs team (2003) found 22
fewer assaults per month against staff after Illinois opened its supermax prison.
Sundt and colleagues (2008) further analyzed the effect of administrative
segregation on systemwide prison order in Illinois. They first examined a
security shakedown in 1996 that, among other effects, increased the number
of segregation cells in the state’s maximum-security prisons by 55 percent and
converted one of the maximum-security prisons to a segregation housing unit.
They then tested the effect of opening the Illinois supermax prison, controlling
for the first set of policy changes in 1996. The team found that the policy
changes implemented in 1996 resulted in approximately three fewer assaults per
10,000 inmates per month but had no effect on violence against staff. Sundt and
colleagues could not determine whether improved inmate safety was attributable
to the dramatic increase in the use of administrative segregation or to some other
aspect of the shakedown, such as better property control, drug testing, or better
security during visitation. The results regarding the effect of Illinois’ supermax
prison were largely unchanged from those reported by Briggs and colleagues
(2003). The analysis confirmed that opening the supermax unit at Tamms had
no effect on inmate safety but was associated with a significant decline in assaults
against staff.
Wooldredge and Steiner (2015) analyzed the direct and total effect of the
proportion of the total inmate population held in administrative and punitive
segregation on rates of assault and nonviolent misconduct in 247 prisons from

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 305

Table 1. Summary of Research on the Effect of Segregation on System-wide Levels
of Prison Violence
State

Intervention Tested

Inmate
Safety

Staff Safety

Opening of 960 Bed Supermax

No Effect

No Effect

Opening of 778 Bed Supermax

No Effect

Temporary (1 month)
increase of 6.5
assaults with injury

Minnesota

Opening of 120 Bed Supermax

No Effect

No Effect

Illinois

Opening of 500 Bed Supermax

No Effect

Gradual, sustained
decline of 22 assaults
per month

% inmate population receiving
disciplinary segregation for
most recent rule violation

No Effect

No Effect

Briggs et al. (2003)
Arizona

Huebner (2003)
National
Sample of
185 Prisons

40 states, using structural equation modeling. Contrary to expectations and
controlling for inmate risk and other organizational characteristics, they found
that segregation had a positive direct effect on rates of assault and nonviolent
misconduct. Specifically, “A greater use of coercive control actually coincided
with larger proportions of inmates who engage in assaults” (p. 244). Moreover,
when examining the pattern of indirect and direct effects, Wooldredge and
Steiner (2015) found that higher prison security was associated with greater
use of segregation, which led to more assaults. The use of segregation was also
positively associated with increased levels of nonviolent misconduct. These
findings, argue Wooldredge and Steiner, call into question assumptions about the
ability of security level and coercive controls to bring about prison order. They
recommend reconsidering the concentration strategy of prison management and
classification, and reevaluating the use of dispersion strategies to distribute highrisk inmates more evenly among facilities.

Summary of Results
It is difficult to draw conclusions from such a preliminary set of studies, most
of which speak only indirectly to the ability of administrative segregation to
achieve its objective of improved systemwide levels of prison safety and order.
Nevertheless, the research reveals some patterns and tentative conclusions about
the systemwide effect of administrative segregation on prison order.

30 6 • National Institute of Justice | NIJ.gov

Table 1. Summary of Research on the Effect of Segregation on System-wide Levels
of Prison Violence (continued)
State

Intervention Tested

Inmate Safety

Staff Safety

55% increase in ad seg. cells
at maximum security prisons;
Pontiac prison converted to
ad seg. facility; statewide
“shakedown”

Decline of 3
assaults/month
per 10,000
inmates

No Effect

Opening of 500 Bed Supermax

No Effect*

Gradual,
sustained
decline of
25 assaults/
month*

Increased rate
of assaults

No Effect

Increased rate
of nonviolent
misconduct

N/A

Sundt et al. (2008)
Illinois

Wooldredge & Steiner (2015)
National
sample of 247
prisons from
40 states

*

Proportion inmate population
in ad seg. and disciplinary seg.

Note: Results replicated findings from Briggs et al. (2003) with controls for 1996 “shakedown.”

First, although it appears that a small number of inmates are responsible
for a disproportionate amount of prison disorder, it is not clear whether
incapacitation can prevent inmate disruption and violence. Inmates incarcerated
within administrative segregation continue to engage in high rates of violence
and misconduct (Bidna, 1975; California Department of Corrections, 1986;
Porporino, 1986; Sorensen et al., 2011). It is possible that administrative
segregation merely concentrates inmate violence in specific locations within
the prison system, but there is also evidence that higher levels of security
exacerbate inmate misconduct and disorder (Gaes & Camp, 2009; Wooldredge
& Steiner, 2015). Administrative segregation may intensify some types of inmate
misconduct (Rhodes, 2004; Toch & Kupers, 2007), particularly among certain
types of offenders.
Second, there is neither support for the deterrent effect of punitive segregation
(Huebner, 2003; Labrecque, 2015; Morris, 2015) nor for administrative segregation.
Third, the effect of administrative segregation on systemwide levels of prison
violence is mixed (see table 1). Most of the evidence suggests that segregation
does not improve systemwide safety (Briggs et al., 2003; Huebner, 2003; Sundt

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 307

et al., 2008) and may contribute to increases in inmate misconduct under
some circumstances or among certain groups of offenders (Briggs et al., 2003;
Labrecque, 2015; Wooldredge & Steiner, 2015). The effect of administrative
segregation on the safety of correctional officers is also inconsistent — sometimes
improving staff safety, sometimes making it worse, but most frequently having no
effect (Briggs et al., 2003; Sundt et al., 2008).
Fourth, Sundt and colleagues (2008) found support for the argument that
administrative segregation normalizes prison systems. An analysis of Illinois
prisons found that opening a supermax facility substantially reduced the use of
lockdown days. More research is needed to determine whether this result can be
generalized and replicated in other locations.

Organizational Culture and Coercive Control
Questions about the effect of culture are woven throughout the canon of research
on prison violence and order. Early prison scholars were particularly interested in
the relationship between prison subcultures and the socialization of inmates and
a set of values and social roles (Clemmer, 1940; Sykes, 1958). A later generation
of scholars extended this work to examine the socialization of correctional
officers and the attitudes and occupational experiences that characterize prison
work (Jacobs & Retsky, 1975; Lombardo, 1981). The 1971 Stanford Prison
Experiment studied prison life by randomly assigning students to roles as guards
or prisoners for a 7- to 14-day trial. The results of this experiment (Haney, Banks,
& Zimbardo, 1973) brought to light the powerful contextual influence that prison
social roles have on the attitudes and behaviors of student “inmates” and “guards”
and their interpersonal interactions. In just a few days, the students began to
create a culture of control and resistance — “us versus them.”
Organizational culture is difficult to define and measure. “You had to be there,”
explains the insider, a phrase that captures the intuitive and latent qualities of
organizational culture (Liebling, 2004). Garland (1990) defined culture as an
idea that “refers to all those conceptions and values, categories and distinctions,
frameworks of ideas and systems of belief which human beings use to construe
their world and render it orderly and meaningful” (p. 194). Contemporary
theorists posit that culture is socially constructed, dynamic, expressive, and
relational (for a discussion, see Stowell & Byrne, 2008).
Organizations form cultures through shared social experiences. Schein (1990)
describes organizational culture as a pattern of assumptions that tells its
members the “correct way to perceive, think, and feel” about organizational
problems. Organizational culture serves an important function, explains Schein:

308 • National Institute of Justice | NIJ.gov

Once a group has learned to hold common assumptions, the resulting automatic
patterns of perceiving, feeling, and behaving provide meaning, stability, and
comfort; the anxiety that results from the inability to understand or predict
events happening around the group is reduced by the shared learning (p. 111).
These insights help to explain how organizational cultures form, their potential
effects, and why they emerge and persist.
Given the importance of the concept of culture to penology, surprisingly
little research has directly studied organizational culture (Byrne, Hummer, &
Taxman, 2008). An important exception is the work by Alison Liebling (2004)
that examined the “moral performance” of prisons. Liebling (2004) identified
unique organizational cultures in five prisons that could be scored on the
emphasis they placed on the values of security and harmony. Security values
included rule enforcement, use of authority, risk management, control practices,
and removal of privileges. Harmony values included respect, humanity, trust,
support, relationships, activity or personal development, and contact with family.
Good prison performance — and, by extension, good organizational cultures
— achieve balance between the two values, whereas “poor” prison performance
overly emphasizes either harmony or security.
Liebling’s (2004) work on the moral performance of prisons shares some
common ideas with Colvin and colleagues’ (2002) theory of differential coercion
and social support. As applied to prisons, Colvin (2007) hypothesizes that
using coercion in the absence of consent and social support increases, rather
than decreases, compliance and safety. Colvin defines coercion as the “force
that compels or intimidates an individual to act because of the fear or anxiety
it creates” (2007, p. 368). He argues that when social support for inmates is low
and coercion is used inconsistently, prisons will see higher levels of violence
and disorder as inmates become angry and direct their frustration at others.
When coercion is used consistently in the absence of low and inconsistent social
support, Colvin predicts that inmates will direct negative emotions inward,
engage in acts of self-harm, and experience mental health problems.
Colvin (2007) analyzed the organizational policies and climate that existed just
prior to the infamous riot at the New Mexico State Penitentiary in 1980.2 He
described an organization that provided low social support and used coercion
inconsistently (including the arbitrary use of punitive segregation), with an
organizational culture characterized by hostility between correctional officers
and inmates. This period of violent unrest is compared with earlier periods when
prison officials maintained a better balance of social support and coercion. Like

2

The 1980 riot at the New Mexico State Penitentiary was one of the most deadly, expensive, and violent prison
riots in U.S. history. Over two days, 33 inmates were killed, 400 were injured, and 12 correctional officers were
held hostage. Rioting inmates took advantage of serious security lapses to beat, rape, torture, and murder (see
Useem, 1985).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 309

Liebling (2004), Colvin (2007) recommends a management strategy that meets
the “Goldilocks test”: neither too lax nor too severe, but just right.
Theorists have also identified the importance of legitimacy and fairness in
achieving prison order (Liebling, 2004; Sparks & Bottoms, 1995; Useem & Piehl,
2006). Prisons are political communities that rely on cooperation and rest on the
belief that the authority of correctional officials and prison regimes is legitimate
and applied fairly. “Prisons cannot operate by force alone,” contend Useem and
Piehl (2006, p. 90). Moreover, when prisons lose legitimacy and prison regimes
are viewed as unfair, the results are despair (Liebling, 2011), violence, disorder
(Beijersbergen, Dirkzwager, Eichelsheim, Van der Laan, & Nieuwbeerta, 2015;
Carrabine, 2005; Bottoms, 1999; Sparks & Bottoms, 1995; Useem & Piehl, 2006)
and, occasionally, political resistance (Reiter, 2014).
The next section considers more fully the organizational cultures found in
administrative segregation. Haney (2008) warns of a culture of harm found in
administrative segregation that has deleterious effects on prison staff and their
ability to work effectively and humanely with inmates. Working in administrative
segregation exposes correctional staff to high levels of stress and trauma, which
may contribute to destructive attitudes, high levels of fear, the loss of professional
detachment and skill, excessive use of force, and burnout. Finally, the potential
effect of organizational culture on efforts to reform the use of administrative
segregation is considered.

The Organizational Culture of Administrative Segregation
The nature of organizational culture makes it a difficult phenomenon to study.
Schein (1990) notes that case studies, clinical descriptions, and ethnographies
are the best tools for its study at the beginning stage of research, because
these methods are more holistic and better able to capture the complexity of
assumptions and values. Only a handful of studies considers the organizational
culture of administrative segregation units. However, some inferences can be
drawn from official mission statements and analysis of the overt values expressed
by the characteristics of the organizational regime. Rhodes’ (2004) ethnography
of Washington state’s administrative segregation unit is a sophisticated and
carefully documented study of one state’s organizational culture. Court cases
and investigations provide another glimpse into some of the worst qualities of
the culture in segregated housing at various times (Cohen, 2008; King et al.,
2008; Simon, 2014). Finally, a handful of scholars has shared and reflected on
their direct observations of supermax culture (Haney, 2008; King, 1999, 2005;
Kurki & Morris, 2001). Among them, Haney’s (2008) discussion is a more
formal description of the causes and effects of the culture of harm inherent in
administrative segregation.

310 • National Institute of Justice | NIJ.gov

Organizational values and language are important makers of culture. An
overriding concern for maintaining security and managing levels of risk
and danger are among the most overt organizational values manifested
by administrative segregation. This aspect of administrative segregation is
consistent with the set of values that Feeley and Simon (1992) describe as the
“new penology,” which makes actuarial risk management of dangerous groups a
priority over individualized treatment or concern for due process. These values
are also consistent with what Liebling (2004) referred to as security values.
Rhodes (2004) characterized the administrative segregation prison as a security
utopia. Like the penitentiaries of the early 1800s, supermax prisons embody an
overriding faith in the ability of technology and architecture to enforce order and
protect staff (also see Reiter, 2012). Emphasis on security values is so dominant
that the supermax prisons in California were built without the physical space
necessary to meet inmates’ basic health care needs (Reiter, 2012; Simon, 2014).
The Colorado Department of Corrections constructed a $200 million supermax
facility that was open for mere months before a court closed it because it lacked
outdoor recreation areas (Prendergast, 2015). The culture of administrative
segregation emphasizes a single-minded concern for staff safety and risk control.
The dominant value of staff safety is institutionalized in ceremonies and rituals
that memorialize officers who were hurt or killed in the line of duty. It also is
powerfully communicated and promulgated in organizational stories about riots,
murders, and horrific acts of inmate brutality — all cited as reasons for needing
administrative segregation.
Technological sterility and efficiency are inherent in contemporary segregation
prisons and are reflected in their names, physical plants, and procedures.
Older names for penal isolation — the hole, the box, and lockup — have been
dropped in favor of technocratic names such as administrative maximum,
intensive management, behavior modification, and, most commonly,
administrative segregation (Metcalf et al., 2013). Newer segregation units
rely on computers to open and close doors, closed-circuit monitors to deliver
educational and other programming, surveillance cameras that constantly
watch inmates, and computer monitors for conducting “visits.” In these
units, staff must follow prescribed protocols consisting of detailed, exacting
procedures. Wall (2016) describes, for example, how canisters of pepper spray
are weighed in Rhode Island segregation units to monitor whether an excessive
amount of the chemical was discharged when used. The California Department
of Corrections and Rehabilitation once required that guards use, as a matter
of policy, a stun gun every time they forcibly removed an inmate from a cell,
regardless of the inmate’s behavior (Simon, 2014). These directives describe a
culture that fixates on technology and formal rules. Rhodes (2007) explains,
“supermax technology offers the cultural gloss of a ‘high-tech solution’ that
helps frame problems — some of them caused by supermax confinement in the
first place — largely in terms of their susceptibility to technical intervention”

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 311

(p. 549). According to Rhodes, the result is an impression of inevitability and
necessity — a sense of progress and sophistication.
The culture of administrative segregation is also embodied in beliefs about
inmates and correctional staff and the way that these belief systems play out
through interactions. Rhodes (2004) shows that administrative segregation is
based on the concept of the hyperrational, irredeemably dangerous inmate.
Prisoners held in administrative segregation are called the worst of the worst,
conjuring up archetypal images of convict revolutionaries and calculating serial
killers (for a discussion, see Reiter, 2012). Rhodes (2004) illustrates how inmate
behavior is interpreted within this belief system. Acts of serious misconduct and
self-harm are interpreted as manipulation and malingering and are viewed as
confirmation that the inmates require isolation and punishment (also see Haney,
2008). Yet, prisoners’ records of good behavior are interpreted as evidence of
their superhuman will and calculated rationality. Drawing on Schein’s (1990)
work, these organizational beliefs can be perceived as strategies for helping staff
to understand the problem of the dangerous inmate and cope with the moral
complexity of extreme deprivation and punishment.
Toughness and hypermasculinity are ascribed to correctional officers who
work in supermax and segregation facilities (Haney, 2008). Supermax prisons,
for example, are said to employ the best of the best to control the worst
of the worst. These attributes are most clearly expressed via uniforms and
security rituals. Haney (2008) describes the paramilitary insignia exhibited
on uniforms, officers donning body armor and gas masks, and guards’ use
of overpowering physical force as examples (also see Rhodes, 2007). Haney
also discusses the cultural value of seeing action and overpowering resistant
inmates, experiences that build guards’ status within the organization.
Irwin (2004) hypothesizes that this type of “us versus them” culture allows
correctional officers to maintain a sense of morality while participating in
practices that may degrade and dehumanize prisoners.

Occupational Roles and Job Performance
During the 1970s and early 1980s, prison reformers advocated for diversifying
and professionalizing prison staff. Recruiting more women, minorities, and
college-educated correctional officers, reformers argued, would enable change
in negative guard cultures characterized by racism, excessive use of force,
and resistance to civil rights reforms. Research, however, typically found no
relationship — or weak and inconsistent relationships — between correctional
officers’ personal backgrounds and work-related attitudes. Reflecting on these
findings, one research team hypothesized that the demands of correctional work
may be “so encompassing and yet so restrictive that all officers, regardless of
gender, race, social backgrounds, and prior beliefs, will develop similar attitudes
toward their jobs” (Jurik, Halemba, Musheno, & Boyle, 1987, p. 109).

312 • National Institute of Justice | NIJ.gov

Research on workplace socialization in prisons suggests that the organizational
cultures of administrative segregation units could affect a variety of workrelated behaviors, attitudes, and emotions. For example, organizational cultures
create expectations among staff about their responsibilities regarding problem
management. In a culture of harm, Haney (2008) argues,
Interventions aimed at de-escalation or compromise may be seen as capitulation,
signs of weakness, or ‘rewarding bad behavior.’ Guards who violate the norms
of punishment by routinely seeking compromise, finding ways to express
encouragement, or showing empathy for the prisoners’ plight face marginalization,
ostracism, and reassignment (p. 972).
Thus, culturally proscribed attitudes and organizational expectations are
likely to shape how employees in segregation units perform their jobs. The
organizational culture may also shape inmates’ expectations and their reactions
to correctional officers.
King (2005) interviewed supermax inmates in Colorado and Minnesota
about their views on staff. In both locations, inmates said the officers were
professional, did not abuse inmates, and could articulate a clear institutional
mission. The officers at Minnesota’s Oak Park Heights facility, however, were
viewed as more helpful and fair, and less racist than the officers at the Colorado
State Penitentiary. King (2005) noted that Oak Park Heights offered more
programming opportunities and that inmates spent more time out of their
cells, which increased opportunities for inmates and staff to communicate and
interact informally. In contrast, inmates and officers at the Colorado facility
rarely interacted, because inmates were locked down and behind barriers. These
observations led King to hypothesize that “in situations where potential contact
between staff and prisoners in public settings is higher, staff will experience a
greater strain towards behaving in helpful, fair, consistent, and nonracist ways”
(2005, p. 135).
Furthermore, King (2005) asked whether the use of administrative segregation
might result in the loss of important skills. For example, the communication, de­
escalation, and conflict-resolution skills of officers working in segregated housing
units may atrophy from lack of use. King also questioned whether these skills
may weaken in all officers who have a greater incentive to move difficult and
troublesome inmates to administrative segregation rather than engaging with
them to solve problems. Reliance on administrative segregation may, therefore,
alter patterns of interaction between officers and inmates throughout a prison
system. Briggs and colleagues (2003) noted, for instance, that they could not
determine whether the lower rates of assault against staff members in Illinois
after the supermax facility opened were due to changes in inmate behavior or
staff behavior, or to a change in interactions between the two groups.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 313

Civilian employees are not immune to the effects of working in administrative
segregation. Doctors, nurses, psychiatrists, and chaplains may experience
heightened conflict and ambiguity regarding their roles in segregation facilities,
where the organizational culture and regime are more starkly at odds with
the orientation and expectations of their professions. “It is hard to imagine a
clinician anywhere else in society even attempting a therapeutic interaction
with a patient who is standing or sitting inside a thick metal cage,” notes Haney
(2008, p. 973). Although he argues that staff cannot alter these security demands,
Rhodes (2004) maintains that the struggle between treatment and custody ideals
is an important check on the hegemony of the administrative segregation regime.
Despite their potential to affect an organization, it is likely that treatment staff
struggle to maintain standards of professional care and clinical empathy when
working in administrative segregation, where security dominates every decision.
Finally, Haney (2008) argues that the culture of harm is particularly vulnerable to
the escalation of punitive practices.
Because guards are encouraged to punish, repress, and forcefully oppose —
by virtue of the fact that they are provided with no alternative strategies for
managing prisoners — they have no choice but to escalate the punishment when
their treatment of prisoners fails to produce the desired results (as it frequently
does). Of course, over time, the correctional staff becomes accustomed to
inflicting a certain level of pain and degradation — it is the essence of the regime
that they control and whose mandates they implement. They naturally become
desensitized to these actions and, in the absence of any alternative approaches
(both the lack of conceptual alternatives or the means to implement them), they
deliver more of the same (p. 970).
U.S. Supreme Court Justice John Paul Stevens, dissenting in Beard v. Banks,
recognized this potential, arguing that the desire to “motivate good behavior” via
deprivation has no principle of limitation (cited in Cohen, 2008). It is reasonable
to expect, then, that management philosophies and practices that place a heavy
emphasis on security and coercion to achieve prison order are vulnerable to
excessive use of force and abuse (Haney, 2008; Useem & Piehl, 2006). Research
examining the use of force, and attitudes toward the use of force, is needed to
learn more about this phenomenon and its relationship to formal and informal
organizational values.

The Emotional Consequences of Working in Administrative Segregation
Haney (2008) describes administrative segregation as operating in an “ecology
of cruelty” that affects all who work and live there. The stark environment
of supermax prisons, in particular, exposes people to stress. The bunker-like
atmosphere, constant vigilance, and wild swings between extreme boredom
and extreme crisis may take a toll on employees’ health. “[P]risoners and

314 • National Institute of Justice | NIJ.gov

guards — are likely to find the outer limits of their psychological tolerance
pressed by these places,” Haney observes (2008, p. 969). Recognizing this
potential, correctional agencies attempt to rotate officers’ assignments within
segregation units, sometimes even rotating employees to other facilities
(Riveland, 1999b). However, their efforts are hampered at supermax prisons,
which are often located in rural locations. Prisons also struggle to manage
high staff turnover and absenteeism, which make shift rotation difficult and
unpredictable. It is important that future research considers how the duration of
an officer’s tenure in segregation may affect work-related outcomes and officer
well-being. Similarly, it is unclear whether shift and institutional rotation are
effective strategies for mitigating the effects of working in segregation.
Researchers and corrections leaders know a great deal about the stresses of
prison work. A well-established body of research demonstrates both the high
level of occupational stress experienced by prison employees and the harmful
effects of chronic occupational stress (Cheek & Miller, 1983; Dowden & Tellier,
2004). For example, research has found particularly high rates of divorce, heart
disease, absenteeism, turnover, and burnout among correctional employees
(Dowden & Tellier, 2004). In interviews with staff working in control units, King
(2005) observed that “[v]irtually without exception staff dealing with the ‘worst
of the worst’ prisoners in England and Wales found their job stressful” (p. 135).
Those who work in administrative segregation may be regularly exposed to
traumatic events, including suicides and disturbing acts of self-mutilation
(Edge & Jones, 2014). Rhodes (2004) described the deep visceral disgust and
anger that staff experience when prisoners engage in dirty protests — throwing
feces and other body fluids on officers or covering themselves and their cells in
feces to force officers to conduct cell extractions. Extended exposure to trauma
and feelings of disgust may contribute to professional detachment and loss of
compassion, causing employees to become numb to emotions or to act out in
anger and frustration.
Vigilance and fear of victimization — feeling constantly on guard for signs
of danger — may also be especially high among employees who work in
administrative segregation. Inmates’ rage and desperation are frequently directed
at the correctional officers who work in segregation units and at the prison
administrators who place inmates in segregation (Rhodes, 2004; Gawande, 2009).
As discussed above, physical assaults against correctional staff occur at a high
rate within segregation units. An exploratory analysis by Sundt and colleagues
(2004) found that Illinois correctional officers at the supermax prison reported
substantially higher levels of work stress and fear of victimization than did
guards working in the general prison population. Supermax officers also were
less committed to and less satisfied with their jobs than were officers working in
the states’ other four maximum-security prisons. Although a low response rate
from the officers working at the supermax prison precludes generalization, the
lack of participation is, perhaps, another indicator of the officers’ disengagement
and alienation.
Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 315

Cultural Change and Reforming Administrative Segregation
In the mid-2000s, the National Institute of Corrections (NIC) developed the
Cultural Change Initiative (CCI) to support corrections officials in their efforts
to change organizational culture. The initiative was designed to address negative
prison culture, which NIC opinion leaders believed contributed to breakdowns in
leadership and problems among staff and inmates such as inadequate responses
to sexual assaults, racist attitudes, and excessive use of force. The initiative
provided training and technical assistance with organizational culture, change
management, and strategic planning (Byrne, Hummer, & Taxman 2008).
Participants in CCI attended a three-day training, “Promoting Positive Prison
Culture,” where they learned about organizational culture, assessed the culture
in their organizations, and began developing workplace improvement plans.
Participating organizations then received technical assistance in writing
mission statements for their facilities, identifying desired values and beliefs,
and developing plans for achieving and monitoring desired outcomes. In a
multisite evaluation of CCI, Byrne and colleagues (2008) found that, despite
implementation problems, participating organizations generally showed
improvement in organizational culture indicators. At one study site that provided
outcome data, CCI was associated with declines in overall inmate misconduct
over a 17-month period. Violent incidents also declined, but the change was
not statistically significant. Although preliminary, the results found by the
Byrne team are promising and support the need for additional research on
organizational culture in prisons and for the development of more robust cultural
change interventions.
The findings are also consistent with theoretical work that points to the
important link between culture and organizational effectiveness. As doubts are
raised about the appropriate use of administrative segregation and its potentially
harmful effects, it will be important to consider how the organizational culture
of prisons, generally — and administrative segregation units, specifically — may
both impede and facilitate change. For example, Mohr and Raemisch (2015)
suggest that too strong a focus on security or efficiency may detract from other
important organizational goals and missions such as improving public safety.
For decades, administrative segregation has been justified as necessary for the
protection of staff members. Moving away from reliance on administrative
segregation and reforming its mission from containment to risk reduction, for
example, will require significant changes in staff culture, belief systems about
inmates, and assumptions about how best to create orderly and safe prisons.
Discussions about the need to reform administrative segregation should
specifically acknowledge the importance of protecting correctional officers from
the negative effects of high stress and continued exposure to high rates of assault
within segregated housing. Focusing exclusively on the well-being of inmates and
abuses documented in administrative segregation could contribute to a hostile
“us versus them” culture and resistance to needed reforms.
316 • National Institute of Justice | NIJ.gov

A. T. Wall (2016), the director of the Rhode Island Department of Corrections,
advocates for the full participation of correctional staff in efforts to reform
the use of administrative segregation. “The success of any such venture will
depend on our ability to win and maintain the trust of corrections personnel,”
he explains. Rather than the paramilitary, top-down organizational model
of management traditionally employed in prisons, Wall argues for a more
horizontal, collaborative approach to reform, wherein employees from all levels
of the organization work to identify challenges and agree on changes. Such a
change in the organizational culture of prisons could have far-reaching benefits.

Toward an Evidence-Based Model of Prison Management
The evidence-based practices movement is transforming the delivery of
correctional programs (Taxman & Belenko, 2012). And for good reason:
theoretically informed and scientifically validated practices deliver better
outcomes with a higher return on public investment (Drake, Aos, & Miller,
2009). Institutional corrections, however, has been slower to adopt an evidencebased orientation and there is little empirical research on the nation’s prison
systems. It is noteworthy that the Department of Justice’s recent Smart on Crime
Initiative included technical assistance grants to support “smart policing,” “smart
prosecution,” “smart defense,” “smart probation,” and “smart supervision” (to
reduce the use of prisons). There was no “smart prison” initiative — there simply
is no evidence base on which to build such a program.
It is worth considering why so little scholarship exists on such an important
public institution. Although this question is worthy of careful evaluation, a few
preliminary observations are offered here with the hope of identifying barriers
to success and opportunities for improvement. Four issues seem particularly
germane: public leadership, data infrastructure, institutional review boards, and
the need for national reporting standards for corrections.
Federal and state governments spent billions of dollars to build supermax
prisons and retrofit other facilities for administrative segregation without a single
independent study documenting either the need for or the utility of this practice
(Mears, 2008a). The use of evidence-based practices requires organizations
to embrace the value of scientific knowledge and incorporate data into their
decision-making. Although academic scholars and professional associations have
for decades called for greater transparency, use of research, and performance
measures, prisons remain a “black box” (Mears, 2008b). Strong leadership at
every level of government is needed to address the problem. Embracing the value
of data-driven policies will facilitate positive organizational change and promote
positive organizational cultures. Promising examples of success in community
corrections exist, where leaders in states such as Oklahoma and Oregon took
assertive roles in promoting greater public accountability and organizational
effectiveness (Latessa, 2004). The political context surrounding incarceration

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 317

is shifting. This opportunity should be embraced, and public leaders should
be urged to support research on prisons that will lead to improved quality,
efficiency, and public safety.
Given the lack of organizational support for research, it is not surprising that the
data infrastructures of departments of corrections are grossly underdeveloped.
As recently as 2012, the Government Accountability Office (2014) found that
the Federal Bureau of Prisons did not document how long inmates spent in
administrative segregation and had never evaluated the effect of long-term
segregation on prison safety or inmate well-being. Although some fault lies with
the hubris of decision-makers who have failed to subject their policies to study
or monitoring, there is also a lack of infrastructure to support data collection and
reporting. There is a tremendous need for technical assistance and funding to
support the creation of robust, modern information technology systems that can
support better decision-making to improve outcomes.
Prisoners are a protected class of research subjects, and scholars who conduct
research on prisoners understandably have the extra burden of demonstrating
that their research complies with all ethical and regulatory requirements
governing research on human subjects. University review boards are often
ill-informed about prisons and criminal justice and wary of exposing their
institutions to any risk. In addition, departments of corrections may require
researchers to submit proposals for research that, if approved, must go through
another internal review board constituted by the prison system. Review boards
within departments of correction also have incentives for denying research
proposals that may expose their organizations to risk by revealing problems
within the prison system or by questioning policy decisions. Moreover, the
members of review boards within departments of corrections typically lack the
scientific credentials and the expertise to make appropriate methodological or
theoretical recommendations.
Current regulatory requirements may discourage scientific social research
on prisons and prisoners. At the least, scholars who engage in research on
prisons and prisoners need additional support to navigate the difficult and
time-consuming institutional review board process. Timelines for conducting
funded research on prisons and inmates, for example, should be more generous
than those set for other types of research. There is also an opportunity for the
Department of Justice to conduct a national assessment of institutional review
boards and recommend best practices to better balance the protection of
prisoners from unethical research practices and the pressing need to develop
a scientific knowledge base about prisons and incarceration. States should also
consider whether departments of corrections have a conflict of interest that
prevents them from objectively reviewing the merits of proposed research.
Finally, the lack of national reporting standards makes it difficult to generalize
from research findings and accumulate knowledge. It is hard to imagine where

318 • National Institute of Justice | NIJ.gov

criminal justice practice and scholarship would be if the Federal Bureau of
Investigation had not created standards to measure crime in the 1920s. Yet, the
corrections field has no standards for reporting performance indicators such as
assaults and recidivism. Similarly, there are no agreed-upon definitions for basic
organizational indicators such as security level or operating capacity. The Prison
Rape Elimination Act of 2003 is an exception and an example of progress. It
created national standards for reporting rape and sexual victimization in prisons.
The corrections field may also need national legislation to establish additional
reporting standards.

Directions for Future Research
Given the state of the knowledge base about prison management, administrative
segregation, organizational culture, and prison order in general, a broad set
of research questions and methodologies is needed. Although experimental
research produces internally valid results, those results are often not suitable
for generalization. Experimental research is not well-suited to developing
new insights and is a poor strategy for describing or understanding complex,
multivariate phenomena and contextual effects. Prison order is a complex,
dynamic phenomenon intimately tied to context; thus, it is premature to establish
a particular methodology or set of research questions. That said, some general
recommendations are offered below.

Basic Research Needs
First, basic research on inmates’ adjustment to incarceration is needed. Too little
is known about the common patterns and causes of inmate behavior. Until more
information about these patterns is gathered and assessed, developing effective
interventions and programs for managing violent and disruptive inmates is
guesswork. Absent a solid understanding of the cause of a problem, there is
a tendency to focus too much on managing symptoms. The corrections field
lacks the research on the etiology of prison violence, victimization, non-serious
misconduct, suicide, self-harm, fear, gang activity, and mental health necessary
to develop a solid basis for identifying effective strategies to manage and prevent
serious and disruptive inmate misconduct. Basic research is the foundation on
which good theory and good practice are built.
In a similar vein, the field lacks research on the trajectories of adjustment and
maladjustment over the course of prisoners’ sentences to better understand
stability and change in inmate behavior. Studies of inmate behavior patterns
before, during, and after placement in administrative segregation — and of
other efforts to manage violent, disturbed, or disruptive inmates — are also
needed to understand more fully the effect of interventions on prison order and
inmate well-being (Toch & Kupers, 2007). Longitudinal research designs may be

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 319

particularly valuable for understanding the dynamic interplay among inmates,
staff, the prison environment, and prison management. Until a better research
base is developed, researchers and practitioners alike will continue to puzzle over
the inconsistent and unpredictable results of efforts to manage and reduce prison
violence and disorder.

Organizational Variation in Prison Order
Second, the field needs research on the causes of variation in institutional order
from one prison system, prison, or cellblock to the next. Too little is known
about the factors that distinguish prisons with high rates of disorder from
those with low rates (Useem & Piehl, 2006). In particular, it is unclear whether
prison security levels prevent disorder or create contexts that exacerbate and
concentrate disorder and future offending (Gaes & Camp, 2009; Wooldredge
& Steiner, 2015). Yet, under the assumption that this fundamental practice is
critical to creating safer prisons, security classifications are used in every prison
system. Studies evaluating the relative effectiveness of the “concentration model”
versus the “dispersion model” are needed, as are studies that examine variation in
management strategies and organizational cultures. It is particularly important
that research in this area develop better measures of key concepts and move away
from general indicators of policies and program participation.

Deterrence, Justice, and Legitimacy
Third, the effects of official responses to inmate grievances and formal sanctions
for misconduct must be better understood. Strong theory and an emerging
body of research point to the important interplay among the use of authority,
perceptions of fairness and legitimacy, and compliance (Liebling, 2004; Sherman,
1993; Sparks & Bottoms, 1995; Tyler, 2003). Research is needed in three areas:
deterrence, restorative justice, and procedural justice. Studies of the use of
sanctions and deterrent strategies commonly yield inconsistent results (Sherman,
1993), such as those seen so far on the use of segregation. Perceptions of fairness,
justice, and legitimacy appear to influence the effect of sanctions and explain
some of the inconsistent findings in the literature. The corrections field should
build on the works of Colvin (2007), Liebling (2004), and Sparks and Bottoms
(1995) to develop an applied knowledge base of the effective exercise of authority
in prisons. This may be a particularly important area of inquiry, because coercive
practices can backfire.
The rise of supermax prisons and the expanded use of administrative segregation
emerged, in part, because corrections officials believed that the court reforms
of the late 1970s and 1980s severely limited their authority and their options for
responding effectively to inmate violence and misconduct. As the field moves
away from using segregation to punish and control violent and disruptive
inmates, it is critical that it develop new options for preventing serious

320 • National Institute of Justice | NIJ.gov

misconduct and responding effectively when it does occur. Corrections officials
need help from researchers and the National Institute of Justice to identify
effective alternatives to administrative and punitive segregation. Swift and certain
deterrence, restorative justice programs, and supervision informed by procedural
justice are promising areas for beginning this research.

Studies of Administrative Segregation
Fourth, more research is needed to understand the effect of administrative
segregation on prison order and safety and to answer questions about when
administrative segregation should be used, how many inmates may require
it, and for how long. Comparative case studies, process evaluations, and
outcome evaluations can reveal more about administrative segregation’s effect
on systemwide prison order and the mechanisms that connect this practice to
various outcomes. Priority should be given to studies that conceptually and
methodologically distinguish between the effects of separating dangerous or
disruptive inmates from the general population and the effects of conditions
of confinement. Studies that systematically examine the various policies and
practices found in administrative segregation will also be revealing.
The field also needs better data about the effect of working in administrative
segregation. Developing and supporting the prison workforce is a longstanding
priority of the Office of Justice Programs as well as the various professional
associations that represent corrections officers. The handful of observations
about how working in supermax prisons affects employees raises concerns about
the heavy burden and risks borne by staff in these facilities. More information
is needed about all that staff experience when they work in administrative
segregation. Then, the field can better consider how to mitigate the potential
effects of chronic occupational stress, fear, and workplace victimization. It is
also important to understand more about how the culture of supermax prisons
affects the use of force, hostility between inmates and staff, and the escalation of
punishment and resistance.

Effective Practices of Inmate Supervision
Fifth, too little is known about how prison employees contribute to inmate
adjustment, prison order, and rehabilitation. There are hints that the quality of
interactions between correctional officers and inmates is consequential (Day,
Brauer, & Butler, 2015; King, 2005; Liebling, 2004). For example, recent research
on the effectiveness of probation and parole officers indicates that both officers’
professional skills and their relationships with the offenders they supervise are
important (Bonta, Bourgon, Rugge, Gress, & Gutierrez, 2013; Kennealy, Skeem,
Manchak, & Eno Louden, 2012; Smith, Schweitzer, Labrecque, & Latessa, 2012).
Research is needed to help the field understand more about interactions between
inmates and officers and how these interactions contribute to prison order, safety,

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 321

fear, perceptions of fairness, and inmate behavioral change. The corrections field
must also learn more about effective (and ineffective) practices for working with
dangerous and disruptive inmates.

Organizational Culture and Effectiveness
Sixth, and finally, additional research is needed to improve the field’s
understanding of the relationship between organizational culture and
organizational effectiveness. The National Institute of Justice’s Cultural Change
Initiative is a promising strategy for developing positive organizational cultures
and improving outcomes such as prison order and safety. Evaluations of
efforts to reform administrative segregation should incorporate assessments of
organizational culture, which may be an important determinant of the Cultural
Change Initiative’s success. More broadly, additional case studies, histories,
and ethnographies are needed to better understand prison cultures and what
distinguishes harmful, destructive prison cultures from just, reintegrative prison
environments. This research will help correctional leaders to think strategically
about the relationships among their management practices, culture, and the
mission of their institutions.

The Well-Managed Prison
The appropriate role of administrative segregation in maintaining an orderly
and safe prison system is not yet known. Certainly, there is very little scientific
evidence supporting its effectiveness and enough contrary evidence to warrant
limiting its use. Cohen (2008) argues that the field could concede to the need to
separate dangerous and disruptive inmates from the general prison population
without also agreeing that extreme social isolation and the harsh conditions
of confinement are legitimate strategies for securing prison safety and order.
Reforms to the regime of administrative segregation are clearly needed, but more
fundamental doubts remain about the capacity of architectures and technologies
of control to impose the habits of order.
In a footnote to his article on trends in prison management, Riveland (1999a)
commented that prisons are one of the few social institutions that measure
themselves by their failures. Departments of corrections report rates of
recidivism, and, occasionally, rates of prison violence, suicide, and major
disruption. These are critically important elements, but they reveal nothing of
the other important organizational goals of prisons. They do not help corrections
officials to understand markers of quality such as inmate productivity, reduction
in criminal risk, increased fairness, and the improved well-being of staff and
inmates. Performance measures are not merely indicators of success or failure;
they also powerfully communicate organizational values and priorities — What
gets counted tends to count.

322 • National Institute of Justice | NIJ.gov

Alison Liebling (2004) wraps up her analysis of prison performance by
identifying what matters in prisons: the quality of the relationships among
inmates, staff, and institutional leaders. Liebling concludes that firm, fair, and
caring relationships are the foundation of moral correctional practices.3 Good
organizational performance, argues Liebling, is characterized by a value balance
between security and harmony that is rooted firmly in the concept of a just
community. Respect, humanity, good and right staff-prisoner relationships,
fairness, effective security and management systems, and strong leadership are
the markers of a well-managed prison.
Improving knowledge about inmate adjustment, prison organizations, and
effective management strategies will help correctional officials and scholars to
develop approaches to achieving prison order and developing a set of effective
practices for inmate supervision. In addition, correctional leaders need data to
inform and support their decisions about the appropriate use of administrative
segregation. The United States spent billions of dollars building the capacity to
incarcerate a large number of inmates in restrictive housing without conducting
a careful study of either the need for or efficacy of segregation (Mears, 2008a).
The corrections field should not compound this mistake by investing in other
unproven strategies. It is time to invest in developing an evidence base to guide
prison management in the use of strategies such as administrative segregation.

References
Andrews, D. A., & Bonta, J. (1994). The psychology of criminal conduct.
Cincinnati, OH: Anderson.
Barak-Glantz, I. L. (1983). Who’s in the hole? Criminal Justice Review, 8(1), 29-37.
Beijersbergen, K. A., Dirkzwager, A. J., Eichelsheim, V. I., Van der Laan, P. H., &
Nieuwbeerta, P. (2015). Procedural justice, anger, and prisoners’ misconduct: A
longitudinal study. Criminal Justice and Behavior, 42(2), 196-218.
Bennett, L. (1976). The study of violence in California prisons: A review with
policy implications. In A. K. Cohen, G. F. Cole, and R. G. Bailey (Eds.), Prison
violence: A sociological perspective (pp. 149-168). Lexington, MA: D.C. Heath.
Bidna, H. (1975). Effects of increased security on prison violence. Journal of
Criminal Justice, 3(1), 33-45.

3

Firm and fair officer-offender relationships may also be the foundation of effective correctional practices
(Andrews & Bonta, 1994; Skeem, Louden, Polaschek, & Camp, 2007).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 323

Bonta, J., Bourgon, G., Rugge, T., Gress, C., & Gutierrez, L. (2013). Taking the
leap: From pilot project to wide-scale implementation of the Strategic Training
Initiative in Community Supervision (STICS). Justice Research and Policy, 15(1), 17-35.
Bottoms, A. E. (1999). Interpersonal violence and social order in prisons. Crime
and Justice, 26, 205-281.
Briggs, C. S., Sundt, J. L., & Castellano, T. C. (2003). The effect of supermaximum
security prisons on aggregate levels of institutional violence. Criminology, 41(4),
1341-1376.
Byrne, J. M., Hummer, D., & Taxman, F. S. (2008). The National Institute of
Corrections’ institutional cultural change initiative: A multisite evaluation. In J.
M. Byrne, D. Hummer, & F. S. Taxman (Eds.), The culture of prison violence
(pp. 137-163). Boston, MA: Pearson.
Carrabine, E. (2005). Prison riots, social order and the problem of legitimacy.
British Journal of Criminology, 45(6), 896-913.
California Department of Corrections. (1986). Task force on violence,
special housing, and gang management. Sacramento, CA: State of California,
Department of Corrections.
Cheek, F. E., & Miller, M. D. S. (1983). The experience of stress for correction
officers: A double-bind theory of correctional stress. Journal of Criminal Justice,
11, 105-120.
Clemmer, D. (1940). The prison community. Boston, MA: Christopher Publishing House.
Cohen, F. (2008). Penal isolation beyond the seriously mentally ill. Criminal
Justice and Behavior, 35(8), 1017-1047.
Colvin, M. (2007). Applying differential coercion and social support theory to
prison organizations: The case of the Penitentiary of New Mexico. The Prison
Journal, 87(3), 367-387.
Colvin, M., Cullen, F. T., & Vander Ven, T. (2002). Coercion, social support, and
crime: An emerging theoretical consensus. Criminology, 40(1), 19-42.
Crouch, B. M., Marquart, J. R., & Marquart, J. W. (2010). An appeal to justice:
Litigated reform of Texas prisons. Austin, TX: University of Texas Press.
Day, J. C., Brauer, J. R., & Butler, H. D. (2015). Coercion and social support
behind bars: Testing an integrated theory of misconduct and resistance in U.S.
prisons. Criminal Justice and Behavior, 42(2), 133-155.

324 • National Institute of Justice | NIJ.gov

De Beaumont, G., & De Tocqueville, A. (1833). On the penitentiary system in the
United States: And its application in France; with an appendix on penal colonies,
and also, statistical notes. Philadelphia, PA: Carey, Lea & Blanchard.
Dowden, C., & Tellier, C. (2004). Predicting work-related stress in correctional
officers: A meta-analysis. Journal of Criminal Justice, 32(1), 31-47.
Drake, E. K., Aos, S., & Miller, M. G. (2009). Evidence-based public policy
options to reduce crime and criminal justice costs: Implications in Washington
state. Victims and Offenders, 4(2), 170-196.
Edge, D., & Jones, E. C. (2014). Solitary nation. Frontline [Television series].
Boston, MA: WBGH Educational Foundation.
Feeley, M. M., & Simon, J. (1992). The new penology: Notes on the emerging
strategy of corrections and its implications. Criminology, 30(4), 449-474.
Flanagan, T. J. (1983). Correlates of institutional misconduct among state
prisoners. Criminology, 21(1), 29-40.
Gaes, G. G., & Camp, S. D. (2009). Unintended consequences: Experimental
evidence for the criminogenic effect of prison security level placement on postrelease recidivism. Journal of Experimental Criminology, 5(2), 139-162.
Garland, D. (1990). Punishment and modern society: A study in social theory.
Chicago, IL: University of Chicago Press.
Gawande, A. (2009, March 30). Hellhole. The United States holds tens of
thousands of inmates in long-term solitary confinement. Is this torture? The New
Yorker. Retrieved from http://www.newyorker.com.
Hallinan, J. (1995, December 10). Supermax prisons: They’re supertough,
superexpensive and supergood for poor communities. The Sunday Patriot, p. G1.
Haney, C. (2008). A culture of harm: Taming the dynamics of cruelty in
supermax prisons. Criminal Justice and Behavior, 35(8), 956-984.
Haney, C., Banks, C., & Zimbardo, P. (1973). Interpersonal dynamics in a
simulated prison. International Journal of Criminology and Penology, 1, 69-97.
Huebner, B. M. (2003). Administrative determinants of inmate violence:
A multilevel analysis. Journal of Criminal Justice, 31(2), 107-117.
Human Rights Watch. (1997). Cold storage: Supermaximum security confinement
in Indiana. New York, NY: Human Rights Watch.
Irwin, J. (2004). The warehouse prison: Disposal of the new dangerous class. New
York, NY: Oxford University Press.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 325

Irwin, J., & Austin, J. (1997). It’s about time: America’s imprisonment binge (2nd
ed.). Belmont, CA: Wadsworth.
Jacobs, J. B., & Retsky, H. G. (1975). Prison guard. Journal of Contemporary
Ethnography, 4, 5-29.
Jurik, N. C., Halemba, G. J., Musheno, M. C., & Boyle, B. V. (1987). Educational
attainment, job satisfaction, and the professionalization of correctional officers.
Work and Occupations, 14(1), 106-125.
Kennealy, P. J., Skeem, J. L., Manchak, S. M., & Eno Louden, J. (2012). Firm, fair,
and caring officer-offender relationships protect against supervision failure. Law
and Human Behavior, 36(6), 496.
King, K., Steiner, B., & Breach, S. R. (2008). Violence in the supermax: A selffulfilling prophecy. The Prison Journal, 88(1), 144-168.
King, R. D. (1999). The rise and rise of supermax: An American solution in
search of a problem? Punishment & Society, 1(2), 163-186.
King, R. D. (2005). The effects of supermax custody. In A. Liebling & S. Maruna
(Eds.), The effects of imprisonment (pp. 118-145). Portland, OR: Willian Publishing.
Kupers, T. A., Dronet, T., Winter, M., Austin, J., Kelly, L., Cartier, W., Morris,
T.J., ... Vincent, L. C. (2009). Beyond supermax administrative segregation:
Mississippi’s experience rethinking prison classification and creating alternative
mental health programs. Criminal Justice and Behavior, 36(10), 1037-1050.
Kurki, L., & Morris, N. (2001). The purposes, practices, and problems of
supermax prisons. Crime and Justice, 28, 385-424.
Labrecque, R. M. (2015). The effect of solitary confinement on institutional
misconduct: A longitudinal evaluation Doctoral dissertation, University of
Cincinnati). Cincinnati. Retrieved from https://etd.ohiolink.edu/!etd.send_
file?accession=ucin1439308329&disposition=inline.
Latessa, E. J. (2004). The challenge of change: Correctional programs and
evidence-based practices. Criminology & Public Policy, 3(4), 547-560.
Liebling, A. (2004). Prisons and their moral performance: A study of values,
quality, and prison life. New York, NY: Oxford University Press.
Liebling, A. (2011). Moral performance, inhuman and degrading treatment and
prison pain. Punishment & Society, 13(5), 530-550.
Liman Program & Association of State Correctional Administrators. (2015).
Time-in-cell: ASCA-Liman 2014 National Survey of Administrative Segregation in
Prison. New Haven, CT: Yale Law School.

326 • National Institute of Justice | NIJ.gov

Lombardo, L. X. (1981). Guards imprisoned: Correctional officers at work. New
York, NY: Elsevier.
Mears, D. P. (2008a). An assessment of supermax prisons using an evaluation
research framework. The Prison Journal, 88(1), 43-68.
Mears, D. P. (2008b). Accountability, efficiency, and effectiveness in corrections:
Shining a light on the black box of prison systems. Criminology & Public Policy,
7(1), 143-152.
Mears, D. P., & Reisig, M. D. (2006). The theory and practice of supermax
prisons. Punishment & Society, 8(1), 33-57.
Mears, D. P., & Watson, J. (2006). Towards a fair and balanced assessment of
supermax prisons. Justice Quarterly, 23(2), 232-270.
Metcalf, H., Morgan, J., Oliker-Friedland, S., Resnik, J., Spiegel, J., Tae, H.,
... Holbrook, B. (2013). Administrative segregation, degrees of isolation, and
incarceration: A national overview of state and federal correctional policies.
Hartford, CT: Liman Public Interest Program, Yale Law School.
Mohr, G. C. & Raemisch, R. (2015, February 13). Restrictive housing: Taking the
lead. Corrections Today. Retrieved from www.aca.org.
Morris, R. G. (2015). Exploring the effect of exposure to short-term solitary
confinement among violent prison inmates. Journal of Quantitative Criminology,
32(1), 1-22.
Porporino, F. J. (1986). Managing violent individuals in correctional settings.
Journal of Interpersonal Violence, 1(2), 213-237.
Pratt, T. C., Cullen, F. T., Blevins, K. R., Daigle, L. E., & Madensen, T. D. (2006).
The empirical status of deterrence theory: A meta-analysis. In F. T. Cullen, J. P.
Wright, & K. R. Blevins (Eds.), Taking stock: The status of criminological theory,
Vol. 15, pp. 367-396.
Prendergast, A. (2015, Dec. 7). Colorado supermax will build rec yards to settle
prisoners’ lawsuit. Westword. Retrieved from www.westword.com.
Ralph, P. H., & Marquart, J. W. (1991). Gang violence in Texas prisons.
The Prison Journal, 71(2), 38-49.
Reiter, K. A. (2012). The most restrictive alternative: The origins, functions, control,
and ethical implications of the supermax prison, 1976-2010. Doctoral dissertation,
University of California, Berkeley. Retrieved from http://digitalassets.lib.berkeley.
edu/etd/ucb/text/Reiter_berkeley_0028E_12273.pdf.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 327

Reiter, K. (2014). The Pelican Bay hunger strike: Resistance within the structural
constraints of a U.S. supermax prison. South Atlantic Quarterly, 113(3), 579-611.
Rhodes, L. A. (2004). Total confinement: Madness and reason in the maximum
security prison (Vol. 7). Berkeley and Los Angeles, CA: University of California Press.
Rhodes, L. A. (2007). Supermax as a technology of punishment. Social Research,
74(2), 547-566.
Riveland, C. (1999a). Prison management trends, 1975-2025. Crime and Justice,
26, 163-203.
Riveland, C. (1999b). Supermax prisons: Overview and general considerations.
Washington, DC: National Institute of Corrections.
Rothman, D. J. (1971). The discovery of the asylum. Piscataway, NJ: Transaction Publishers.
Rubin, A. T. (2015). Resistance or friction: Understanding the significance of
prisoners’ secondary adjustments. Theoretical Criminology, 19(1), 23-42.
Sherman, L. W. (1993). Defiance, deterrence, and irrelevance: A theory of the
criminal sanction. Journal of Research in Crime and Delinquency, 30(4), 445-473.
Simon, J. (2014). Mass incarceration on trial: The remarkable court decision and
the future of prisons in America. New York, NY: The New York Press.
Schein, E. H. (1990). Organizational culture. American Psychologist, 45(2), 109-119.
Skeem, J. L., Louden, J. E., Polaschek, D., & Camp, J. (2007). Assessing
relationship quality in mandated community treatment: Blending care with
control. Psychological Assessment, 19(4), 397-410.
Smith, P., Schweitzer, M., Labrecque, R. M., & Latessa, E. J. (2012). Improving
probation officers’ supervision skills: An evaluation of the EPICS model. Journal
of Crime and Justice, 35(2), 189-199.
Sorensen, J. R., Cunningham, M. D., Vigen, M. P., & Woods, S. O. (2011). Serious
assaults on prison staff: A descriptive analysis. Journal of Criminal Justice, 39(2),
143-150.
Sparks, J. R., & Bottoms, A. E. (1995). Legitimacy and order in prisons. British
Journal of Sociology, 46(1), 45-62.
Stowell, J. I., & Byrne, J.M. (2008). Does what happens in prison stay in prison?
Examining the reciprocal relationship between community and prison culture. In
J. M. Byrne, D. Hummer, and F. S. Taxman (Eds.), The culture of prison violence
(pp. 27-39). Boston, MA: Pearson.

328 • National Institute of Justice | NIJ.gov

Suedfeld, P. (1974). Solitary confinement in the correctional setting: Goals,
problems, and suggestions. Corrective and Social Psychiatry and Journal of
Behavior Technology, Methods and Therapy, 20(3), 10-20.
Sundt, J. L., Castellano, T. C., & Briggs, C. S. (2008). The sociopolitical context of
prison violence and its control: A case study of supermax and its effect in Illinois.
The Prison Journal, 88(1), 94-122.
Sundt, J. L., Castellano, T. C., & Briggs, C. S. (2004). Modern prison work:
Organizational and contextual determinates of job quality. Washington, DC:
U.S. Department of Justice, National Institute of Justice.
Sykes, G. M. (1958). The society of captives: A study of a maximum security prison.
Princeton, NJ: Princeton University Press.
Taxman, F. S., & Belenko, S. (2012). Implementation of evidence based community
corrections and addiction treatment. New York, NY: Springer.
Toch, H. (1997). Corrections: A humanistic approach. Guilderland, NY: Harrow
and Heston.
Toch, H. (2001). The future of supermax confinement. The Prison Journal, 81(3),
376-388.
Toch, H. (2003). The contemporary relevance of early experiments with
supermax reform. The Prison Journal, 83(2), 221-228.
Toch, H. (2007). Sequestering gang members, burning witches, and subverting
due process. Criminal Justice and Behavior, 34(2), 274-288.
Toch, H., & Adams, K. (1986). Pathology and disruptiveness among prison
inmates. Journal of Research in Crime and Delinquency, 23(1), 7-21.
Toch, H., & Kupers, T. A. (2007). Violence in prisons, revisited. Journal of
Offender Rehabilitation, 45(3-4), 1-28.
Travis, J., Western, B., & Redburn, F. S. (2014). The growth of incarceration in the
United States: Exploring causes and consequences. Washington, DC: The National
Academies Press.
Tyler, T. R. (2003). Procedural justice, legitimacy, and the effective rule of law.
Crime and Justice, 30, 283-357.
United States Government Accountability Office. (2013). Improvements needed
in Bureau of Prisons’ monitoring and evaluation of impact of segregated housing.
Washington, DC: Author.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 329

Useem, B. (1985). Disorganization and the New Mexico prison riot of 1980.
American Sociological Review, 50(5), 677-688.
Useem, B., & Piehl, A. M. (2006). Prison buildup and disorder. Punishment &
Society, 8(1), 87-115.
Wall, A. T. (2016). Time-in-cell: A practitioner’s perspective. Yale Law Journal
Forum, 125, 246-252.
Wooldredge, J., & Steiner, B. (2015). A macro-level perspective on prison inmate
deviance. Punishment & Society, 17(2), 230-257.

330 • National Institute of Justice | NIJ.gov

CHAPTER 9

Toward an Understanding
of “What Works” in
Segregation: Implementing
Correctional Programming and
Re-Entry-Focused Services in
Restrictive Housing Units
Paula Smith, Ph.D.
School of Criminal Justice, University of Cincinnati

Introduction

I

ssues related to prison management have been the topic of heated debate
over the years (Gendreau & Keyes, 2001). Within this context, the use of
“solitary confinement” — also known by terms such as restrictive housing,
administrative segregation, and disciplinary segregation — has generated special
attention and controversy, with repeated calls to abolish the practice. It is
instructive, however, that such housing units have been used since the inception
of the prison to isolate inmates from the general population of offenders for both
punitive (i.e., disciplinary) and administrative (i.e., safety) purposes (Labrecque &
Smith, 2013). In fact, as the incarceration rate increased dramatically between the
1970s and 2008 in the United States (Glaze & Kaeble, 2014; Simon, 2007), many

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 331

prison managers felt that they had few options for controlling the institutional
environment but to place extreme limits on prisoners’ activities (Mears &
Watson, 2006; Pizarro & Stenius, 2004). Therefore, it is perhaps not surprising
that several recent studies have indicated that the use of segregation has also
increased over the past two decades in the United States (Haney, 2008; O’Keefe,
2008). Coupled with the fact that segregation is a more restrictive and very
expensive option for housing inmates, it is critical to understand both the utility
and efficacy of solitary confinement (Frost & Monteiro, 2016).
At the moment, there is an active debate about the constitutional and
humanitarian concerns related to the use of segregation, as well as its
effectiveness as a tool for managing prisons (Labrecque & Smith, 2013). On one
side, corrections officials underscore the daily challenge of managing disruptive
inmates in crowded prisons. They cite the very practical reality that segregation
is a necessary tool in the continuum of placement options within correctional
institutions, particularly for inmates who pose a threat to themselves or others.
On the other side, critics assert that segregation is one of the most restrictive
and stressful settings for offenders to experience (Arbour, 1996; Human Rights
Watch, 2000). From this perspective, solitary confinement is held to violate basic
human rights, with harsh conditions that are unduly severe and disproportionate
to legitimate security and inmate management objectives (Human Rights Watch,
2000). For these reasons, advocates for prison inmates have characterized the
practice as “cruel and unusual punishment” (Grassian, 1983), citing a lengthy
list of objectionable conditions: lack of windows, 24-hour lighting, minimal
opportunities for exercise and recreation, restricted interpersonal contact,
removal of privileges, denial of personal items, and limited therapeutic services.
Support for finding alternatives to managing prisons safely and humanely has
gained momentum. In fact, several recent legislative changes have been announced
to limit the use of solitary confinement in adult inmate populations, and
completely eliminate its use with juvenile inmate populations in the United States
(e.g., Eilperin, 2016). Recently, the U.S. Department of Justice (2016) released a
document entitled Report and Recommendations Concerning the Use of Restrictive
Housing to delineate guidelines, or “best practices,” for correctional agencies to
consider. In general, there is renewed interest in developing and implementing
theoretically informed and empirically valid approaches to support rehabilitation
for inmates in restrictive housing units (Frost & Monteiro, 2016).
Given concerns over whether solitary confinement constitutes “cruel and
unusual punishment,” it is not surprising that the vast majority of empirical
research conducted to date has focused on whether or not segregation produces
any adverse physiological or psychological effects on inmates (see Gendreau
& Labrecque, in press; Labrecque & Smith, 2013). Far less attention has been
paid to the impact of segregation on subsequent institutional behavior and
post-release recidivism, despite the considerable benefit that such knowledge
would represent for correctional agencies worldwide (Labrecque & Smith, 2013).

332 • National Institute of Justice | NIJ.gov

Simply put, segregation has historically been viewed as a form of punishment
and incapacitation within the prison system rather than as a mechanism for
facilitating rehabilitation. Moreover, there are even fewer evaluations of offender
rehabilitation programs or services in restrictive housing units. This neglect is
especially consequential because those receiving what is typically an institution’s
most severe sanction are often the very inmates in critical need of services to
support both their short-term compliance with institutional rules and long­
term behavioral change. Furthermore, inmates admitted to restrictive housing
facilities often possess certain characteristics (e.g., mental illness, gang affiliation,
low self-control) that create significant barriers to managing them in the general
population of offenders (Gendreau & Thériault, 2011).
Although some disagreement exists over the short-term effects of brief periods
of isolation, there is a general consensus that solitary confinement for prolonged
periods is inhumane and causes long-term harm. Some of the most vulnerable
inmate populations (e.g., offenders with mental illness) are at the highest risk
for lengthy periods of incarceration in restrictive housing units. Advocates of
offender rehabilitation and prison reform contend that solitary confinement
represents a passive correctional intervention that often reinforces short-term
solutions to problems with inmates (Gendreau & Labrecque, in press). This is
especially prescient when administrative policies, clinical assessment protocols,
and treatment interventions are in place that can limit the use of segregation
while maintaining institutional safety and promoting behavioral change
(Gendreau & Thériault, 2011). In short, prisons cannot expect to rehabilitate
offenders merely by confining and restricting them (Cullen & Gilbert, 1982).
In fact, segregating inmates may ultimately undermine legitimate attempts
to rehabilitate them (Rothman, 1980). Moreover, the results of research by
Lovell, Johnson, and Cain (2007) suggest that inmates released directly from
administrative segregation into the community have higher recidivism rates
compared to offenders released from the general prison population. This finding
is consistent with a handful of other studies that examine the post-release
outcomes of inmates who are released directly from segregation (e.g., Mears &
Bales, 2009; Seale, Atkinson, Grealish, Fitzgerald, Grassel, & Viscuso, 2011; Ward
& Werlich, 2003).
For all of these reasons, the need for services to assist offenders in segregation
cannot be understated. Several jurisdictions have applied evidence-based
correctional practices within the context of administrative segregation to reduce
subsequent institutional misconduct and post-release recidivism (e.g., North
Dakota Department of Correction and Rehabilitation, Ohio Department of
Rehabilitation and Correction, Washington State Department of Corrections).
Many other prison authorities are engaged in similar initiatives (e.g., The Vera
Institute of Justice; see Shames, Wilcox, & Subramanian, 2015).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 333

Within this context, it is important to understand “what works” in solitary
confinement to improve inmate behavioral outcomes. This white paper addresses
the issue from an evidence-based perspective. The first section undertakes a brief
review of what is known about the impact of segregation on inmate institutional
adjustment.1 The second section summarizes the principles of effective
intervention and provides a framework for how correctional programming and
re-entry-focused services might be integrated into restrictive housing units.
The third section presents specific recommendations to guide the design and
implementation of evidence-based services in segregation. It provides specific
examples from select jurisdictions to demonstrate how some of these concepts
are being implemented.
As a prelude to this discussion, it is important to define the terminology used
in this white paper. Within correctional contexts, the terms used to describe
segregation policies and practices vary greatly across jurisdictions (Frost
& Monteiro, 2016; Labrecque & Smith, 2013). For example, it is difficult
to separate the literature on disciplinary segregation from the literature on
administrative segregation; the former refers to short-term confinement after
a specific infraction, whereas the latter refers to long-term classification to a
restrictive housing unit. Furthermore, researchers have tended to study solitary
confinement without carefully distinguishing the various types of segregation
(Frost & Monteiro, 2016). There is no universal definition of segregation, nor is
there consensus about who should be placed in such settings (Frost & Monteiro,
2016; Riveland, 1999). Nevertheless, most of these units have enough distinctive
features in common to be analyzed together. For this white paper, the use of the
term administrative segregation has been avoided in favor of the more general
term segregation or restrictive housing units.

The Effects of Solitary Confinement
As previously mentioned, whether segregation produces any harmful effects
has been a longstanding debate in the field of corrections (Gendreau &
Thériault, 2011). The literature reveals two very different perspectives. While
some researchers have characterized segregation as psychologically harmful
(Grassian, 1983; Grassian & Friedman, 1986; Haney, 2008, 2009; Jackson,
2001; Smith, 2006), others have argued that the empirical literature suggests
that segregation produces minimal, if any, negative psychological effects when
used for relatively short periods and under reasonable conditions of confinement
(Bonta & Gendreau, 1995; Ecclestone, Gendreau, & Knox, 1974; Gendreau &
Bonta, 1984; Gendreau & Thériault, 2011; Gendreau, Freedman, Wilde, & Scott,
1972; O’Keefe, 2008; O’Keefe, Klebe, Stucker, Sturm, & Leggett, 2010; Suedfeld,
1984; Zinger, Wichmann, & Andrews, 2001). It is important to emphasize here
1

This topic is considered in detail by other contributions commissioned by the National Institute of Justice.

33 4 • National Institute of Justice | NIJ.gov

that both perspectives generally agree that prolonged periods of segregation
should be avoided, and that inmates should be housed in the least restrictive
setting necessary for maintaining the safety and security of the institution (U.S.
Department of Justice, 2016).
It is critical that practitioners understand the short-term effects of segregation,
as they have important implications for assessing, treating, and delivering
services in restrictive housing units. If short-term placement in segregation does
not produce dramatic adverse effects under certain conditions, then it seems
reasonable to further investigate how this context might be used to deliver more
individualized and intensive interventions to inmates in need. This white paper
briefly reviews the available research on the effects of segregation (for a detailed
review, see Frost & Monteiro, 2016; Gendreau & Labrecque, in press).
The empirical literature on restricted environmental stimulation (Suedfeld, 1980),
or sensory deprivation, is relevant to the discussion here about the physiological and
psychological effects of segregation (Zubek, 1969). The sensory deprivation literature
was the first to suggest that such environments could be harmful. It also reveals a
crucial methodological problem that is present when sensory deprivation is enforced.
Some of the first sensory-deprivation experiments were conducted at McGill
University in the 1950s, and the researchers reported dramatic cognitive
deterioration and perceptual impairment in samples of college students
(e.g., Bexton, Heron, & Scott, 1954). However, subsequent studies failed to
replicate these findings (e.g., Zubek, Bayer, & Shephard, 1969). The reason for
this inconsistency was eventually recognized in a landmark study by Orne and
Scheibe (1964); namely, a strong placebo effect occurred when care was not taken
regarding how information was elicited from participants (see also Zubek, 1969;
Hunt & Chefurka, 1976). Specifically, it was noted that, “… subjects’ behavior
can be differentially manipulated by altering the implicit and explicit clues in the
experimental situation, and further (they) may react to social cues or demand
characteristics in such a way as to confound experimental results” (Orne &
Scheibe, 1964, p. 10).
In the early 1960s, researchers with the Canadian Penitentiary Service noted that
solitary confinement cells had some physical resemblance to the conditions of
sensory deprivation in previous experimental studies. As a result, the researchers
were interested in how inmates responded to isolation in solitary confinement
cells. Clearly, the policy implications of this research for the field of corrections
would be profound. Gendreau, Freedman, Wilde, and Scott (1968, 1972) as
well as Eccelstone and colleagues (1974) conducted a series of studies in which
inmates spent between two and eight days in segregation. The results indicated
that inmates in segregation exhibited lower EEG frequency as measured by
visually evoked potentials (which is indicative of lowered sensory arousal and
cortical activity as well as a need for sensory stimulation), and lower stress levels
as indicated by plasma cortisol levels. Other studies by this group also failed

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 335

to find adverse effects on inmates’ physical health, auditory functioning, and
discrimination learning ability (for a review, see Bonta & Gendreau, 1990).
Subsequently, studies were conducted with greater ecological validity in which
inmates were admitted to segregation involuntarily (Andersen, Sestoft, Lillebaek,
Gabrielsen, & Hemmingsen, 2003; Suedfeld, Ramirez, Deaton, & Baker-Brown,
1982; Wormith, Tellier, & Gendreau, 1988; Zinger et al., 2001). Collectively, these
studies involved longitudinal assessments, repeated measure comparison group
designs, multisite replications, different forms of segregation, male and female
samples, and standardized assessments. The strongest effects were reported for
increases in hostility and depression, but the effect sizes were still rather small.
Two decades after the publication of Orne and Scheibe’s (1964) critique of the
sensory deprivation research, two studies commanded considerable attention in
the prison literature. First, Grassian (1983) claimed that segregation produced
psychological harm (e.g., hallucinations, overt psychotic disorganization, massive
free-floating anxiety, primitive aggressive fantasies, paranoia, and lack of impulse
control leading to random violence). His assessment protocol consisted of openended interviews and an interview style that actively encouraged disclosure and
provided reassurance. Second, Haney (2003) generated similar results almost 20
years later in a sample of 100 inmates in a supermax prison. Again, data were
collected through interviews, although it is unclear exactly what measures were
used. Furthermore, there was no indication as to whether the prevalence of the
symptoms reported by the sample existed prior to incarceration, or how long
these effects persisted after the assessment.
More recently, other scholars have noted that the methodological shortcomings
of the research reporting harmful effects (e.g., selection bias, response bias,
inadequate or no control groups) has limited the credibility of the results (see
Gendreau & Labrecque, in press; Labrecque & Smith, 2013; Suedfeld et al.,
1982; Zinger et al., 2001). Several primary studies and two recent independent
meta-analytic reviews on this topic have been completed (Smith, Gendreau, &
Labrecque, 2015; Morgan, Van Horn, MacLean, Bolanos, Gray, Batastini,
& Mills, 2014). Both meta-analytic reviews conclude that the mean effect size for
psychological outcomes tends to be in the small to moderate range.
In short, if there are outliers in the empirical literature, they appear to be from
studies that claim segregation produced dramatic adverse psychological effects
(Gendreau & Labrecque, in press). The vast majority of the empirical studies on
segregation point to a similar conclusion: the negative effects associated with
relatively brief periods of segregation are not nearly as dramatic as once feared
(Smith et al., 2015; Gendreau & Thériault, 2011; Morgan et al., 2014). At the
same time, more empirical research is needed on the psychological effects of
segregation. If the conditions of confinement are humane, however, there is good
reason to support the idea that restrictive housing units can be managed in a

336 • National Institute of Justice | NIJ.gov

manner that allows for the delivery of intensive interventions to inmates in need
of services for successful transition into the general population of offenders.
It is also important to note that a very limited number of studies have been
conducted to examine the impact of segregation on behavioral outcomes.2
Within this limited research base, there are three types of behavioral outcomes
of interest: institutional violence, post-release recidivism, and institutional
misconduct. Once again, two independent meta-analyses have summarized the
available studies (Smith et al., 2015; Morgan et al., 2014), and small mean effect
sizes were reported for behavioral outcomes. The paucity of research on this topic
is rather alarming given that segregation is often described as an important tool
in ensuring systemwide order in prison systems (Mears & Watson, 2006). These
behavioral outcomes must be a priority for future research. This issue will be
further discussed in a later section of this white paper.

The Principles of Effective Intervention
Correctional rehabilitation refers to planned interventions that target for change
some characteristic of the offender that causes criminality (e.g., attitudes,
cognitive processes, personality factors or mental health, social relationships,
educational and vocational skills, and employment), and intend to make
the offender less likely to recidivate (see Andrews & Bonta, 2010; Cullen &
Gendreau, 2000).3 This requires the specification of what to target (i.e., dynamic
risk factors, or criminogenic needs), who to target (i.e., higher-risk offenders),
and how to target (i.e., cognitive-behavioral and social learning treatment
modalities). Collectively, these fundamental concepts are referred to as the
principles of effective intervention.

Setting the Stage for the Principles of Effective Intervention
The principles of effective intervention were established as a result of a threestage research agenda. Initially, researchers conducted narrative literature
reviews, and generated recommendations in consultation with colleagues who
had conducted successful programs. Second, demonstration projects were
conducted to evaluate the effectiveness of correctional treatments (e.g., Andrews,
1979, 1980; Andrews & Keissling, 1980; Gendreau & Ross, 1979, 1987; Ross &
Fabiano, 1985). Finally, meta-analytic techniques were applied to the corrections

2

See, for example, Briggs, Sundt, & Castellano, 2003; Butler, Steiner, Makarios, & Travis, in press; Huebner, 2003;
Lovell et al., 2007; Labrecque, 2015; Mears & Bales, 2009; Mears & Castro, 2006; Morris, 2015; Seale et al.,
2011; Sundt, Castellano, & Briggs, 2008; Ward & Werlich, 2003.

3

Correctional rehabilitation does not include interventions that seek to suppress criminal behavior through specific
deterrence (i.e., use of punishment and sanctions).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 337

literature to generate a more precise estimate of the empirical support for
the principles of risk, need, and responsivity (RNR). Compared to narrative
reviews, meta-analysis involves the quantitative synthesis of research, and thus
yields a more precise estimate of the overall mean effect size (for a review, see
Gendreau & Smith, 2007). Meta-analysis is the review method of choice for
many disciplines, including corrections, which has more than 44 meta-analyses
of correctional treatment effectiveness (see McGuire, 2004; Smith, Gendreau,
& Swartz, 2009). Taken together, the results of these studies provide strong
empirical support for the principle of effective intervention. This section presents
the findings from three categories of meta-analyses: (1) those which affirm
that, overall, correctional treatment programs reduce recidivism; (2) those that
identify some general principles of “what works” in reducing offender recidivism;
and (3) those that search for more specific clinically and psychologically relevant
criteria (for a detailed review, see Smith, 2013).
Garrett (1985) and Davidson, Gottschalk, Gensheimer, and Mayer (1984)
published the first meta-analyses in the field of corrections. Garrett (1985)
synthesized 433 effect sizes from studies of 13,000 juvenile offenders and reports
a mean effect size of r = .12. Furthermore, the results indicate that cognitivebehavioral interventions are associated with the largest mean effect size
(r = .22). Davidson and colleagues (1984) produced similar results in that
behavioral interventions (e.g., positive reinforcement, token economies,
behavioral contracts) are associated with the greatest reductions in recidivism.
Subsequently, Lipsey (1992) analyzed the results of a large database of juvenile
interventions (i.e., a total of 443 effect sizes). Sixty-four percent of these estimates
are in the expected direction (i.e., reduced recidivism), and the average reduction
in recidivism varies from 5 percent to 9 percent depending on statistical
adjustments. Similarly, Lösel (1995) provides a comprehensive assessment of 13
meta-analyses of juvenile and adult offenders published between 1985 and 1995,
and reports that mean effect sizes ranged from r = .05 to .18 with an overall mean
of about r = .10. This basic pattern of results remains even after controlling for
the effects of several other factors, including subject attrition, methodological
quality, length of follow-up, and study publication status.
In summary, the meta-analyses referenced above identify the most effective
treatment programs as those which are cognitive-behavioral in nature, have a
high degree of structure, are demonstration programs (rather than “real world”
or “routine” correctional programs), and are delivered in the community rather
than in institutional settings (see also Cleland, Pearson, Lipton, & Yee, 1997;
Izzo & Ross, 1990; Lipsey, 1999; Lipsey, Chapman, & Landenberger, 2001; Lösel,
1995; Redondo, Sanchez-Meca, & Garrido, 1999). Although skeptics might argue
that the 10 percent reduction in recidivism found by Lösel (1995) is of little
practical value, several authors have demonstrated that the finding represents a
meaningful and cost-effective reduction in recidivism (Lipsey & Wilson, 1993;

338 • National Institute of Justice | NIJ.gov

Rosenthal & DiMatteo, 2001; Aos, Lieb, Mayfield, Miller, & Pennucci, 2004;
Cohen, 1998; Welsh & Farrington, 2000).

Developing the Principles of Effective Intervention
The next series of meta-analyses searched for more specific “clinically relevant
and psychologically informed” principles of effective offender treatment
(Andrews, Zinger, Hoge, Bonta, Gendreau, & Cullen, 1990). A meta-analysis
conducted by the Andrews team (1990) coded the treatment literature (a total of
154 effect sizes) along various dimensions that provided the basis for developing
the RNR framework. This database was subsequently extended to 374 effect sizes
(see also Andrews & Bonta, 2010, pp. 365-369). In short, the results indicate
that there was considerable heterogeneity in the effectiveness of correctional
interventions; that is, correctional programs that have certain characteristics
yield much larger effect sizes when compared to approaches that do not
(Andrews et al., 1990). This section reviews the three main principles of effective
intervention (for a detailed review of the meta-analytic evidence for the RNR
framework, see Smith et al., 2009).

The Need Principle
To develop a comprehensive theoretical framework, it is necessary to first
identify the covariates of crime — that is, the biological, personal, interpersonal,
situational, and social variables that are statistically associated with antisocial
behavior (Andrews & Bonta, 2010). These variables include both static predictors
(e.g., criminal history) as well as dynamic factors (e.g., antisocial attitudes,
pro-criminal peers, substance abuse). The latter criminogenic needs are the
appropriate targets for intervention because they are amenable to change.
Several meta-analyses have demonstrated that certain criminogenic needs are
robust predictors of recidivism (Bonta, Law, & Hanson, 1998; Gendreau, Little,
& Goggin, 1996; Hanson & Morton-Bourgon, 2004). These predictors include
(1) an antisocial personality pattern (e.g., aggression, hostility, impulsivity, lack
of self-control, poor emotion regulation); (2) antisocial attitudes, values, and
beliefs; (3) the presence of antisocial peers and associates; (4) substance abuse;
(5) problematic circumstances within family/marital relationships; (6) difficulties
with education and employment; and (7) lack of prosocial leisure and recreation
activities. Together with criminal history, the first three criminogenic needs
identified are referred to as first-tier predictors because the predictive validities
associated with these covariates are especially robust (Andrews & Bonta,
2010). The most effective treatment programs target criminogenic needs and
prioritize the first-tier predictors in this regard — the need principle in the RNR
framework. In fact, Andrews and Bonta (2010) report that treatment programs
targeting criminogenic needs reduce recidivism by 20 percent more than
programs that do not target them. Moreover, these meta-analyses also found that

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 339

other factors had weak predictive validities (e.g., low self-esteem, depression,
anxiety, fear of official punishment) and should therefore not be the primary
targets for intervention (for a detailed review, see Gendreau et al., 1996).
The predictors of institutional misconduct are very similar to the predictors
of post-release recidivism (French & Gendreau, 2006). Furthermore, the same
correctional interventions that are effective in decreasing post-release recidivism
also lead to meaningful reductions in institutional misconduct (see French
& Gendreau, 2006). This is particularly relevant for determining how to best
implement programming in segregation, knowing that the appropriate treatment
targets for improving institutional adjustment (and therefore post-release
recidivism) can be identified and targeted through structured interventions
(French & Gendreau, 2006).

The Risk Principle
Research consistently indicates that higher-risk offenders derive the most benefit
from treatment (Andrews & Bonta, 2010). In fact, treatment programs that target
higher-risk samples reduce recidivism by 7 percent more than programs that
target lower-risk offenders. From a theoretical viewpoint, this finding makes
sense; higher-risk offenders, by definition, are likely to have more criminogenic
needs and therefore require more intense treatment. In contrast, participation
in treatment services can increase the failure rates of lower-risk samples by
disrupting protective factors and exposing them to their higher-risk counterparts
(see Lowenkamp, Latessa, & Smith, 2006). This is referred to as the risk principle.

The Responsivity Principle
Finally, the general responsivity principle describes how to best target criminogenic
needs. The meta-analyses of earlier studies have consistently found that the most
effective interventions are those that were cognitive-behavioral in nature. In fact,
Andrews and Bonta (2010) reported that cognitive-behavioral interventions
produced 19 percent greater reductions in recidivism when compared to other
models of offender treatment.
In addition to the general responsivity principle, Andrews and Bonta (2010) also
underscore the importance of specific responsivity factors. This refers to the need
for corrections practitioners to match the mode and style of service delivery with
key offender characteristics (e.g., offenders with lower IQs derive more benefit
from behavioral approaches than cognitive strategies; Cullen, Gendreau, Jarjoura,
& Wright, 1997).

The Effectiveness of RNR Treatment Programs
Previous research demonstrates that adhering to the RNR framework has a
cumulative effect (for a review, see Andrews & Bonta, 2010). When treatment

3 4 0 • National Institute of Justice | NIJ.gov

programs are categorized by whether they followed all three RNR principles in
contrast to those that did not, Andrews and Bonta (2010) report a 23 percent
difference in recidivism. These principles also apply to a variety of corrections
populations, including female offenders, minority groups, youthful offenders,
mentally disordered, violent, and sex offenders (Andrews, Dowden, & Rettinger,
2001; Dowden & Andrews, 2000).

Core Correctional Practices
The clinical skills related to effective service delivery with offender populations
are referred to as core correctional practices (Andrews & Bonta, 2010; see also
Gendreau, Andrews, & Thériault, 2010). They include effective reinforcement,
disapproval, and use of authority; relationship practices; structured skill-building
(including problem-solving); and cognitive restructuring. These therapeutic
practices are consistent with the cognitive-behavioral model of treatment, and
are associated with reductions in recidivism that range from 19 percent to 27
percent for programs that apply them versus those that do not (Andrews &
Bonta, 2010). These competencies can be used by all front-line staff members —
in the daily interactions between officers and inmates — and in clinical sessions,
group interventions, and case-management meetings (Andrews & Bonta, 2010).
Each of these competencies is briefly described below.
Corrections professionals must be capable of using high-level social reinforcement
to encourage prosocial behaviors, as well as effective disapproval to discourage
antisocial behaviors (Gendreau et al., 2010). Effective reinforcement involves
providing specific praise and acknowledgment for desirable behaviors, and
requires the offender to think about both the short-term and long-term
benefits associated with its continued use. Effective disapproval involves
providing statements of non-support for undesirable behaviors, and requires
the offender to think about both the short-term and long-term costs associated
with its continued use. Once the undesirable behavior has been corrected
and the offender engages in an appropriate prosocial behavior, it is important
that the staff member immediately terminate disapproval and provide social
reinforcement for the change.
Most corrections professionals are in positions of power relative to the offender,
and must use their authority to respectfully guide the offender toward compliance
(Gendreau et al., 2010). Staff members are encouraged to focus their message
on the behavior exhibited (and not on the person performing it), to be direct
and specific concerning their demands, and to specify the offender’s choices and
attendant consequences in any given situation. The guidelines associated with the
effective use of authority are particularly important in segregation units where
inmates are often not compliant with rules and staff requests.
In addition, staff should adopt several important relationship practices to help
them develop a collaborative working relationship (also referred to as the

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 3 41

therapeutic alliance in the clinical counseling literature) with the offender
(Gendreau et al., 2010). From this perspective, the most effective front-line
staff members are open, warm, respectful, nonjudgmental, empathic, flexible,
enthusiastic, and engaging. Furthermore, it is important for corrections
professionals to use humor and express optimism, and to be solution-focused,
structured, and directive. Moreover, front-line staff members should avoid
arguments and power struggles with offenders, and instead work to enhance
internal motivation and self-efficacy within the offender (Gendreau et al., 2010).
Another core correctional practice involves structured skill building (Gendreau et
al., 2010). Goldstein (1986) identified five main components of this process:
1. Define the skill to be learned by describing it in discrete steps.
2. Model or demonstrate the skill for the client.
3. Have the client practice the new skill by role playing it, and provide
corrective feedback.
4. Use homework assignments to generalize use of the skill beyond the
treatment setting.
5. Have the offender practice the skill in increasingly difficult situations, and
provide feedback (i.e., graduated rehearsal).
Previous research has underscored the importance of problem solving as a specific
social skill that should be taught to offenders because, once mastered, they can
apply it to a wide variety of high-risk situations (see Trotter, 1999).
Finally, corrections professionals should be thoroughly trained in cognitive
restructuring. Front-line staff members should be able to teach clients how to
generate descriptions of problematic situations, as well as the associated thoughts
and feelings that accompany them. Corrections professionals must then help
offenders identify risky thinking and practice replacing this self-talk with more
prosocial alternatives. Many correctional programs use thinking reports (e.g.,
Bush, Bilodeau, & Kornick, 1995) to assist clients in identifying risky thoughts
and feelings and how these affect their behavior.
In summary, it is important to acknowledge that beyond a theoretical
understanding of the variation in criminal behavior and the principles of
effective intervention is a need for a pragmatic “how to do it” that they can teach
offenders in order to change offenders’ behaviors. Research on the principles
of effective intervention has led to the development of numerous composite
offender risk and need assessments, structured treatment interventions, and
program evaluation instruments. Many of these same tools can be adapted
for use in restrictive housing units, and will be further discussed in the
recommendations provided later in this document.

3 42 • National Institute of Justice | NIJ.gov

Understanding the Limits of Deterrence and Punishment
Despite a plethora of research on the principles of effective intervention,
corrections practitioners continue to implement strategies that are ineffective —
and that might even cause greater harm to offenders than good (Latessa, Cullen,
& Gendreau, 2002).4 The term correctional quackery describes programs that
are developed without considering the principles of effective intervention, and
instead rely on common sense, personal experience, and conventional wisdom
(Latessa et al., 2002). For example, punishment-oriented strategies
(i.e., intermediate sanctions such as intensive supervision, house arrest,
electronic monitoring, boot camps, Scared Straight programs) have not been
determined to be effective in reducing recidivism, and yet continue to be
frequently implemented (Gendreau, Goggin, Cullen, & Andrews, 2000).
An increasing amount of research has also challenged the notion that
incarceration functions as an effective deterrent. To illustrate, the empirical
literature on offender re-entry has documented high levels of parole failure
for inmates released from prison (see Hughes, Wilson, & Beck, 2001; Langan
& Levin, 2002; Petersilia, 2003; Pew Center on the States, 2011). Furthermore,
a number of methodologically rigorous studies with diverse samples have
demonstrated that the effect of prison is, if anything, a slight to moderate
increase in post-release recidivism (Cid, 2009; Jonson, 2010; Nagin, Cullen,
& Jonson, 2009; Nieuwbeerta, Nagin, & Blokland, 2009; Sampson & Laub,
1993; Spohn & Holleran, 2002; Villettaz, Killias, & Zoder, 2006). Perhaps
even more relevant to the segregation debate is a smaller amount of literature
that has examined the relationship between the conditions of confinement
and recidivism. Chen and Shapiro (2007) measured the harshness of prison
conditions by level of security (e.g., minimum versus maximum). Controlling
for offender risk level, they find that harsher prison conditions do not lead to
higher levels of deterrence and, “if anything … may lead to more post-release
crime” (Chen & Shapiro, 2007, p. 1). Gaes and Camp (2009) report similar results
with a sample of offenders randomly assigned to higher- versus lower-security
correctional institutions. Inmates in the higher classification had a hazard rate
of reincarceration that was 31 percent higher than that for inmates in the lower
classification. A handful of other studies also find results that are consistent with
this general pattern that harsher prison conditions are associated with higher
post-release recidivism rates (Drago, Galbiati, & Vertova, 2008; Listwan, Sullivan,
Agnew, Cullen, & Colvin, 2013; Windzio, 2006).
How does this research relate to the restrictive housing units? It can be argued
that segregation routinely involves two separate but interrelated components:
(1) isolation (e.g., confinement in a single cell, restriction of social interaction);
and (2) deprivation (e.g., removal of personal items, denial of privileges).
4

Several reasons have been offered to explain why ineffective programs are so frequently implemented in the field of
corrections. See Gendreau, Goggin, and Smith (2001) for a more detailed consideration of the topic.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 3 4 3

Corrections officials often contend that both conditions are necessary to
maintain the safety and security of the institution. Although this is a legitimate
consideration in certain cases (e.g., threat of suicide or extreme violence), the
deprivation inherent in many segregation settings is more often intended as
punishment (see Mears & Castro, 2006). In other words, the harsh conditions
introduced in segregation are intended to be aversive, and therefore produce a
greater deterrent effect.5 In many cases, however, isolated confinement could be
accomplished with far less deprivation. This is an important point because the
introduction of deprivation and harsh conditions of confinement might even
undermine legitimate attempts at rehabilitation, in much the same way that
institutional climate can create barriers to effective service delivery in the general
population of inmates (Rothman, 1980).

Translating Research into Practical Recommendations
As previously discussed, there is a well-developed literature base on “what
works” to reduce offender recidivism. The principles of effective intervention
have now been extensively applied in both institutional and community-based
settings, and with diverse samples of offenders (for a review, see Andrews &
Bonta, 2010). Similarly, there is also a substantial literature base on “what doesn’t
work” to rehabilitate offenders (Gendreau et al., 2000). Taken together, the “what
works” and “what doesn’t work” discoveries lead to a better solution — the RNR
framework, which provides a blueprint for how services should be designed and
delivered in correctional settings. However, attempts to systematically use this
information to inform policies and practices within restrictive housing units have
been far less frequent. In fact, it is obvious from this review of the literature that
the integration of evidence-based practices within the context of segregation is
still in the very early stages of development, and published outcome evaluations
of treatment programs based on RNR principles are virtually nonexistent.
Instead, this section considers how the principles of effective intervention might
be applied to inmates in restrictive housing units.6 This does not constitute a
list of “best practices” given the lack of empirical evidence on the topic, but
it offers 10 recommendations that are theoretically relevant and grounded in
the “what works” literature. The recommendations included in this section
have implications for both research and practice. The reader should note that a
handful of these guidelines are based on those previously discussed by Gendreau
and Thériault (2011). Furthermore, all of these recommendations are consistent
with the recently published guiding principles issued by the U.S. Department of
Justice (2016) concerning the contemporary use of segregation.

5

This is further underscored by the term restrictive housing unit, implying the removal of privileges and liberties.

6

Examples of innovative programs that have piloted components of the recommendations included in this section
are highlighted. In most cases, formal process and outcome evaluations are ongoing, and the results have not
been published.

3 4 4 • National Institute of Justice | NIJ.gov

Recommendation 1: Adopt a meta-analytic perspective to
encourage knowledge cumulation.
It is difficult to achieve clarity in a field where scholars are divided (Hunt, 1997).
This is certainly true in the empirical literature on the short-term effects of
segregation. Moreover, conflicts within civil rights, moral, and political agendas
can lead to further confusion (Gendreau & Thériault, 2011; Jackson, 2002; Mears
& Watson, 2006). Gendreau and Thériault (2011) note that debates and literature
reviews in the field of corrections have frequently been framed in narrow — and
often ideological — frames of reference, and convergent validity is compromised
(see also Gendreau, 1996; Gendreau & Ross, 1979). The discourse that follows
from this state of affairs is often antithetical to systematic attempts at knowledge
cumulation (Hunt, 1997).
The previous section referred to meta-analysis as a quantitative review of the
literature that is the review method of choice in most disciplines, including
corrections (Smith et al., 2009). Single studies offer limited information; useful
policies in the social sciences are based on replication with diverse samples in
multiple jurisdictions before sound conclusions are reached (Hunter & Schmidt,
1996; Schmidt, 1992). The importance of meta-analysis in this regard cannot be
overstated; the results of systematic quantitative reviews can have a significant
impact on policy and practice. The principles of effective intervention presented
in the previous section are based on the results of meta-analysis, and the main
findings have been replicated with remarkable consistency (Smith et al., 2009).
In comparison, the literature on the effects of segregation has only recently been
summarized using meta-analytic techniques (see Smith et al., 2015; Morgan
et al., 2014), and the conclusions drawn were limited by the fact that relatively
few studies were eligible for inclusion. Furthermore, information on important
moderators could not be systematically analyzed due to small sample sizes and
missing data.
There is an urgent need for research to investigate segregation as a correctional
policy and rehabilitative practice. Future studies should examine the effects of
segregation on behavioral outcomes (i.e., institutional violence, post-release
recidivism, institutional misconduct) and for special populations of offenders
(e.g., those with mental illness, juvenile offenders) to determine the traits of
offenders who do not respond well to segregation. Precious few evaluations of
correctional treatment services in restrictive housing units have been published
to date (see Batastini, 2015; Pizarro, Zgoba, & Haugebrook, 2014). As the field
accrues more primary research, it will be exceedingly important to continue to
support knowledge cumulation and meta-analyses of the empirical literature
(Gendreau & Thériault, 2011; Labrecque & Smith, 2013). Meta-analysis has
another important advantage — it allows academics to systematically document
gaps in the literature to recommend future priorities for research. In short, a
meta-analytic perspective on topics related to segregation can inform priorities
for both research and practice.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 3 4 5

Recommendation 2: Monitor the prison environment and institutional
climate to prevent misconduct and reduce the need for segregation.
Crime prevention strategies that originate in community settings also can be
applied to correctional institutions (Gendreau & Thériault, 2011). For example,
structural features of a prison (e.g., pod designs, lighting, camera surveillance)
can be used to analyze patterns in the time and location of institutional
misconduct — to be proactive about the situations and environments where
incidents are most likely to occur. For example, the systematic analysis of data on
institutional misconduct might reveal that incidents are more likely to occur in
certain locations within the prison or during specific shifts under the supervision
of particular front-line staff members. This information can then provide prison
administrators with the knowledge they need to prevent these incidents. From
the perspective of the cognitive-behavioral model, these data allow corrections
professionals to better understand the risky situations that inmates will likely
encounter while incarcerated. Specific skills can then be taught to offenders to
equip them to make prosocial choices.
Similarly, information gained from analyzing aggregate statistics like inmate
and staff turnover rates can be used to understand fluctuations in institutional
misconduct (French & Gendreau, 2006; Porporino, 1986; Wortley, 2002). With
this information, prison administrators can identify when their institutions
might be particularly vulnerable to fluctuations in the rate of institutional
misconducts, and then implement measures to counteract predicted increases
in incidents. For example, prison administrators might use this information to
strategically inform staff assignments and inmate placements.
Perhaps even more significant is the observation that institutional misconduct
and the use of segregation cannot be understood without considering the
institutional climate and behavior management practices of the correctional
facility at large. A correctional institution with a therapeutic environment and
high-quality programs is much more likely to have lower rates of institutional
misconduct because its offenders are actively engaged in learning skills that they
can apply to avoid and manage risky situations (Andrews & Bonta, 2010; French
& Gendreau, 2006). Similarly, offenders who have privileges and other incentives
in the general population will be more motivated to comply with institutional
rules and progress in their treatment — and less motivated to spend time in
segregation settings. Therefore, prison administrators are advised to design their
institution’s schedules in a manner that ensures that inmates are consistently
engaged in meaningful, prosocial activities. For example, it is generally
recommended that inmates participate in therapeutic tasks for at least 35 hours
each week, and have access to a wide range of reinforcers to encourage program
participation and prosocial behavior (Gendreau et al., 2010).

3 4 6 • National Institute of Justice | NIJ.gov

Recommendation 3: Screen inmates at intake to determine risk for
placement in segregation.
Efforts to prevent or divert offenders from segregation can be greatly enhanced
by identifying offenders who are at risk of placement in segregation. Although
there has been little research on the topic, the available literature examining
individual-level and institutional-level predictors generally suggests that the
predictors of segregation may be similar to the predictors of other outcomes,
including institutional misconduct and post-release recidivism. Institutions
use many different types of risk measures or scales to predict institutional
adjustment. Some of the noteworthy scales include the Static Factor Assessment
(SFA; Motiuk, 1993), the Dynamic Factors Identification and Analysis (DFIA/
DFIA-R), the Custody Rating Scale (CRS; Solicitor General Canada, 1987), and
the Statistical Information on Recidivism-Revised (SIR-R1) scale (Nafekh &
Motiuk, 2002).
The construction of a new actuarial assessment scale for predicting placement
in segregation merits special comment here. Using data from the Offender
Management System maintained by the Correctional Service of Canada, Helmus
(2015) developed an instrument, the Risk of Administrative Segregation Tool
(RAST), specifically to predict placement in administrative segregation. The
study includes both a development sample (n = 11,110) and a validation sample
(n = 5,591) of offenders incarcerated in Canadian federal institutions. The tool
contains six static items (age, prior convictions, prior segregation placement,
sentence length, criminal versatility, and prior violence) with scores ranging
from 0 to 13. Results indicate that the RAST scale can predict placement in
administrative segregation for both security concerns and protective custody,
as well as placements in segregation within periods of one and two years after
admission. Additionally, the instrument has adequate predictive accuracy for
both non-Aboriginal and Aboriginal men and women (AUC = 0.80). The tool
also exhibits superior performance when compared to other risk scales used by
the Correctional Service of Canada for similar purposes (Helmus, 2015).
In the process of developing and validating the RAST scale, Helmus (2015)
provides several options for practitioners to use to designate nominal risk
categories (e.g., low, moderate, high). The options available depend on the
agency’s goals or criteria, but the selection of risk categories should be clearly
articulated, not arbitrary, and related to how the scale will be used in practice
(Helmus, 2015). Helmus also notes that the RAST scale is useful to correctional
agencies because all of the information required to score the items is available at
admission to the institution, and it does not require a significant investment of
time or resources to administer.
An important cautionary note is warranted here. The RAST scale is undoubtedly
an important contribution to the empirical literature because it identifies
offenders at higher risk for placement in segregation; however, it is a static

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 3 47

instrument. Therefore, it cannot be used to measure reductions in risk over time.
Although it can be applied as a screening tool, research into dynamic predictors
should continue to help prison officials further understand the criminogenic
needs that should take priority.

Recommendation 4: Implement programs and services based on
RNR principles to prevent misconduct.
By identifying inmates who are at higher risk for placement in segregation,
corrections administrators can provide interventions to teach offenders skills that
might prevent them from engaging in institutional misconduct, and therefore
reduce subsequent placements in administrative segregation. It is critical that
such programming and services have a solid basis in the RNR principles —
integrating cognitive-behavioral and social learning approaches to target the
known predictors of institutional misconduct. For example, a brief intervention
module is being developed for this purpose (Smith, 2016a). Through structured
sessions, inmates learn skills they can use to establish a prosocial support
network within the correctional institution. They also learn how to interact
with peers and staff members (including how to deal with authority and avoid
negative peer pressure). Additional sessions are also included to help offenders
learn to regulate their emotions (e.g., frustration tolerance) and solve problems.
Compelling meta-analytic evidence now shows that participation in general
cognitive-behavioral treatment reduces prison misconduct in addition to postrelease recidivism (French & Gendreau, 2006). Specifically, cognitive-behavioral
interventions that teach prosocial skills for addressing high-risk situations in
the community can also be used to enhance skills that offenders can apply in
correctional institutions. This also underscores the importance of the timing
of correctional programming, particularly for inmates with an elevated risk for
placement in segregation.

Recommendation 5: Transform segregation from a deprivation
environment to a therapeutic environment.
As discussed earlier, restrictive housing units have historically been defined
as environments involving both isolation and deprivation. However, previous
research has called into question the conventional wisdom that the harshness
of the prison condition functions as an effective deterrent (for a review, see
Listwan et al., 2013). Except in circumstances where the removal of personal
items is demonstrated to be necessary, there is good reason to believe that such
conditions of deprivation interfere with the delivery of effective correctional
programs (see Smith & Schweitzer, 2013). Arguably, the instances where there is
a legitimate concern that requires extreme restriction are relatively rare.

3 4 8 • National Institute of Justice | NIJ.gov

While most previous evaluations focused on the content of treatment, it is also
important to consider the context of the intervention (Smith & Schweitzer,
2013). Correctional officials must devote significant attention to transforming
segregation from a deprivation environment to a therapeutic environment.
Previous studies have underscored the importance of normalization within
the prison setting to create a more humanized environment (e.g., Centen &
Sampson, 1991). Although this will be challenging for many jurisdictions, such
efforts can create a context that is more conducive to offender rehabilitation.
Such initiatives should consider the aspects of the physical milieu, the availability
of correctional programming and rehabilitative services, access to meaningful
social interactions and other activities, access to privileges, and the content
of interactions between staff and inmates. Some departments (e.g., the North
Dakota Department of Corrections and Rehabilitation) are addressing this issue
in numerous innovative ways within restrictive housing units. Some of these
strategies will be discussed in conjunction with later recommendations.

Recommendation 6: Select the least restrictive option and limit the
use of segregation for prolonged periods.
Inmates should be housed in the least restrictive setting necessary to ensure
safety and security (U.S. Department of Justice, 2016). Correctional agencies
must be capable of articulating the specific reasons for an inmate’s placement
and retention in segregation (U.S. Department of Justice, 2016). Furthermore,
the development of a detailed case plan and regular reviews of inmate progress
should occur to ensure that this setting remains the most appropriate placement
option over time. Such reviews should be objective and based on documented,
observable behaviors.
In some cases, inmates may spend months or years in restrictive housing units
without the opportunity to engage in correctional programming or other
services. These inmates are deprived of human interaction, with the exception
of limited interaction with correctional officers and other corrections personnel
(e.g., mental health and medical professionals). To reiterate, there is little debate
that segregation for prolonged periods is inhumane, counterproductive, and
should be avoided (see Pizarro et al., 2014). One such initiative to limit the
use of segregation deserves particular comment here. The Vera Institute of
Justice is dedicated to developing a fairer, more humane, and more effective
criminal justice system (Browne, Cambier, & Agha, 2011). One of its efforts
working toward this goal is the Segregation Reduction Project and the Safe
Alternative to Segregation Initiative. As part of this initiative, the Vera Institute
of Justice partners with state departments of corrections to (1) reduce the
number of inmates in segregation, (2) improve the conditions of confinement
in segregation, and (3) enhance programming and support for safe transitions
back to the general prison population. Several recent partnerships have led to

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 3 49

the implementation of policies and strategies that have dramatically reduced
reliance on prolonged periods of segregation.7 Specific strategies include using
alternative sanctions for minor rule violations, reducing segregation time for
certain types of rule violations, reducing segregation time as an incentive for
sustained good behavior, and introducing step-down programs to facilitate the
inmate’s reintegration into the general population (Browne et al., 2011). Such
initiatives are critical because they can examine both system-level factors as well
as individual service delivery variables.

Recommendation 7: Divert inmates who cannot cope with
segregation to other placement options to reduce harmful effects.
Very little is known about the specific traits of inmates who cannot tolerate
segregation. Although the results are tentative, some examples of offender
characteristics associated with poor patterns of adjustment include high
stimulation seeking, impulsivity, low conceptual level, and low adrenal
functioning (Gendreau & Bonta, 1984; Gendreau & Thériault, 2011; Zubek,
1969). Collecting this type of data should not be onerous, as most prison systems
should have considerable information available in client files to determine the
types of inmates who have exhibited problems in the past (Gendreau & Thériault,
2011). These data are critical to identifying the specific types of offenders who
should be diverted from restrictive housing units to other placement options
within the system (e.g., secure prison hospital wards).
Inmates with mental health needs merit special consideration here. Correctional
institutions should use psychiatric screening measures to identify offenders
with mental health needs (Gendreau & Thériault, 2011). The specific measures
selected should have demonstrated adequate reliability and validity. Inmates with
serious mental illness pose unique challenges for restrictive housing units, and
services must be made available to prevent critical incidents, including self-harm
and suicide (Gendreau & Thériault, 2011).
Two views exist as to how inmates with mental illness might react to segregation.
The traditional criminological perspective (e.g., Mears & Watson, 2006) suggests
that such inmates are adversely affected by periods of isolation. On the other
hand, the psychiatric literature suggests that some inmates with mental illness
might react positively to solitary confinement because of the need for less
stimulation (Grassian & Friedman, 1986). In fact, corrections professionals
frequently observe that inmates with mental illness seek out solitary confinement
(see Scott & Gendreau, 1969). Similarly, previous studies have demonstrated that
offenders with mental illness often respond best to environments with reduced

7

Browne et al. (2011) describe partnerships with agencies in Illinois, Maryland, and Washington that have
implemented changes to reduce the use of segregation.

350 • National Institute of Justice | NIJ.gov

sensory input (Smith, Gendreau, & Goggin, 2008). An understanding of how
inmates with mental illness respond to segregation can inform the delivery of
rehabilitation programs and re-entry-focused services. Although more research
is needed in this area, it remains clear that assessment and services are critical
considerations for offenders with mental illness who are placed in restrictive
housing units.

Recommendation 8: Ensure that all front-line staff members are
trained and skilled in core correctional practices to facilitate both
short-term compliance with rules and long-term behavioral change.
Restrictive housing units are often populated with inmates who are prone to
behavioral infractions (i.e., misconduct) and institutional violence. Given their
propensity for rule violations, many are sent to segregation where some will
continue to exhibit antisocial behaviors. The practical reality is that some inmates
can be very difficult to manage, and often test the patience of front-line staff
members (Gendreau & Thériault, 2011). Nevertheless, advocates of offender
rehabilitation underscore that, “… the guiding principle of any human service
enterprise is to treat individuals with respect and humane care” (Gendreau &
Thériault, 2011, p. 7). It is therefore crucial for the front-line staff members who
work in restrictive housing units to be skilled in core correctional practices
related to effective service delivery with offender populations. Ideally, corrections
professionals would be equipped to deliver structured interventions to modify
target behaviors in a relatively short time. These front-line staff members could
also take advantage of teachable moments to reinforce and extend inmates’
coping skills and prosocial behaviors. For this reason, it is advisable for prison
authorities to assign dedicated staff (e.g., correctional officers, case managers,
clinicians, supervisors, and others) to restrictive housing units to ensure better
communication, consistency, stability, and on-site supervision for implementing
services and interventions.
Finally, cross training should be provided in mental health, substance abuse, and
criminogenic needs for correctional officers working in segregation. Corrections
professionals who can balance the dual roles of security and rehabilitation
are particularly valuable in working with offenders (see Skeem, Eno, Louden,
Polaschek, & Camp, 2007; Soderstrom, 2007). At an even more basic level, all
correctional staff should receive formal training on restrictive housing policies.
Furthermore, data on compliance with these policies should be collected and
analyzed, and the information should be reflected on employee performance
evaluations, as appropriate (U.S. Department of Justice, 2016).

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 351

Recommendation 9: Develop an individualized treatment plan and
measure inmate progress.
Among the more common (and rational) recommendations made for
segregation settings are improving assessment protocols (Bottos, 2007; Gendreau
& Thériault, 2011; Human Rights Watch, 2000; Lanes, 2011; Wormith et al.,
1988) and adding more rehabilitative services (Arrigo & Bullock, 2008; Lovell,
2008; O’Keefe, 2008; Scott & Gendreau, 1969). Although the field does not have
established criteria for “best practices” in segregation, scholars and correctional
officials have started to contemplate these issues in an attempt to design and
implement evidence-based services. This work must involve close partnerships
between practitioners and researchers.
In addition to the obvious involvement of psychologists in intervention activities
in segregation, these professionals could also provide useful contributions to
several of the research/program evaluations noted in other recommendations
(Gendreau & Thériault, 2011). One area where this is particularly relevant
involves the application of applied behavior analysis (ABA) within segregation,
which involves applying learning theories and behavioral interventions to change
specific target behaviors (Cooper, Heron, & Heward, 2007). Specifically, it is
an individualized approach to behavior modification that requires identifying
individual target behaviors, maintaining conditions to elicit those specific
behaviors, developing a schedule of reinforcement and punishment, and
then following that schedule to elicit the desired changes.8 The North Dakota
Department of Corrections and Rehabilitation recently initiated a pilot ABA
program in one restrictive housing unit to provide a focused, structured
treatment program to target discrete problem behaviors under the close
supervision of licensed psychologists. Although preliminary, the initial results are
very promising (personal communication, K. Wolfer, November 20, 2015). A more
comprehensive empirical evaluation of the program is currently in progress.
In addition to using ABA (particularly in the early stages of segregation),
corrections officials are advised to consider a combination of treatment strategies to
target the criminogenic needs of offenders by using evidence-based strategies, and
particularly for those dynamic factors that are linked to problems with institutional
adjustment. The RNR framework provides clear guidance about the approaches
most likely to be beneficial in this regard: (1) radical behavioral approaches
that are based on the principles of classical and operant conditioning; (2) social
learning approaches that involve modeling and behavioral rehearsal techniques
that engender self-efficacy; and (3) cognitive approaches that include cognitive
skills training, problem-solving therapy, self-control procedures, self-instructional
training, and stress inoculation training (Gendreau & Thériault, 2011).
8

The term maintaining conditions refers to the specific antecedents and consequences that cause a person to
perform a behavior (see Spiegler & Guevremont, 2010, for a detailed discussion).

352 • National Institute of Justice | NIJ.gov

In terms of radical behavioral approaches (including contingency management),
correctional agencies should develop a range of appropriate reinforcers that
include, at a minimum, tangible, token, and social reinforcers and activities (see
Spiegler & Guevremont, 2010).9 It is also advisable for correctional programs to
identify reinforcers that are most meaningful to offenders by surveying inmates.
Furthermore, prisons should develop a detailed written protocol to ensure
that reinforcers and punishers are administered consistently and immediately.
Similarly, it is important to identify a range of consequences such as fines, loss of
tokens or points, time out from generalized reinforcers, and social disapproval
(for a detailed review, see Spiegler & Guevremont, 2010). Finally, front-line
staff members should assess whether the punishment produces any negative
effects after administration, including emotional reactions (e.g., anxiety, anger),
withdrawal or avoidance behaviors, and perpetuation effects (i.e., when an
inmate learns to use punishment to control others’ behaviors). Punishment
should never interfere with new learning, lead to response substitution, or
disrupt social relationships. Although research has suggested that the number
of reinforcers should far outweigh the number of punishers, it is very common
for most correctional programs to spend more time and effort on developing
protocols related to sanctions. It is important to emphasize that the use of
reinforcers and sanctions applies to both the general population and within the
context of segregation.
Group interventions can also be used to teach specific skills to small groups
of inmates in segregation. In most cases, this treatment involves some form
of special restraint apparatus to limit physical interactions between inmates.
Adaptations of structured curricula have also been used in several jurisdictions
(e.g., Washington State Department of Corrections) to expose inmates to
treatment concepts prior to increased congregate time and transition back into
the general population of offenders. Another structured curriculum for use in
this specific manner is being developed by researchers in collaboration with
practitioners in the field (e.g., Smith, 2016b).
Ideally, structured interventions are combined with treatment packages
to address the individual needs of inmates. The interventions also may be
organized into a phase or level system in which inmates can progressively earn
privileges and advance through treatment by demonstrating desired behaviors.
For example, two correctional institutions operated by the Washington State
Department of Corrections have developed a contingency management system
for inmates in segregation to encourage prosocial behaviors in preparation
for re-entry into the general population of inmates. The phases are then tied
to the curriculum content and the types of reinforcers that inmates can earn.
9

A simple definition of reinforcement is the application of a stimulus to increase the likelihood that a behavior will
occur again. Reinforcers can take several forms, including tangible items and intangible incentives (e.g., activities
or special privileges). The reader is referred to Spiegler and Guevremont (2010) for a comprehensive discussion
of the selection and application of reinforcers.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 353

Preliminary results provide evidence that the contingency management system is
an important component of motivating offenders to learn the skills they need for
successful adjustment in the general population.

Recommendation 10: Implement aftercare and re-entry-focused
services to improve outcomes for inmates post-release.
Inmates released directly from segregation into the community have a higher risk
for recidivism compared to those released from the general population (Lovell
et al., 2007). For this reason, policies and practices should be developed to
gradually introduce segregated offenders back into the general population before
they are released from custody. This finding is consistent with other research
in the area of corrections that has underscored the importance of re-entry­
focused services and phase systems that transition offenders from higher levels
of supervision and structure to lower levels based on demonstrated progress.
Developing, implementing, and evaluating such programs in restrictive housing
settings specifically should be a priority for the field of corrections to improve
outcomes for inmates.
Meta-analyses of the treatment literature have found that institutional programs
consistently produce smaller effect sizes in comparison with community-based
programs (see Andrews & Bonta, 2010). This is attributable, at least in part, to
the fact that community-based programs have the distinct advantage of offering
interventions in vivo, that is, in more naturalistic environments where offenders
can immediately practice new skills. Prisons, however, are by definition artificial
environments where inmates have more limited opportunities to use skills in
their own high-risk situations. Relapse-prevention plans are useful as offenders
consider how they might immediately apply these skills within the prison context
(e.g., interactions with staff members and other inmates) as well as post-release.
To be successfully discharged from the institution, therefore, offenders should
meet clearly defined completion criteria and be trained to observe and manage
problem situations. The rehearsal of alternative, prosocial behaviors should
include initial practice in a safe environment (e.g., treatment group session)
using relatively simple scenarios. Eventually, offenders should practice their
newly acquired skills in increasingly difficult situations (e.g., in the housing
unit with peers). When clients demonstrate a new behavior, their improved
competency should be rewarded to encourage them to exhibit the response
again. Participating in aftercare and booster sessions can also improve treatment
outcomes. For all of these reasons, it is critical for inmates who are nearing the
end of their sentence and are still in segregation to be re-socialized, whenever
possible, into the general prison population prior to release into the community.
Pizarro et al. (2014) followed a sample of inmates released from supermax
institutions to compare the characteristics of offenders who were successful with
offenders who recidivated. The results suggest that successful inmates were more

35 4 • National Institute of Justice | NIJ.gov

likely to have been released on parole, and were more likely to have completed
behavioral and psychological treatment while incarcerated (Pizarro et al., 2014).
In North Dakota, inmates in segregation are moved into a transition unit where
they participate in daily activities with the general population of offenders, but
then return to the secure unit in the evening. Special group sessions are held
to discuss and troubleshoot risky situations that might have occurred during
the day. This allows inmates a more gradual transition back into the general
population. After inmates have been released from segregation, they are placed on
a specialized caseload for a period to ensure that adequate services and resources
are available to them in the general prison population. Such efforts are expected to
reduce subsequent placements in segregation. Data collection is ongoing, but the
preliminary results suggest that there is remarkable, measurable change in target
behaviors (personal communication, K. Wolfer, November 20, 2015).

Conclusion
The use of segregation remains a controversial issue in prison management
literature. It is common practice in prisons nationwide, but significant gaps in the
empirical literature remain. Future research should further investigate the effects
of segregation to ensure that correctional institutions are managed safely and
humanely. Furthermore, the implementation of treatment programs within the
context of restrictive housing units can be an important component of efforts to
reduce institutional misconduct and enhance post-release behavioral outcomes.
Many of the strategies discussed in the recommendations of this white paper
attempt to reconfigure restrictive housing units into placement options that can
support the goals of offender rehabilitation.

References
Andersen, H. S., Sestoft, D., Lillebaek, T., Gabrielsen, G., & Hemmingsen, R.
(2003). A longitudinal study of prisoners on remand: Repeated measures of
psychopathology in the initial phase of solitary versus nonsolitary confinement.
International Journal of Law and Psychiatry, 26(2), 165-177.
Andrews, D. A. (1979). The dimensions of correctional counseling and
supervision process in probation and parole. Toronto, Ontario, Canada: Ontario
Ministry of Correctional Services.
Andrews, D. A. (1980). Some experimental investigations of the principles of
differential association through deliberate manipulations of the structure of
service systems. American Sociological Review, 45(3), 448-462.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 355

Andrews, D. A., & Bonta, J. (2010). The psychology of criminal conduct (5th ed.).
New Providence, NJ: Anderson/LexisNexis.
Andrews, D. A., Dowden, C., & Rettinger, J. (2001). Special populations within
corrections. In J. A. Winterdyk (Ed.), Corrections in Canada: Social reactions to
crime (pp. 170-212). Toronto, Ontario, Canada: Prentice Hall.
Andrews, D. A., & Kiessling, J. J. (1980). Program structure and effective correctional
practices: A summary of the CaVIC research. In R. R. Ross & P. Gendreau
(Eds.), Effective correctional treatment (pp. 439-463). Toronto, Ontario, Canada:
Butterworths.
Andrews, D. A., Zinger, I., Hoge, R. D., Bonta, J., Gendreau, P., & Cullen,
F. T. (1990). Does correctional treatment work? A clinically-relevant and
psychologically informed meta-analysis. Criminology, 28(3), 369-404.
Aos, S., Lieb, R., Mayfield, J., Miller, M., & Pennucci, A. (2004). Benefits and
costs of prevention and early intervention programs for youth. Olympia, WA:
Washington State Institute for Public Policy.
Arbour, L. (1996). Commission of Inquiry into certain events at the Prison for
Women in Kingston. Ottawa, Ontario, Canada: Public Works and Government
Services of Canada.
Arrigo, B. A., & Bullock, J. L. (2008). The psychological effects of solitary
confinement on prisoners in supermax units: Reviewing what we know and
recommending what should change. International Journal of Offender Therapy
and Comparative Criminology, 52(6), 622-640.
Batastini, A. B. (2015). Comparing group psychotherapy via telepsychology and
in-person service modalities for inmates in long-term segregation (unpublished
doctoral dissertation). Lubbock, TX: Texas Tech University.
Bexton, W. H., Heron, W., & Scott, T. H. (1954). Effects of decreased variation in
the sensory environment. Canadian Journal of Psychology, 8(2), 70-76.
Bonta, J., & Gendreau, P. (1990). Reexamining the cruel and unusual punishment
of prison life. Law and Human Behavior, 14(4), 347-372.
Bonta, J., & Gendreau, P. (1995). Reexamining the cruel and unusual punishment
of prison life. In T.J. Flanagan (Ed.), Long-Term Imprisonment: Policy, Science,
and Correctional Practice (pp. 75-94). Thousand Oaks, CA: Sage.
Bonta, J., Law, M., & Hanson, K. (1998). The prediction of criminal and violent
recidivism among mentally disordered offenders: A meta-analysis. Psychological
Bulletin, 123(2), 123-142.

356 • National Institute of Justice | NIJ.gov

Bottos, S. (2007). Profile of offenders in administrative segregation: A review of
the literature (research report No. B-39). Ottawa, Ontario, Canada: Correctional
Service of Canada.
Briggs, C. S., Sundt, J. L, & Castellano, T. C. (2003). The effect of super-maximum
security prisons on aggregate levels of institutional violence. Criminology, 41(4),
1341-1376.
Browne, A., Cambier, A., & Agha, S. (2011). Prisons within prisons: The use of
segregation in the United States. Federal Sentencing Reporter, 24(1), 46-49.
Bush, J., Bilodeau, B., & Kornick, M. (1995). Options: A cognitive change program.
Longmont, CO: National Institute of Corrections.
Butler, H. D., Steiner, B., Makarios, M., & Travis, L. (in press). Assessing the
effects of exposure to supermax confinement on offender post-release behaviors.
The Prison Journal.
Centen, G., & Sampson, E. (1991). Architecture, operations, and change. Forum
on Corrections, 3(2), 27-30.
Chen, M. K., & Shapiro, J. M. (2007). Do harsher prison conditions reduce
recidivism? A discontinuity-based approach. American Law and Economic
Review, 9(1), 1-29.
Cid, J. (2009). Is imprisonment criminogenic? A comparative study of recidivism
rates between prison and suspended prison sanctions. European Journal of
Criminology, 6(6), 459-480.
Cleland, C. M., Pearson, F. S., Lipton, D. S., & Yee, D. (1997). Does age make a
difference? A meta-analytic approach to reductions in criminal offending for
juveniles and adults. Paper presented at the Annual Meeting of the American
Society of Criminology, San Diego, CA.
Cohen, M. A. (1998). The monetary value of saving a high-risk youth. Journal of
Quantitative Criminology, 14(1), 5-33.
Cooper, J. O., Heron, T. E., & Heward, W. L. (2007). Applied behavior analysis.
Columbus, OH: Merrill Publishers.
Cullen, F. T., & Gendreau, P. (2000). Assessing correctional rehabilitation:
Policy, practice, and prospects. In J. Horney (Ed.), Criminal Justice 2000, volume
3: Policies, processes, and decisions of the criminal justice system (pp. 109-175).
Washington, DC: National Institute of Justice.
Cullen, F. T., Gendreau, P., Jarjoura, G. R., & Wright, J. P. (1997). Crime and the
bell curve: Lessons from intelligent criminology. Crime & Delinquency, 43(4),
387-411.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 357

Cullen, F. T., & Gilbert, K. E. (1982). Reaffirming rehabilitation. Cincinnati,
OH: Anderson.
Davidson, W. S., III, Gottschalk, R., Gensheimer, L., & Mayer, J. (1984). Interventions
with juvenile delinquents: A meta-analysis of treatment efficacy. Washington, DC:
National Institute of Juvenile Justice and Delinquency Prevention.
Dowden, C., & Andrews, D. A. (2000). Effective correctional treatment and
violent reoffending: A meta-analysis. Canadian Journal of Criminology, 42(4),
449-467.
Drago, F., Galbiati, R., & Vertova, P. (2008). Prison conditions and recidivism (IZA
discussion paper no. 3395). Bonn, Germany: IZA Institute for the Study of Labor.
Ecclestone, C. E., Gendreau, P., & Knox, C. (1974). Solitary confinement of
prisoners: An assessment of its effects on inmates’ personal constructs and
adrenocortical activity. Canadian Journal of Behavioural Science, 6(2), 178-191.
Eilperin, J. (2016, January 26). Obama bans solitary confinement for juveniles in
federal prisons. The Washington Post. Available at http://www.washingtonpost.com.
French, S. A., & Gendreau, P. (2006). Reducing prison misconducts: What works!
Criminal Justice and Behavior, 33(2), 185-218.
Frost, N. A., & Monteiro, C. E. (2016). Administrative segregation in prison.
Washington, DC: U.S. Department of Justice.
Gaes, G. G., & Camp, S. D. (2009). Unintended consequences: Experimental
evidence for the criminogenic effect of prison security level placement on postrelease recidivism. Journal of Experimental Criminology, 5(2), 139-162.
Garrett, C. J. (1985). Effects of residential treatment of adjudicated delinquents:
A meta-analysis. Journal of Research in Crime and Delinquency, 22(4), 287-308.
Gendreau, P. (1996). Offender rehabilitation: What we know and what needs to
be done. Criminal Justice and Behavior, 23(1), 144-161.
Gendreau, P., Andrews, D., & Thériault, Y. L. (2010). Correctional program
assessment inventory-2010. Saint John, New Brunswick, Canada: University of
New Brunswick.
Gendreau, P., & Bonta, J. L. (1984). Solitary confinement is not cruel and unusual
punishment: Sometimes people are! Canadian Journal of Criminology, 26(4),
467-478.
Gendreau, P., Freedman, N., Wilde, G. J. S., & Scott, G. D. (1968). Stimulation
seeking after seven days of perceptual deprivation. Perceptual and Motor Skills,
26(2), 547-550.

358 • National Institute of Justice | NIJ.gov

Gendreau, P., Freedman, N. L., Wilde, G. J., & Scott, G. D. (1972). Changes in
EEG alpha frequency and evoked response latency during solitary confinement.
Journal of Abnormal Psychology, 79(1), 54-59.
Gendreau, P., Goggin, C., Cullen, F. T., & Andrews, D. A. (2000). Effects of
community sanctions and incarceration on recidivism. Forum on Corrections,
12(2), 10-13.
Gendreau, P., Goggin, C., & Smith, P. (2001). Obstacles to effective correctional
program delivery. In L. L. Motiuk & R. C. Serin (Eds.), Compendium 2000
on effective correctional programs (pp. 37-46). Ottawa, Ontario, Canada:
Correctional Service of Canada, Solicitor General of Canada.
Gendreau, P., & Keyes, D. (2001). Making prisons safer and more humane
environments. Canadian Journal of Criminology, 43(1), 123-130.
Gendreau, P., & Labrecque, R. M. (in press). The effects of administrative
segregation: A lesson in knowledge cumulation. In J. Wooldredge & P. Smith
(Eds.), Oxford handbook on prisons and imprisonment. Oxford, England: Oxford
University Press.
Gendreau, P., Little, T., & Goggin, C. (1996). A meta-analysis of the predictors of
adult offender recidivism: What works! Criminology, 34(4), 575-608.
Gendreau, P., & Ross, B. (1979). Effective correctional treatment: Bibliotherapy
for cynics. Crime & Delinquency, 25(4), 463-489.
Gendreau, P., & Ross, R. R. (1987). Revivification of rehabilitation: Evidence
from the 1980s. Justice Quarterly, 4(3), 349-407.
Gendreau, P., & Smith, P. (2007). Influencing the “people who count”: Some
perspectives on the reporting of meta-analytic results for prediction and treatment
outcomes with offenders. Criminal Justice and Behavior, 34(12), 1536-1559.
Gendreau, P., & Thériault, Y. (2011). Bibliotherapy for cynics revisited:
Commentary on one year longitudinal study of the psychological effects of
administrative segregation. Washington, DC: National Institute of Corrections.
Glaze, L. E., & Kaeble, D. (2014). Correctional populations in the United States,
2013. NCJ 248479. Available at http://www.bjs.gov.
Goldstein, A. P. (1986). Psychological skill training and the aggressive adolescent.
In S. Apter & A. Goldstein (Eds.), Youth violence: Program and prospects (pp. 89­
119). New York, NY: Pergamon Press.
Grassian, S. (1983). Psychopathological effects of solitary confinement. American
Journal of Psychiatry, 140(11), 1450-1454.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 359

Grassian, S., & Friedman, N. (1986). Effects of sensory deprivation in psychiatric
seclusion and solitary confinement. International Journal of Law and Psychiatry,
8(1), 49-65.
Haney, C. (2003). Mental health issues in long-term solitary and “supermax”
confinement. Crime and Delinquency, 49(1), 124-156.
Haney, C. (2008). A culture of harm: Taming the dynamics of cruelty in
supermax prisons. Criminal Justice and Behavior, 35(8), 956-984.
Haney, C. (2009). The social psychology of isolation: Why solitary confinement is
psychologically harmful. Prison Service Journal, 181(1), 12-20.
Hanson, R. K., & Morton-Bourgon, K. E. (2004). Predictors of sexual recidivism:
An updated meta-analysis (report no. 2004-02). Ottawa, Ontario, Canada: Public
Safety and Emergency Preparedness of Canada.
Helmus, L. M. (2015). Developing and validating a risk assessment scale to
predict inmate placements in administrative segregation in the Correctional
Service of Canada (unpublished doctoral dissertation). Ottawa, Ontario, Canada:
Carleton University.
Huebner, B. M. (2003). Administrative determinants of inmate violence: A
multilevel analysis. Journal of Criminal Justice, 31(2), 107-117.
Hughes, T., Wilson, D., & Beck, A. (2001). Trends in state parole, 1990-2000. NCJ
184735. Washington, DC: Bureau of Justice Statistics.
Human Rights Watch. (2000). Out of sight: Supermaximum security confinement
in the United States. New York, NY: Human Rights Watch.
Hunt, M. (1997). How science takes stock: The story of meta-analysis. New York,
NY: Russell Sage Foundation.
Hunt, H. T., & Chefurka, C. M. (1976). A test of the psychedelic model of altered
states of consciousness: The role of introspective sensitization in eliciting unusual
subjective reports. Archives of General Psychiatry, 33(7), 867-876.
Hunter, J. E., & Schmidt, F. L. (1996). Cumulative research knowledge and social
policy formulation: The critical role of meta-analysis. Psychology, Public Policy,
and Law, 2(2), 324-347.
Izzo, R. L., & Ross, R. R. (1990). Meta-analysis of rehabilitation programs for
juvenile delinquents. Criminal Justice and Behavior, 17(1), 134-142.
Jackson, M. (2001). The psychological effects of administrative segregation.
Canadian Journal of Criminology, 43(1), 109-116.

360 • National Institute of Justice | NIJ.gov

Jackson, M. (2002). Justice behind walls. Vancouver, British Columbia, Canada:
Douglas & McIntyre.
Jonson, C. L. (2010). The impact of imprisonment on reoffending: A meta-analysis
(unpublished doctoral dissertation). Cincinnati, OH: University of Cincinnati.
Labrecque, R. M. (2015). The effect of solitary confinement on institutional misconduct:
A longitudinal evaluation. Washington, DC: U.S. Department of Justice.
Labrecque, R. M., & Smith, P. (2013). Advancing the study of solitary confinement.
In J. Fuhrmann & S. Baier (Eds.), Prisons and prison systems: Practices, types, and
challenges (pp. 57-70). Hauppage, NY: Nova Science Publishers.
Lanes, E. C. (2011). Are the “worst of the worst” self-injurious prisoners more
likely to end up in long-term maximum-security administrative segregation?
International Journal of Offender Therapy and Comparative Criminology, 55(7),
1034-1050.
Langan, P. A., & Levin, D. J. (2002). Recidivism of prisoners released in 1994.
NCJ 193427. Washington, DC: Bureau of Justice Statistics.
Latessa, E. J., Cullen, F. T., & Gendreau, P. (2002). Beyond correctional quackery:
Professionalism and the possibility of effective treatment. Federal Probation,
66(2).
Lipsey, M. W. (1992). Juvenile delinquency treatment: A meta-analytic inquiry
into the variability of effects. In T. D. Cook, H. Cooper, D. S. Cordray, H.
Hartmann, L. V. Hedges, R. J. Light, T. A. Louis, & F. Mostseller (Eds.), Metaanalysis for explanation: A casebook (pp. 83-127). New York, NY: Russell Sage.
Lipsey, M. W. (1999). Can rehabilitative programs reduce the recidivism of
juvenile offenders? An inquiry into the effectiveness of practical programs.
Virginia Journal of Social Policy and Law, 6(3), 611-641.
Lipsey, M. W., Chapman, G. L., & Landenberger, N. A. (2001). Cognitivebehavioral programs for offenders. The Annals of the American Academy of
Political and Social Science, 578(1), 144-157.
Lipsey, M. W., &, Wilson, D. B. (1993). The efficacy of psychological, educational
and behavioral treatment. American Psychologist, 48(12), 1181-1209.
Listwan, S. J., Sullivan, C. J., Agnew, R., Cullen, F. T., & Colvin, M. (2013). The
pains of imprisonment revisited: The impact of strain on inmate recidivism.
Justice Quarterly, 30(1), 144-168.
Lösel, F. (1995). The efficacy of correctional treatment: A review and synthesis of
meta-evaluations. In J. McGuire (Ed.), What works: Reducing reoffending
(pp. 79-111). Oxford, England: John Wiley & Sons.

Restrictive Housing in the U.S.: Issues, Challenges, and Future Directions • 361

Lovell, D. (2008). Patterns of disturbed behavior in a supermax population.
Criminal Justice and Behavior, 35(8), 985-1004.
Lovell, D., Johnson, L. C., & Cain, K. C. (2007). Recidivism of supermax
prisoners in Washington State. Crime & Delinquency, 53(4), 633-656.
Lowenkamp, C. T., Latessa, E. J., & Smith, P. (2006). Does correctional program
quality really matter? The impact of adhering to the principles of effective
intervention. Criminology and Public Policy, 5(3), 575-594.
McGuire, J. (2004). Understanding psychology and crime: Perspectives on theory
and action. Maidenhead, Berkshire, England: University Press.
Mears, D. P., & Bales, W. D. (2009). Supermax incarceration and recidivism.
Criminology, 47(4), 1131-1166.
Mears, D. P., & Castro, J. L. (2006). Wardens’ views on the wisdom of supermax
prisons. Crime and Delinquency, 52(3), 398-431.
Mears, P., & Watson, J. (2006). Towards a fair and balanced assessment of
supermax prisons. Justice Quarterly, 23(2), 232-270.
Morgan, R. D., Van Horn, S. A., MacLean, N., Bolanos, A., Gray, A. L., Batastini,
A., & Mills, J. F. (2014). Administrative segregation: Is it a harmful correctional
practice? Presented at the annual meeting of the American Psychology-Law
Society in New Orleans, LA.
Morris, R. G. (2015). Exploring the effect of exposure to short-term solitary
confinement among violent prison inmates. Journal of Quantitative Criminology,
32(1), 1-22.
Motiuk, L. L. (1993). Where are we in our ability to assess risk? Forum on
Corrections Research, 5(2), 14-18.
Nafekh, M., & Motiuk, L. (2002). The statistical information on recidivism –
revised 1 (SIR-R1) scale: A psychometric examination (research report no. R-126).
Ottawa, Ontario, Canada: Correctional Service of Canada.
Nagin, D. S., Cullen, F. T., & Jonson, C. L. (2009). Imprisonment and reoffending.
In M. Tonry (Ed.), Crime and justice: A review of research, vol. 38 (pp. 115-200).
Chicago, IL: University of Chicago Press.
Nieuwbeerta, P., Nagin, D. S., & Blokland, A. A. (2009). Assessing the impact of
first-time imprisonment on offenders’ subsequent criminal career development: A
matched samples comparison. Journal of Quantitative Criminology, 25(3), 227-257.
O’Keefe, M. L. (2008). Administrative segregation from within: A corrections
perspective. The Prison Journal, 88(1), 123-143.

362 • National Institute of Justice | NIJ.gov

O’Keefe, M. L., Klebe, K. J., Stucker, A., Sturm, K., & Leggett, W. (2010). One
year longitudinal study of the psychological effects of administrative segregation.
Colorado Springs, CA: Colorado Departm