Skip navigation

Identifying the Culprit - Assessing Eyewitness Identification, National Academy of Sciences, 2014

Download original document:
Brief thumbnail
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
This PDF is available from The National Academies Press at http://www.nap.edu/catalog.php?record_id=18891

Identifying the Culprit: Assessing Eyewitness Identification

ISBN
978-0-309-31059-8
160 pages
6x9
PAPERBACK (2014)

Committee on Scientific Approaches to Understanding and Maximizing the
Validity and Reliability of Eyewitness Identification in Law Enforcement and
the Courts; Committee on Science, Technology, and Law; Policy and
Global Affairs; Committee on Law and Justice; Division of Behavioral and
Social Sciences and Education; National Research Council

Visit the National Academies Press online and register for...
Instant access to free PDF downloads of titles from the
NATIONAL ACADEMY OF SCIENCES
NATIONAL ACADEMY OF ENGINEERING
INSTITUTE OF MEDICINE
NATIONAL RESEARCH COUNCIL

10% off print titles
Custom notification of new releases in your field of interest
Special offers and discounts

Distribution, posting, or copying of this PDF is strictly prohibited without written permission of the National Academies Press.
Unless otherwise indicated, all materials in this PDF are copyrighted by the National Academy of Sciences.
Request reprint permission for this book

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

Identifying the Culprit
Assessing Eyewitness Identification

Committee on Scientific Approaches to Understanding and
Maximizing the Validity and Reliability of Eyewitness
Identification in Law Enforcement and the Courts
Committee on Science, Technology, and Law
Policy and Global Affairs
Committee on Law and Justice
Division of Behavioral and Social Sciences and Education

THE NATIONAL ACADEMIES PRESS
Washington, D.C.
www.nap.edu

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
THE NATIONAL ACADEMIES PRESS 500 Fifth Street, NW Washington, DC 20001

NOTICE: The project that is the subject of this report was approved by the Governing Board of the National
Research Council, whose members are drawn from the councils of the National Academy of Sciences, the National
Academy of Engineering, and the Institute of Medicine. The members of the committee responsible for the report
were chosen for their special competences and with regard for appropriate balance.
This study was funded by a grant between the National Academy of Sciences and the Laura and John Arnold
Foundation. Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the
author and do not necessarily reflect the views of the organization that provided support for the project.
International Standard Book Number 13: XXX- X-XXX-XXXX-X
International Standard Book Number 10: X-XXX-XXXXX-X
Additional copies of this report are available from the National Academies Press, 500 Fifth Street, NW, Room 360,
Washington, DC 20001; (800) 624-6242 or (202) 334-3313; http://www.nap.edu.
Copyright 2014 by the National Academy of Sciences. All rights reserved. Printed in the United States of America

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

The National Academy of Sciences is a private, nonprofit, self-perpetuating society of distinguished scholars engaged in
scientific and engineering research, dedicated to the furtherance of science and technology and to their use for the general
welfare. Upon the authority of the charter granted to it by the Congress in 1863, the Academy has a mandate that requires it to
advise the federal government on scientific and technical matters. Dr. Ralph J. Cicerone is president of the National Academy of
Sciences.
The National Academy of Engineering was established in 1964, under the charter of the National Academy of Sciences, as a
parallel organization of outstanding engineers. It is autonomous in its administration and in the selection of its members,
sharing with the National Academy of Sciences the responsibility for advising the federal government. The National
Academy of Engineering also sponsors engineering programs aimed at meeting national needs, encourages education and
research, and recognizes the superior achievements of engineers. Dr. C. D. Mote, Jr., is president of the National Academy of
Engineering.
The Institute of Medicine was established in 1970 by the National Academy of Sciences to secure the services of eminent
members of appropriate professions in the examination of policy matters pertaining to the health of the public. The Institute acts
under the responsibility given to the National Academy of Sciences by its congressional charter to be an adviser to the federal
government and, upon its own initiative, to identify issues of medical care, research, and education. Dr. Victor J. Dzau is
president of the Institute of Medicine.
The National Research Council was organized by the National Academy of Sciences in 1916 to associate the broad community
of science and technology with the Academy’s purposes of furthering knowledge and advising the federal government.
Functioning in accordance with general policies determined by the Academy, the Council has become the principal operating
agency of both the National Academy of Sciences and the National Academy of Engineering in providing services to the
government, the public, and the scientific and engineering communities. The Council is administered jointly by both Academies
and the Institute of Medicine. Dr. Ralph J. Cicerone and Dr. C. D. Mote, Jr., are chair and vice chair, respectively, of the National
Research Council.
www.national-academies.org

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
COMMITTEE ON THE SCIENTIFIC APPROACHES TO UNDERSTANDING AND
MAXIMIZING THE VALIDITY OF EYEWITNESS IDENTIFICATION IN LAW
ENFORCEMENT AND THE COURTS
Co-Chairs
THOMAS D. ALBRIGHT (NAS), Professor and Director, Vision Center Laboratory and
Conrad T. Prebys Chair in Vision Research, Salk Institute for Biological Studies
JED S. RAKOFF, Senior Judge, United States District Court for the Southern District of New York
Members
WILLIAM G. BROOKS III, Chief of Police, Norwood (MA) Police Department
JOE S. CECIL, Project Director, Division of Research, Federal Judicial Center
WINRICH FREIWALD, Assistant Professor, Laboratory of Neural Systems, The Rockefeller University
BRANDON L. GARRETT, Roy L. and Rosamond Woodruff Morgan Professor of Law, University
of Virginia Law School
KAREN KAFADAR, Commonwealth Professor and Chair of Statistics, University of Virginia
A.J. KRAMER, Federal Public Defender for the District of Columbia
SCOTT McNAMARA, Oneida County (NY) District Attorney
CHARLES ALEXANDER MORGAN III, Associate Clinical Professor of Psychiatry, Yale University
School of Medicine
ELIZABETH A. PHELPS, Silver Professor of Psychology and Neural Science, New York University
DANIEL J. SIMONS, Professor, Department of Psychology, University of Illinois
ANTHONY D. WAGNER, Professor of Psychology and Neuroscience and Co-Director, Center for
Cognitive and Neurobiological Imaging, Stanford University; Director, Stanford Memory Laboratory
JOANNE YAFFE, Professor of Social Work and Adjunct Professor of Psychiatry, University of Utah
Staff
ANNE-MARIE MAZZA, Study Director and Director, Committee on Science, Technology, and Law,
National Academy of Sciences
ARLENE F. LEE, Director, Committee on Law and Justice, National Academy of Sciences
STEVEN KENDALL, Program Officer, Committee on Science, Technology, and Law, National
Academy of Sciences
KAROLINA KONARZEWSKA, Program Assistant, Committee on Science, Technology, and Law,
National Academy of Sciences
ANJALI SHASTRI, Christine Mirzayan Science and Technology Policy Graduate Fellow
SARAH WYNN, Christine Mirzayan Science and Technology Policy Graduate Fellow

v
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
COMMITTEE ON SCIENCE, TECHNOLOGY, AND LAW
Co-Chairs
DAVID BALTIMORE (NAS/IOM), President Emeritus and Robert Andrews Millikan Professor
of Biology, California Institute of Technology
DAVID S. TATEL, Judge, U.S. Court of Appeals for the District of Columbia Circuit
Members
THOMAS D. ALBRIGHT (NAS), Professor and Director, Vision Center Laboratory and
Conrad T. Prebys Chair in Vision Research, Salk Institute for Biological Studies
ANN ARVIN (IOM), Lucile Packard Professor of Pediatrics and Microbiology and Immunology;
Vice Provost and Dean of Research, Stanford University
BARBARA E. BIERER, Professor of Medicine, Harvard Medical School
CLAUDE CANIZARES (NAS), Vice President and the Bruno Rossi Professor of Physics,
Massachusetts Institute of Technology
ARTURO CASADEVALL, Leo and Julia Forchheimer Professor of Microbiology and Immunology;
Chair, Department of Biology and Immunology; and Professor of Medicine, Albert Einstein
College of Medicine
JOE S. CECIL, Project Director, Program on Scientific and Technical Evidence, Division of
Research, Federal Judicial Center
R. ALTA CHARO (IOM), Warren P. Knowles Professor of Law and Bioethics, University of
Wisconsin at Madison
HARRY T. EDWARDS, Judge, U.S. Court of Appeals for the District of Columbia Circuit
DREW ENDY, Associate Professor, Bioengineering, Stanford University and President,
The BioBricks Foundation
MARCUS FELDMAN (NAS), Burnet C. and Mildred Wohlford Professor of Biological Sciences,
Stanford University
JEREMY FOGEL, Director, The Federal Judicial Center
HENRY T. GREELY, Deane F. and Kate Edelman Johnson Professor of Law and Professor, by
courtesy, of Genetics, Stanford University
MICHAEL GREENBERGER, Law School Professor and Director, Center for Health and
Homeland Security, University of Maryland
BENJAMIN W. HEINEMAN, JR., Senior Fellow, Harvard Law School and Harvard Kennedy
School of Government
MICHAEL IMPERIALE, Arthur F. Thurnau Professor of Microbiology and Immunology University
of Michigan
GREG KISOR, Chief Technologist of Global Licensing, Intellectual Ventures
GOODWIN LIU, Associate Justice, California Supreme Court
JENNIFER MNOOKIN, David G. Price and Dallas P. Price Professor of Law, University of
California, Los Angeles School of Law
R. GREGORY MORGAN, Vice President and General Counsel, Massachusetts Institute of Technology
ALAN B. MORRISON, Lerner Family Associate Dean for Public Interest and Public Service Law,
George Washington University Law School
CHERRY MURRAY (NAS/NAE), Dean, School of Engineering and Applied Sciences,
Harvard University
ROBERTA NESS (IOM), Dean and M. David Low Chair in Public Health, University of Texas
School of Public Health

vi
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
HARRIET RABB, Vice President and General Counsel, The Rockefeller University
DAVID RELMAN (IOM), Thomas C. and Joan M. Merigan Professor, Departments of Medicine,
and of Microbiology and Immunology, Stanford University and Chief, Infectious Disease Section,
VA Palo Alto Health Care System
RICHARD REVESZ, Lawrence King Professor of Law; Dean Emeritus; and Director, Institute for
Policy Integrity, New York University School of Law
MARTINE A. ROTHBLATT, Chairman and Chief Executive Officer, United Therapeutics
DAVID VLADECK, Professor and Co-Director, Institute for Public Representation, Georgetown
Law School
Staff
ANNE-MARIE MAZZA, Director
STEVEN KENDALL, Program Officer
KAROLINA KONARZEWSKA, Program Assistant

vii
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
COMMITTEE ON LAW AND JUSTICE
Chairs
JEREMY TRAVIS (Chair), President, John Jay College of Criminal Justice, The City University
of New York
RUTH D. PETERSON (Vice-Chair), Professor of Sociology and Director, Criminal Justice Research
Center, Ohio State University
Members
CARL C. BELL, Staff Psychiatrist - St. Bernard’s Hospital; Staff Psychiatrist - Jackson Park Hospital's
Outpatient Family Practice Clinic; Professor of Psychiatry and Public Health - University of Illinois
at Chicago
JOHN J. DONOHUE III, C. Wendell and Edith M. Carlsmith Professor of Law, Stanford Law School
MINDY FULLILOVE, Professor of Clinical Psychiatry and Professor of Clinical Sociomedical Sciences
and Co-Director, Community Research Group, NYSPI and Mailman SPH, Columbia University
MARK KLEIMAN, Professor of Public Policy, University of California, Los Angeles
GARY LAFREE, Director, National Consortium for the Study of Terrorism and Responses to
Terrorism (START) and Professor, Criminology and Criminal Justice University of Maryland
JANET L. LAURITSEN, Professor, Department of Criminology and Criminal Justice, University
of Missouri
GLENN C. LOURY, Merton P. Stoltz Professor of the Social Sciences, Department of Economics,
Brown University
JAMES P. LYNCH, Professor and Chair, Department of Criminology and Criminal Justice,
University of Maryland
CHARLES F. MANSKI (NAS), Board of Trustees Professor in Economics, Department of
Economics, Northwestern University
DANIEL S. NAGIN, Teresa and H. John Heinz III University, Professor of Public Policy and
Statistics, Carnegie Mellon University
ANNE MORRISON PIEHL, Associate Professor, Department of Economics and Program in
Criminal Justice, Rutgers University
DANIEL B. PRIETO, Director, Cybersecurity and Technology and Director, Defense Industrial
Base Cyber Security/Information Assurance, Office of the Secretary of Defense, CIO
SUSAN B. SORENSON, Professor, University of Pennsylvania
DAVID WEISBURD, Distinguished Professor, Department of Criminology, Law and Society
Director, Center for Evidence-Based Crime Policy George Mason University Walter E. Meyer
Professor of Law and Criminal Justice, Faculty of Law, The Hebrew University
CATHY SPATZ WIDOM, Distinguished Professor, Psychology Department, John Jay College
of Criminal Justice, The City University of New York
PAUL K. WORMELI, Executive Director, Integrated Justice Information Systems
Staff
ARLENE F. LEE, Director
EMILY BACKES, Research Associate
MALAY MAJMUNDAR, Senior Program Officer
STEVE REDBURN, Scholar
JULIE SCHUCK, Senior Program Associate
DANIEL TALMAGE, Program Officer
TINA M. LATIMER, Program Coordinator
viii
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

Acknowledgments
ACKNOWLEDGMENT OF PRESENTERS
The committee gratefully acknowledges the contributions of the following individuals:
Karen L. Amendola, Police Foundation; Steven E. Clark, University of California, Riverside; Rob
Davis, Police Executive Research Forum; Kenneth Deffenbacher, University of Nebraska at Omaha;
Paul DeMuniz, Oregon Supreme Court; Shari Seidman Diamond, Northwestern University and
American Bar Foundation; John Firman, International Association of Chiefs of Police; Ronald
Fisher, Florida International University; Geoffrey Gaulkin, Special Master, State v. Henderson
(NJ); Kristine Hamann, National District Attorney’s Association; Barbara Hervey, Texas Court of
Criminal Appeals; Robert J. Kane, Supreme Judicial Study Group on Eyewitness Identification
(MA); Saul Kassin, John Jay College of Criminal Justice; Peter Kilmartin, State of Rhode Island;
David LaBahn, Association of Prosecuting Attorneys; Elizabeth F. Loftus, University of California,
Irvine; Roy S. Malpass, University of Texas at El Paso; Sheri Mecklenburg, U.S. Department of
Justice; Christian A. Meissner, Iowa State University; John Monahan, University of Virginia; Steven
Penrod, John Jay College of Criminal Justice; P. Jonathon Phillips, National Institute of Standards
and Technology; Joseph Salemme, Chicago Police Department; Daniel L. Schacter, Harvard
University; Barry Scheck, The Innocence Project; Jessica Snowden, Federal Judicial Center; Nancy
Steblay, Augsburg College; Gary L. Wells, Iowa State University; John T. Wixted, University of
California, San Diego; David V. Yokum, University of Arizona.
ACKNOWLEDGMENT OF REVIEWERS
This report has been reviewed in draft form by individuals chosen for their diverse perspectives and
technical expertise, in accordance with procedures approved by the National Academies’ Report Review
Committee. The purpose of this independent review is to provide candid and critical comments that will
assist the institution in making its published report as sound as possible and to ensure that the report
meets institutional standards for objectivity, evidence, and responsiveness to the study charge. The review
comments and draft manuscript remain confidential to protect the integrity of the process.
We wish to thank the following individuals for their review of this report: Art Acevedo, Austin,
Texas Police Department; Aaron Benjamin, University of Illinois at Urbana-Champaign; Vicki Bruce,
Newcastle University; Jules Epstein, Widener University; Jeremy Fogel, The Federal Judicial Center;
Constantine Gatsonis, Brown University; Hank Greely, Stanford University; Peter Imrey, Cleveland
Clinic; Robert Kane, Massachusetts Supreme Court; Timothy Koller; Office of the Richmond County
District Attorney; Elizabeth Loftus, University of California, Irvine; Robert Masters, Office of the Queens
County District Attorney; Geoffrey Mearns, Northern Kentucky University; and Hal Stern, University of
California, Irvine.
Although the reviewers listed above have provided many constructive comments and suggestions,
they were not asked to endorse the conclusions or recommendations, nor did they see the final draft of the
report before its release. The review of this report was overseen by David Korn, Harvard Medical School
and Massachusetts General Hospital and Stephen E. Fienberg, Carnegie Mellon University. Appointed by
the National Academies, they were responsible for making certain that an independent examination of this
report was carried out in accordance with institutional procedures and that all review comments were
carefully considered. Responsibility for the final content of this report rests entirely with the authoring
committee and the institution.

ix
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

Contents
PREFACE ........................................................................................................................................................ xiii

SUMMARY ........................................................................................................................................................ 1
1

INTRODUCTION ..................................................................................................................................... 7

2

EYEWITNESS IDENTIFICATION PROCEDURES ..................................................................... 15

3

THE LEGAL FRAMEWORK FOR ASSESSMENT OF EYEWITNESS
IDENTIFICATION EVIDENCE ........................................................................................................ 21

4

BASIC RESEARCH ON VISION AND MEMORY ........................................................................ 31

5

APPLIED EYEWITNESS IDENTIFICATION RESEARCH ....................................................... 49

6

FINDINGS AND RECOMMENDATIONS ....................................................................................... 71
APPENDIXES

A

BIOGRAPHICAL INFORMATION OF COMMITTEE AND STAFF ............................................. 83

B

COMMITTEE MEETING AGENDAS .............................................................................................. 93

C

CONSIDERATION OF UNCERTAINTY IN DATA ON THE CONFIDENCEACCURACY RELATIONSHIP AND THE RECEIVER OPERATING
CHARACTERISTIC (ROC) CURVE ................................................................................................ 97
BOXES, FIGURES, AND TABLES

Boxes
Box 1-1 The Ronald Cotton Case, 8
Box 1-2 Charge to the Committee, 9
Box 2-1 Blinding, 18
Box 5-1 The Influences of Discriminability and Response Bias on Human Decisions, 56
Box 5-2 Analysis of Receiver Operating Characteristics (ROC), 58
Figures
Figure 1-1 Memory Accuracy and Time, 12
Figure 5-1 Receiver Operating Characteristic (ROC) Curve, 53
Figure C-1 Data Inferred from Juslin et al., 102
Figure C-2 Data from Brewer and Wells, 103
Figure C-3 Data from Experiment 1A in Mickes, Flowe, and Wixted, 103
Figure C-4 Data from Experiment 2 in Mickes, Flowe, and Wixted, 104
Tables
Table 5-1 Contingency Table for Possible Eyewitness Identification Outcomes, 53
Table C-1 Conditions and Logarithms of Reported pAUC Values, 106
Table C-2 Analysis of Variance Table for log(pAUC), 106
xi
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

Preface
Eyewitness identifications play an important role in the investigation and prosecution of crimes, but
they have also led to erroneous convictions. In the fall of 2013, the Laura and John Arnold Foundation
called upon the National Academy of Sciences (NAS) to assess the state of research on eyewitness
identification and, when appropriate, make recommendations. In response to this request, the NAS
appointed an ad hoc study committee that we have been privileged to co-chair.
The committee’s review analyzed relevant published and unpublished research, external
submissions, and presentations made by various experts and interested parties. The research examined fell
into two general categories: (1) basic research on vision and memory and (2) applied research directed at
the specific problem of eyewitness identification. Basic research has progressed for many decades, is of
high quality, and is largely definitive. Research of this category identifies principled and insurmountable
limits of vision and memory that inevitably affect eyewitness accounts, bear on conclusions regarding
accuracy, and provide a broad foundation for the committee’s recommendations.
Through its review, the committee came to recognize that applied eyewitness identification research
has identified key variables affecting the accuracy of eyewitness identifications. This research has been
instrumental in informing law enforcement, the bar, and the judiciary of the frailties of eyewitness
identification testimony. Such past research has appropriately identified the variables that may affect an
individual’s ability to make an accurate identification. However, given the complexity of the problem,
the practical difficulties it poses for experimental research, and the still ongoing evolution of statistical
procedures in the field of eyewitness identification research, there remains at the time of this review
substantial uncertainty about the effect and the interplay of these variables on eyewitness identification.
Nonetheless, a range of practices has been validated by scientific methods and research and represents a
starting place for efforts to improve eyewitness identification procedures.
In this report, the committee offers recommendations on how law enforcement and the courts may
increase the accuracy and utility of eyewitness identifications. In addition, the committee identifies areas
for future research and for collaboration between the scientific and law enforcement communities.
We are indebted to those who addressed the committee and to those who submitted materials to the
committee; and we are particularly indebted to the members of the committee. These individuals devoted
untold hours to their reviews of materials, meetings, conference calls, analyses, and report writing. This
report is very much the result of the enormous contributions of an engaged community of scholars and
practitioners who reached their findings and recommendations after many vigorous and thoughtful
discussions. We also would like to thank the project staff, Karolina Konarzewska, Steven Kendall, Arlene
Lee, and Anne-Marie Mazza and editor Susanna Carey for their dedication to the project and to the work
of the committee.
Thomas D. Albright and Jed S. Rakoff
Committee Co-chairs

xiii
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

Summary
Eyewitnesses play an important role in criminal cases when they can identify culprits.1 Yet it is well
known that eyewitnesses make mistakes, and their memories can be affected by various factors including
the very law enforcement procedures designed to test their memories. For several decades, scientists have
conducted research on the factors that affect the accuracy of eyewitness identification procedures. Basic
research on the processes that underlie human visual perception and memory have given us an
increasingly clear picture of how eyewitness identifications are made and, more important, an improved
understanding of the principled limits on vision and memory that may lead to failures of identification.
Basic research has been complemented by a growing body of applied research on eyewitness
identification, which has examined those variables that particularly affect eyewitnesses to crimes: system
variables (conditions such as the procedures followed to obtain identifications that can be controlled by
law enforcement) and estimator variables (conditions associated with the actual crime, such as viewing
conditions, or factors specific to the eyewitness, such as the race of the victim relative to that of the
perpetrator, that cannot be controlled by law enforcement).
Through such scientific research, we have learned that many factors influence the visual perceptual
experience: dim illumination and brief viewing times, large viewing distances, duress, elevated emotions,
and the presence of a visually distracting element such as a gun or a knife. Gaps in sensory input are filled
by expectations that are based on prior experiences with the world. Prior experiences are capable of
biasing the visual perceptual experience and reinforcing an individual’s conception of what was seen. We
also have learned that these qualified perceptual experiences are stored by a system of memory that is
highly malleable and continuously evolving, neither retaining nor divulging content in an informational
vacuum. The fidelity of our memories to actual events may be compromised by many factors at all stages
of processing, from encoding to storage to retrieval. Unknown to the individual, memories are forgotten,
reconstructed, updated, and distorted. Therefore, caution must be exercised when utilizing eyewitness
procedures and when relying on eyewitness identifications in a judicial context.
In 2013, the Laura and John Arnold Foundation called on the National Academy of Sciences (NAS)
to appoint an ad hoc study committee to:
1. critically assess the existing body of scientific research as it relates to eyewitness identification;
2. identify any gaps in the existing body of literature and suggest appropriate research questions to
pursue that will further our understanding of eyewitness identification and that might offer
additional insight into law enforcement and courtroom practice;
3. provide an assessment of what can be learned from research fields outside of eyewitness
identification;
4. offer recommendations for best practices in the handling of eyewitness identifications by law
enforcement;
5. offer recommendations for developing jury instructions;
6. offer advice regarding the scope of a Phase II consideration of neuroscience research as well as
any other areas of research that might have a bearing on eyewitness identification; and
7. write a consensus report with appropriate findings and recommendations.
The committee heard from numerous experts, practitioners, and stakeholders and reviewed relevant
published and unpublished literature as well as submissions provided to the committee. In this report, the
committee offers its findings and recommendations for:
1

Throughout this report, the term identification denotes person recognition. Eyewitness identification refers to
recognition by a witness to a crime of a culprit unknown to the witness.

1
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
•
•
•

identifying and facilitating best practices in eyewitness procedures for the law enforcement
community;
strengthening the value of eyewitness identification evidence in court; and
improving the scientific foundation underpinning eyewitness identification.
OVERARCHING FINDINGS

The committee is confident that the law enforcement community, while operating under
considerable pressure and resource constraints, is working to improve the accuracy of eyewitness
identifications. These efforts, however, have not been uniform and often fall short as a result of
insufficient training, the absence of standard operating procedures, and the continuing presence of actions
and statements at the crime scene and elsewhere that may intentionally or unintentionally influence
eyewitness’ identifications.
Basic scientific research on human visual perception and memory has provided an increasingly
sophisticated understanding of how these systems work and how they place principled limits on the
accuracy of eyewitness identification.2 Basic research alone is insufficient for understanding conditions in
the field, and thus has been augmented by studies applied to the specific practical problem of eyewitness
identification. Applied research has identified key variables that affect the accuracy and reliability of
eyewitness identifications and has been instrumental in informing law enforcement, the bar, and the
judiciary of the frailties of eyewitness identification testimony.
A range of best practices has been validated by scientific methods and research and represents a
starting place for efforts to improve eyewitness identification procedures. A number of law enforcement
agencies have, in fact, adopted research-based best practices. This report makes actionable
recommendations on, for example, the importance of adopting “blinded” eyewitness identification
procedures. It further recommends that standardized and easily understood instructions be provided to
eyewitnesses and calls for the careful documentation of eyewitness’ confidence statements. Such
improvements may be broadly implemented by law enforcement now. It is important to recognize,
however, that, in certain cases, the state of scientific research on eyewitness identification is unsettled.
For example, the relative superiority of competing identification procedures (i.e., simultaneous versus
sequential lineups) is unresolved.
The field would benefit from collaborative research among scientists and law enforcement personnel
in the identification and validation of new best practices that can improve eyewitness identification
procedures. Such a foundation can be solidified through the use of more effective research designs (for
example, those that consider more than one variable at a time, and in different study populations to ensure
reproducibility and generalizability), more informative statistical measures and analyses (i.e., methods
from statistical machine learning and signal detection theory to evaluate the performance of binary
classification tasks), more probing analyses of research findings (such as analyses of consequences of
data uncertainties), and more sophisticated systematic reviews and meta-analyses (that take account of
current guidelines, including transparency and reproducibility of methods).
In view of the complexity of the effects of both system and estimator variables and their interactions
on eyewitness identification accuracy, better experimental designs that incorporate selected combinations
of these variables (e.g., presence or absence of a weapon, lighting conditions, etc.) will elucidate those
variables with meaningful influence on eyewitness performance, which can, in turn, inform law
enforcement practice of eyewitness identification procedures. To date, the eyewitness literature has
2

Basic research on vision and memory seeks a comprehensive understanding of how these systems are
organized and how they operate generally. The understanding derived from basic research includes principles that
enable one to predict how a system (such as vision or memory) might behave under specific conditions (such as
those associated with witnessing a crime), and to identify the conditions under which it will operate most effectively
and those under which it will fail. Applied research, by contrast, empirically evaluates specific hypotheses about
how a system will behave under a particular set of real-world conditions.

2
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
evaluated procedures mostly in terms of a single diagnosticity ratio or an ROC (Receiver Operating
Characteristic) curve; even if uncertainty is incorporated into the analysis, many other powerful tools for
evaluating a “binary classifier” are available and worthy of consideration.3 Finally, syntheses of
eyewitness research has been limited to meta-analyses that have not been conducted in the context of
systematic reviews. Systematic reviews of stronger research studies need to conform to current standards
and be translated into terms that are useful for decision makers.
The committee here offers a summary of its key recommendations to strengthen the effectiveness of
policies and procedures used to obtain accurate eyewitness identifications.
RECOMMENDATIONS TO ESTABLISH BEST PRACTICES
FOR THE LAW ENFORCEMENT COMMUNITY
The committee’s review of law enforcement practices and procedures, coupled with its
consideration of the scientific literature, has identified a number of areas where eyewitness identification
procedures could be strengthened. The practices and procedures considered here involve acquisition of
data that reflect a witness’ identification and the contextual factors that bear on that identification. A
recurrent theme underlying the committee’s recommendations is development of and adherence to
guidelines that are consistent with scientific standards for data collection and reporting.
Recommendation #1: Train All Law Enforcement Officers in Eyewitness Identification
The committee recommends that all law enforcement agencies provide their officers and agents
with training on vision and memory and the variables that affect them, on practices for minimizing
contamination, and on effective eyewitness identification protocols.
Recommendation #2: Implement Double-Blind Lineup and Photo Array Procedures
The committee recommends blind (double-blind or blinded) administration of both photo arrays
and live lineups and the adoption of clear, written policies and training on photo array and live lineup
administration.
Recommendation #3: Develop and Use Standardized Witness Instructions
The committee recommends the development of a standard set of easily understood instructions to
use when engaging a witness in an identification procedure.
Recommendation #4: Document Witness Confidence Judgments
The committee recommends that law enforcement document the witness’ level of confidence
verbatim at the time when she or he first identifies a suspect.
Recommendation #5: Videotape the Witness Identification Process
The committee recommends that the video recording of eyewitness identification procedures
become standard practice.

3

T. Hastie, R. Tibshirani, and J.H. Friedman, The Elements of Statistical Learning: Data Mining, Inference,
and Prediction (New York: Springer, 2009).

3
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
RECOMMENDATIONS TO STRENGTHEN THE VALUE
OF EYEWITNESS IDENTIFICATION EVIDENCE IN COURT
The best guidance for legal regulation of eyewitness identification evidence comes not from
constitutional rulings, but from the careful use and understanding of scientific evidence to guide factfinders and decision-makers. The Manson v. Brathwaite test under the Due Process Clause of the U.S.
Constitution for assessing eyewitness identification evidence was established in 1977, before much
applied research on eyewitness identification had been conducted. This test evaluates the “reliability” of
eyewitness identifications using factors derived from prior rulings and not from empirically validated
sources. As critics have pointed out, the Manson v. Brathwaite test includes factors that are not diagnostic
of reliability. Moreover, the test treats factors such as the confidence of a witness as independent markers
of reliability when, in fact, it is now well established that confidence judgments may vary over time and
can be powerfully swayed by many factors. While some states have made minor changes to the due
process framework, wholesale reconsideration of this framework is only a recent development.
Recommendation #6: Conduct Pretrial Judicial Inquiry
The committee recommends that, as appropriate, a judge make basic inquiries when eyewitness
identification evidence is offered.
Recommendation #7: Make Juries Aware of Prior Identifications
The committee recommends that judges take all necessary steps to make juries aware of prior
identifications, the manner and time frame in which they were conducted, and the confidence level
expressed by the eyewitness at the time.
Recommendation #8: Use Scientific Framework Expert Testimony
The committee recommends that judges have the discretion to allow expert testimony on relevant
precepts of eyewitness memory and identifications.
Recommendation #9: Use Jury Instructions as an Alternative Means to Convey Information
The committee recommends the use of clear and concise jury instructions as an alternative means
of conveying information regarding the factors that the jury should consider.
RECOMMENDATIONS TO IMPROVE THE SCIENTIFIC FOUNDATION
UNDERPINNING EYEWITNESS IDENTIFICATION RESEARCH
Basic scientific research on visual perception and memory provides important insight into the
factors that can limit the fidelity of eyewitness identification. Research targeting the specific problem of
eyewitness identification complements basic scientific research. However, this strong scientific
foundation remains insufficient for understanding the strengths and limitations of eyewitness
identification procedures in the field. Many of the applied studies on key factors that directly affect
eyewitness performance in the laboratory are not readily applicable to actual practice and policy. Applied
research falls short because of a lack of reliable or standardized data from the field, a failure to include a
range of practitioners in the establishment of research agendas, the use of disparate research
methodologies, failure to use transparent and reproducible research procedures, and inadequate reporting
of research data. The task of guiding eyewitness identification research toward the goal of evidence-based
policy and practice will require collaboration in the setting of research agendas and agreement on
methods for acquiring, handling, and sharing of data.
4
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Recommendation #10: Establish a National Research Initiative on Eyewitness Identification
The committee recommends the establishment of a National Research Initiative on Eyewitness
Identification.
Recommendation #11: Conduct Additional Research on System and Estimator Variables
The committee recommends broad use of statistical tools that can render a discriminability measure
to evaluate eyewitness performance and a rigorous exploration of methods that can lead to more
conservative responding. The committee further recommends that caution and care be used when
considering changes to any existing lineup procedure, until such time as there is clear evidence for the
advantages of doing so.
CONCLUSION
Eyewitness identification can be a powerful tool. As this report indicates, however, the malleable
nature of human visual perception, memory, and confidence; the imperfect ability to recognize
individuals; and policies governing law enforcement procedures can result in mistaken identifications
with significant consequences. New law enforcement training protocols, standardized procedures for
administering lineups, improvements in the handling of eyewitness identification in court, and better data
collection and research on eyewitness identification, can improve the accuracy of eyewitness
identifications.

5
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

1
Introduction
Accurate eyewitness identifications1 may aid in the apprehension and prosecution of the perpetrators
of crimes. However, inaccurate identifications may lead to the prosecution of innocent persons while the
guilty party goes free. It is therefore crucial to develop eyewitness identification procedures that achieve
maximum accuracy and reliability.
Eyewitness evidence is not infallible. In 1932, Yale University law professor Edwin M. Borchard
documented nearly seventy cases of miscarriage of justice caused by eyewitness errors in his book,
Convicting the Innocent.2 Years later, in 1967, the U.S. Supreme Court highlighted the danger of
erroneous eyewitness identification in United States v. Wade, stating, “The vagaries of eyewitness
identification are well-known; the annals of criminal law are rife with instances of mistaken
identification.”3
The Federal Bureau of Investigation (FBI) estimates that U.S. law enforcement made 12,196,959
arrests in 2012. The FBI estimates that 521,196 of these arrests were for violent crimes.4 Accurate data on
the number of crimes observed by eyewitnesses are not available. If only a fraction of the violent crimes
in the United States involve an eyewitness, the number must be sizable. One estimate based on a 1989
survey of prosecutors suggests that at least 80,000 eyewitnesses make identifications of suspects in
criminal investigations each year.5
Recently, post-conviction DNA exonerations6,7 of innocent persons have dramatically highlighted
the problems with eyewitness identifications. In the United States, more than 300 exonerations have
resulted from post-conviction DNA testing since 1989.8 According to the Innocence Project, at least one
mistaken eyewitness identification was present in almost three-quarters of DNA exonerations.9 In many
of these cases, eyewitness identification played a significant evidentiary role, and almost without
exception, the eyewitnesses who testified expressed complete confidence that they had chosen the
perpetrator. Many eyewitnesses testified with high confidence despite earlier expressions of uncertainty.10
For example, in the well-known case of Ronald Cotton (see Box 1-1), Jennifer Thompson (the victim) has

1

Throughout this report, the term identification denotes person recognition. Eyewitness identification refers to e
recognition by a witness to a crime of a culprit unknown to the witness.
2
Edwin M. Borchard, Convicting the Innocent: Sixty-Five Actual Errors of Criminal Justice (New York: Garden
City Publishing Company, Inc., 1932).
3
United States v. Wade, 388 U.S. 230, 288 (1967).
4
Federal Bureau of Investigation, “Crime in the United States 2012: Persons Arrested,” available at:
http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2012/crime-in-the-u.s.-2012/persons-arrested/persons-arrested.
5
A. G. Goldstein, J. E. Chance, and G. R. Schneller, “Frequency of Eyewitness Identification in Criminal Cases:
A Survey of Prosecutors,” Bulletin of the Psychonomic Society 27(1): 71, 73 (January 1989).
6
CNN, “Exonerated: Cases by the numbers,” December 4, 2013, available at:
http://www.cnn.com/2013/12/04/justice/prisoner-exonerations-facts-innocence-project/.
7
Taryn Simon, “Freedom Row,” New York Times Magazine, January 26, 2003.
8
The Innocence Project, “DNA Exoneree Case Profiles,” available at: http://www.innocenceproject.org/know/.
9
The Innocence Project, “Eyewitness Identification,” available at:
http://www.innocenceproject.org/fix/Eyewitness-Identification.php
10
Brandon L. Garrett, Convicting the Innocent: Where Criminal Prosecutions Go Wrong 63–68 (Cambridge,
MA: Harvard University Press, 2011).

7
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

BOX 1-1 The Ronald Cotton Casea
In 1984, a college student named Jennifer Thompson was raped in her apartment in Burlington, North
Carolina. The police asked her to help create a composite sketch of the rapist. The police then received a tip that a
local man named Ronald Cotton resembled the composite, and shortly after the crime, Thompson was shown a
photo array containing six photos. With some difficulty, she chose two pictures, one of which was of Cotton.
Finally, she said, “I think this is the guy,” pointing to Cotton. “You’re sure,” the lead detective asked, and she
responded, “Positive.” Thompson asked, “Did I do OK?” The detectives responded, “You did great.” She has
described how those encouraging remarks had the effect of making her more confident in her identification.
The police then showed Thompson a live lineup. Cotton was the only person repeated from the prior photo
array. This sequence would make Cotton more familiar and might suggest that he was the prime suspect.
Nevertheless, Thompson remained hesitant and was having trouble deciding between two people. After several
minutes, she told the police that Cotton “looks the most like him.” The lead detective asked “if she was certain,”
and she said, “Yes.” Again, the detectives further reinforced her decision. The lead detective told Thompson, “It’s
the same person you picked from the photos.” She later described feeling a “huge amount of relief” when told
that she had again picked the right person.
At Ronald Cotton's criminal trial, Thompson agreed she was “absolutely sure” that he was the rapist. Cotton
was sentenced to life in prison plus fifty-four years. He served ten and a half years before DNA tests exonerated
him and implicated another man, Bobby Poole. Not only did the identification procedures increase her confidence
in the mistaken memory event, they resulted in her rejection of the actual culprit. Poole had been presented to
Thompson at a post-trial hearing, and she could not recognize him. “I have never seen him in my life,” she said at
the time.
In response to this error, the lead detective in the case, Mike Gauldin, later as police chief, was the first in the
state to institute a series of new practices, including double-blind lineup procedures. In the years that followed,
North Carolina adopted such practices statewide. Ronald Cotton and Jennifer Thompson have since written a
book, “Picking Cotton,” that describes their case and experiences.
a

See generally, http://www.cbsnews.com/news/eyewitness-how-accurate-is-visual-memory/and http://www.slate.com/
articles/news_and_politics/jurisprudence/features/2011/getting_it_wrong_convicting_the_innocent/how_eyewitnesses_
can_send_innocents_to_jail.html.

described how she was initially quite unsure of her eyewitness identification of Cotton, a man later
exonerated by DNA testing. She became certain it was Cotton only after the police made confirmatory
remarks and had her participate in two identification procedures where Cotton was the only person shown
both times.
Erroneous eyewitness identifications can occur across the range of criminal convictions in which
eyewitness evidence is presented, but most of these cases lack the biological material that can be tested
for DNA and used for exoneration purposes. While eyewitness misidentifications may have been a
dominant factor in some erroneous convictions, it is important to note that other factors, including errors
at various stages of the legal and judicial processes, may have contributed to the erroneous convictions.
CHARGE TO THE COMMITTEE
In 2013, the Laura and John Arnold Foundation called on the National Research Council (NRC) to
assess the state of scientific research on eyewitness identification and to recommend best practices11 for
handling eyewitness identifications by law enforcement and the courts. The goal of this effort was to
11

For the purposes of this report, the committee characterizes best practice as the adoption of standardized
procedures based on scientific principles. The committee does not make any endorsement of practices designated as
best practices by other bodies.

8
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
evaluate the scientific basis for eyewitness identification, to help establish the scientific foundation for
effective real-world practices, and to facilitate the development of policies to improve eyewitness
identification validity in the context of the American justice system.
In response to this charge, the NRC appointed an ad hoc committee, the Committee on Scientific
Approaches to Understanding and Maximizing the Validity and Reliability of Eyewitness Identification in
Law Enforcement and the Courts (hereinafter, the committee), to undertake this study (see Box 1-2 for the
committee’s charge). The committee met three times, held numerous conference calls, heard from various
stakeholders (see Appendix B), and reviewed extensive research on eyewitness identification before
reaching its findings and recommendations.
SCIENCE AND LAW
Law enforcement officers12 investigating crimes rely on eyewitness identification procedures to
verify that a suspect is the individual seen by an eyewitness. Such procedures can take place under
conditions that may have significant effects on the accuracy and reliability of an eyewitness’
identification. Unlike officers in the field, laboratory researchers have, in theory, greater control over
influences that might contaminate the visual perceptual experience and memory of an eyewitness.
Science is a self-correcting enterprise. Researchers formulate and test hypotheses using observations
and experiments, which are then subject to independent review. In science, evidence and data are
analyzed and experiments are repeated to ensure that biases or other factors do not lead to incorrect
conclusions. Scientific progress results from the review and revision of earlier results and conclusions.
The culture of scientific research is markedly different from a legal culture that must seek definitive
results in individual cases. In 1993, in Daubert v. Merrell Dow Pharmaceuticals, Inc., the U.S. Supreme
Court ruled that, under Rule 702 of the Federal Rules of Evidence (which covers both civil and criminal
trials in the federal courts), a “trial judge must ensure that any and all scientific testimony or evidence
admitted is not only relevant, but reliable.”13

BOX 1-2 Charge to the Committee
The charge to the NRC was to:
1.
2.
3.
4.
5.
6.
7.

critically assess the existing body of scientific research as it relates to eyewitness identification;
identify any gaps in the existing body of literature and suggest, as appropriate, research questions to
pursue that will further our understanding of eyewitness identification and that might offer additional
insight into law enforcement and courtroom practice;
provide an assessment of what can be learned from research fields outside of eyewitness identification;
offer recommendations for best practices in the handling of eyewitness identifications by law
enforcement;
offer recommendations for developing jury instructions;
offer advice regarding the scope of a Phase II consideration of neuroscience research as well as any other
areas of research that might have a bearing on eyewitness identification; and
write a consensus report with appropriate findings and recommendations.

12

For ease of reading, throughout the report the committee will use the term officer to mean law enforcement
officials and professionals.
13
Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). The Court also noted that “there are
important differences between the quest for truth in the courtroom and the quest for truth in the laboratory. Scientific
conclusions are subject to perpetual revision. Law, on the other hand, must resolve disputes finally and quickly.”

9
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Criminal justice and legal personnel have come to rely on eyewitness evidence. Law enforcement
officials have first-hand experience with eyewitnesses in criminal investigations and trials, and over the
years, some juridictions have implemented and strengthened eyewitnesses practices and procedures in an
attempt to improve acccuracy. Consequently, the law enforcement and legal communities have made
important contributions to our understanding of eyewitness identifications and improvements to practices
in the field. Researchers have become increasingly involved in assessing eyewitness identification
procedures as law enforcement, lawyers, and judges have themselves sought more accurate procedures
and approaches. In the 2009 National Research Council report, Strengthening Forensic Science in the
United States: A Path Forward, the committee noted, “in addition to protecting innocent persons from
being convicted of crimes that they did not commit, we are also seeking to protect society from persons
who have committed criminal acts.”14 This shared common goal of protecting innocent persons and
society makes collaboration between the scientific, law enforcement, and legal communities critically
important to society.
IDENTIFYING THE CULPRIT
Officers typically use three procedures to identify a perpetrator whose identity is unknown: (1)
showups; (2) presentations of photo arrays; and (3) live lineups. A showup is a procedure in which
officers present a single criminal suspect to a witness. This procedure usually occurs near the crime
location and immediately or shortly after the crime has occurred. Officers also use photo arrays and live
lineups, in which they ask the witness to view numerous individuals, one of whom may be the
perpetrator. The suspect is presented along with fillers (known non-suspects). Currently, photo arrays are
used more often than live lineups.15,16
If the eyewitness makes a positive identification during a showup, a photo array, or a lineup, the
identification may constitute evidence about a suspect’s involvement in the crime. The eyewitness
identification may, when considered with other available evidence, establish probable cause to support an
arrest. Such evidence may play a pivotal role in enabling the prosecution to meet its burden of proof in a
subsequent trial.
In recent years, more law enforcement agencies have created written eyewitness identification
policies and have adopted formalized training. However, there are many agencies that do not have
standard written policies or formalized training for the administration of identification procedures or for
ongoing interactions with witnesses.17
VISION AND MEMORY
At its core, eyewitness identification relies on brain systems for visual perception and memory: The
witness perceives the face and other aspects of the perpetrator’s physical appearance and bearing, stores
that information in memory, and later retrieves the information for comparison with the visual percept of
14

National Research Council, Strengthening Forensic Science in the United States: A Path Forward
(Washington, D.C.: The National Academies Press, 2009), p. 12.
15
Police Executive Research Forum, “A National Survey of Eyewitness Identification Procedures in Law
Enforcement Agencies,” March 2013, p. 48. The survey indicates that 94.1 percent of responding law enforcement
agencies reported that they use photo arrays, while only 21.4 percent reported using live lineups. 61.8 percent of
agencies reported that they use showups. See also J. S. Neuschatz et al., “Comprehensive evaluation of showups,”
in Advances in Psychology and Law, ed. M. Miller and B. Bornstein (New York: Springer, in press).
16
Throughout the report, unless otherwise specified, references to lineups refer to both photo arrays and live
lineups.
17
Police Executive Research Forum, p.65.

10
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
an individual in a lineup. Recent years have seen great advances in our scientific understanding of the
basic mechanisms, operational strategies, and limitations of human vision and memory. These advances
inform our understanding of the accuracy of eyewitness identification.
Human vision does not capture a perfect, error-free “trace” of a witnessed event. What an individual
actually perceives can be heavily influenced by bias18 and expectations derived from cultural factors,
behavioral goals, emotions, and prior experiences with the world. For eyewitness identification to take
place, perceived information must be encoded in memory, stored, and subsequently retrieved. As time
passes, memories become less stable. In addition, suggestion and the exposure to new information may
influence and distort what the individual believes she or he has seen.
Several factors are known to affect the fidelity of visual perception and the integrity of memory. In
particular, vision and memory are constrained by processing bottlenecks and various sources of noise.19
Noise comes from a variety of sources, some associated with the structure of the visual environment,
some inherent to the optical and neuronal processes involved, some reflecting sensory content not
relevant to the observer’s goals, and some originating with incorrect expectations derived from memory.
The concept of noise has profound significance for understanding eyewitness identification, as it helps us
realize that the accuracy of information about the environment gained through vision and stored in
memory is necessarily, and often sharply, limited by noise.
The recognition of one person by another—a seemingly commonplace and unremarkable everyday
occurrence—involves complex processes that are limited by noise and subject to many extraneous
influences. Eyewitness identification research confronts methodological challenges that some other basic
experimental sciences do not encounter, as well as practical challenges in establishing adequate
experimental controls over the numerous variables that affect visual perception and memory.
APPLIED RESEARCH ON EYEWITNESS IDENTIFICATION:
SYSTEM AND ESTIMATOR VARIABLES
Our understanding of the underlying processes and limits of eyewitness identification, derived from
basic research on vision and memory, is complemented by research directed specifically at the problem of
eyewitness identification. The modern era of eyewitness identification research began in the 1970s.
Today, eyewitness identification is generally viewed as a behavioral output. The accuracy and reliability
of eyewitness identification are critically modulated by variables that include a witness’ extant cognition
and memory and related psychological and situational factors at the time of the event, over the ensuing
intervals, and at all stages of recall (see Figure 1-1). Because a crime is an unexpected event, one can
draw a natural distinction between variables that reflect the witness’ unplanned situational or cognitive
state at the time of the crime and the variables that reflect controllable conditions and internal states

18

Bias is defined as any tendency that prevents unprejudiced consideration of a question (see Dictionary.com;
http://dictionary.reference.com/browse/bias). Response bias is a general term for a wide range of influences that
moderate the responses of participants away from an accurate or truthful response. Response bias can be induced or
caused by a number of factors, all relating to the idea that humans do not respond passively to stimuli, but rather
actively integrate multiple sources of information to generate a response in a given situation [(see M. Orne,“On the
social psychology of the psychological experiment: with particular reference to demand characteristics and their
implications,” American Psychologist 17: 776–783, (1962)]. In research, bias is seen in sampling or testing when
circumstances select or encourage one outcome or answer over another (see Merriam-Webster.com;
http://www.merriam-webster.com/dictionary/bias).
19
“Noise” refers here to factors that cause uncertainty on the part of an individual about whether a particular
signal (e.g. a specific visual stimulus) is present. This use of the term follows the definition used in electronic signal
transmission, in which noise refers to random or irrelevant elements that interfere with detection of coherent and
informative signals.

11
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
following the witnessed events. Researchers categorize these factors, respectively, as estimator variables
and system variables.20
System variables describe the characteristics of specific procedures and practices (e.g., the content
and nature of instructions given to witnesses who are asked if they are able to make an identification).
The criminal justice system can exert some control over system variables by following standardized
procedures that are based on scientific knowledge and strengthened through education and training.
One important category of system variables concerns the conditions and protocols for lineup
identification. Under current law enforcement practice, eyewitness identification procedures involve
having a witness view individuals or images of individuals. Research indicates that accuracy and
reliability of eyewitness identifications may be influenced by the type of presentation (e.g., lineup) used,
the likeness of non-suspect participants (lineup “fillers”) to the suspect, the number of fillers, and the
suspect’s physical location in the presentation.21,22 Eyewitness performance may be affected by how the
lineup images are presented—simultaneously (as a group) or sequentially (one at a time). System
variables, such as the nature of the instructions and feedback provided before and after the identification
procedure, may also affect the eyewitness’ identification.
Estimator variables affect the accuracy of eyewitness identification, but they are beyond the control
of the criminal justice system. Estimator variables tend to be associated with characteristics of the witness
or factors that are operating either at the time of the criminal event (perhaps relating to memory encoding)
or the retention interval (the time between witnessing an event and the identification process). Specific
examples include the eyewitness’ level of stress or trauma at the time of the incident, the light level and
nature of the visual conditions that affect visibility and the clarity of a perpetrator’s features, and the
physical distance between the witness and the perpetrator. Both system and estimator variables will be
discussed in detail in subsequent chapters.

FIGURE 1-1 Memory Accuracy and Time
SOURCE: Courtesy of Thomas D. Albright.

20

G. L. Wells, “Applied Eyewitness-Testimony Research: System Variables and Estimator Variables,” Journal
of Personality and Social Psychology 36(12):1546–1557 (1978).
21
N. K. Steblay et al., “Eyewitness Accuracy Rates in Police Showup and Lineup Presentations: A MetaAnalytic Comparison,” Law and Human Behavior 27(5): 523–540 (October 2003).
22
R. J. Fitzgerald et al., “The Effect of Suspect-Filler Similarity on Eyewitness Identification Decisions: A
Meta-analysis,” Psychology, Public Policy, and Law 19(2): 151–164 (May 2013).

12
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
EFFORTS AT IMPROVEMENT
In response to insights gained from research on erroneous convictions, there have been attempts to
provide recommendations for improving the reliability and validity of eyewitness identifications. An
effort of particular note is the National Institute of Justice’s (NIJ) Technical Working Group for
Eyewitness Evidence (TWGEYEE). Called together by then-U.S. Attorney General Janet Reno in 1998,
members of the working group were asked to develop and publish guidance for improving eyewitness
identification procedures.23 The working group recognized the role that memory plays in the mistaken
interpretation and remembrance of events and offered guidance based on the practical experiences of the
law enforcement community and insights gained from behavioral and psychological research. The NIJ
provided detailed instructions for each step of the eyewitness identification procedure to the
approximately 18,000 state and local law enforcement agencies across the nation. After the report was
issued, only a few states conducted evaluations and engaged in improvement efforts, including the
implementation of new laws and the issuance of corrective guidelines and policies. Consequently,
eyewitness identification policies remain fragmented by jurisdiction, except in a minority of states that
have adopted state-wide policies. At present, the United States does not have a uniform national set of
protocols.24
JUDICIAL CONSIDERATION OF EYEWITNESS IDENTIFICATION EVIDENCE
The U.S. Supreme Court’s 1977 ruling in Manson v. Brathwaite provides the current framework for
judicial review of eyewitness identification under the Due Process Clause of the U.S. Constitution.25 The
Manson v. Brathwaite test asks judges to evaluate the “reliability” of eyewitness identifications using
factors derived from prior rulings and not from empirically validated sources. The Manson v. Brathwaite
ruling was not based on much of the research conducted by scientists on visual perception, memory, and
eyewitness identification, and it fails to include important advances that have strengthened standards for
judicial review of eyewitness identification evidence at the state level.
In 2011, the Justices of the Massachusetts Supreme Judicial Court convened the Study Group on
Eyewitness Identification to “offer guidance as to how our courts can most effectively deter unnecessarily
suggestive identification procedures and minimize the risk of a wrongful conviction.” The report made
five recommendations to minimize inaccurate identifications: (1) acknowledge variables affecting
identification accuracy; (2) develop a model policy and implement best practices for police departments;
(3) expand use of pretrial hearings; (4) expand use of improved jury instructions; and (5) offer continuing
education.26
In 2011, the New Jersey Supreme Court issued a unanimous decision in State v. Larry R.
Henderson. The opinion revised the legal framework for evaluating and admitting eyewitness
identification evidence and directed that improved jury instructions be prepared to help jurors evaluate
such evidence. Henderson drew on an extensive review of scientific evidence regarding human vision,
memory, and the various factors that can affect the reliability of eyewitness identifications. In July of

23

U.S. Department of Justice, Office of Justice Programs, Eyewitness Evidence: A Guide for Law Enforcement
(Washington, D.C.:, 1999).
24
Police Executive Research Forum, p. 65.
25
Manson v. Brathwaite, 432 U.S. 98, 114 (1977).
26
Massachusetts Supreme Judicial Court Study Group on Eyewitness Identification, Report and
Recommendations to the Justices, July 24, 2013, available at: http://www.mass.gov/courts/docs/sjc/docs/eyewitnessevidence-report-2013.pdf.

13
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
2012, the court released expanded jury instructions, and revised court rules relating to eyewitness
identifications in criminal cases.27
In fall 2012, the Oregon Supreme Court also established a new procedure for evaluating whether
eyewitness identifications could be used in court. In State v. Lawson, the Court reviewed eyewitness
identification research conducted over the past thirty years, determined that the Manson v. Brathwaite test
“does not accomplish its goal of ensuring that only sufficiently reliable identifications are admitted into
evidence,” and offered a revised procedure that requires the court to make a determination of whether
investigators used “suggestive” tactics to get an identification and the extent to which other information
supports the identification.28
Despite these improvements and judicial decisions, policies and practices across the country remain
inconsistent.
ORGANIZATION OF THE REPORT
This report begins with a description of law enforcement protocols for eyewitness identification
(Chapter 2). Chapter 3 presents the legal framework for eyewitness identification evidence. A discussion
of the current scientific understanding of visual perception and memory follows in Chapter 4. In Chapter
5, the committee provides an analysis of eyewitness identification research. The report concludes with the
committee’s findings and recommendations (Chapter 6).

27

New Jersey Judiciary, “Supreme Court Releases Eyewitness Identification Criteria for Criminal Cases,” July
19, 2012, available at: http://www.judiciary.state.nj.us/pressrel/2012/pr120719a.htm.
28
State v. Lawson, 352 Or. 724 (Or. 2012).

14
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

2
Eyewitness Identification Procedures
Police in the United States investigate millions of crimes each year.1 Only a small percentage of the
police-investigated crimes involve the use of police-arranged identification procedures. Identification
procedures are unnecessary when, for example, the perpetrator is caught during the commission of the
criminal act, as in the crime of driving while intoxicated, or when the victim knows the perpetrator, as in
crimes of domestic violence.2
Police use identification procedures for numerous reasons. In some circumstances, the police
identify a suspect during an investigation and use the identification procedure to test a witness’ ability to
identify the suspect as the perpetrator. In other instances, the identification procedure is used as an
investigative tool to further an investigation. A positive identification might form probable cause for a
search warrant or the apprehension and subsequent questioning of a suspect, or both. Most significant for
the purposes of this report are the circumstances in which a witness positively identifies the police suspect
as the perpetrator, and the identification serves as compelling evidence in the prosecution of a case.
Data on the number of eyewitness identification procedures are not systematically or uniformly
collected. While the exact number of eyewitness identification procedures related to crimes involving
strangers is unknown, mistaken identifications have disastrous effects for those wrongly accused of
crimes and for society should a guilty person goes free. Mistaken identifications may also erode public
confidence in the criminal justice system as a whole.3 Recently, some police departments and prosecutors
have implemented stringent eyewitness identification procedures in an effort to reduce erroneous
convictions.4
Police-arranged eyewitness identification procedures vary greatly depending on the nature of the
case. In some cases, a police-arranged identification is conducted at the very early stages of an
investigation. For instance, consider the circumstance in which police respond to a bank robbery in
progress. The perpetrator is described as a white male, approximately 6 feet, 2 inches in height wearing an
orange shirt. As the police arrive at the crime scene, an officer observes and apprehends a man fleeing the
bank wearing an orange shirt and exhibiting similar physical characteristics. In this situation, a policearranged identification procedure may be conducted on the scene and prior to any significant
investigation. At the other extreme are, for example, lengthy homicide or rape cases that include extensive
investigations, forensic testing, and eyewitness interviews conducted over a protracted period of time.
Such efforts may culminate in the identification of a suspect and the suspect’s inclusion in a photo array
identification procedure. In such a circumstance, an eyewitness may not be asked to identify a perpetrator
1

Federal Bureau of Investigation, “Crime in the United States 2012: Violent Crime,” available at:
http://www.fbi.gov/about-us/cjis/ucr/crime-in-the-u.s/2012/crime-in-the-u.s.-2012/violent-crime/violent-crime.
2
Throughout Chapter 2, the terms law enforcement and police are used interchangeably and refer to all law
enforcement agencies at the local, state, and federal levels.
3
See, generally, The International Association of Chiefs of Police, “National Summit on Wrongful Convictions:
Building a Systemic Approach to Prevent Wrongful Convictions,” August 2013.
4
See The Innocence Project, Eyewitness Identification, available at:
http://www.innocenceproject.org/fix/Eyewitness-Identification.php; U.S. Department of Justice, Office of Justice
Programs, Eyewitness Evidence: A Guide for Law Enforcement (Washington, D.C.:, 1999); Metropolitan Police –
District of Columbia, General Order – Procedures for Obtaining Pretrial Eyewitness Identification, April 18, 2013;
New York State District Attorneys Association Best Practice Committee, New York State Photo Identification
Guidelines, October 2010; Rhode Island Police Chiefs Association, “Lineup and Showup Procedures (Eyewitness
Identification), November 2011; and Innocence Project of Texas, Eyewitness Identification Reform, available at:
http://www.ipoftexas.org/eyewitness-id.

15
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
until months after the commission of the crime—and often after repeated probes of their memory by, for
example, police, family members, and others.
Identification procedures may be used in different ways for different purposes. They are not always
used to identify an unknown perpetrator of a crime. The police may, for example, use photo arrays and
confirmatory single photographs to clarify the legal identity (birth name/government name) of an
individual who is well known to a witness, but only by a street name. In such examples, a witness may
know (and may have known) the perpetrator for years but may only be able to identify him by a common
street name, like “Prince.” The police typically will use an identification procedure to identify to which
“Prince” the witness is referring before they make an arrest or take other investigative measures such as
the execution of a search warrant.
This chapter reviews the eyewitness identification procedures commonly used by the police and
concludes with a brief discussion of situations in which citizens engage in identifying perpetrators
without police assistance.
PHOTOGRAPHIC ARRAY
The photo array is the most common police-arranged identification procedure used in the United
States.5 A photo array consists of six to nine photographs displayed to a witness. An officer might create
an array by selecting photographs of persons deemed to resemble the perpetrator.6 Officers might then
display the photographs one at a time to the witness and ask whether she or he recognizes each one. This
method is known as a sequential procedure. Officers might also create photo arrays by cutting six square
holes in a folder and taping the photographs to the back of the folder so that the faces of the fillers (nonsuspects) and suspect are displayed together. When such photographs are presented simultaneously as a
two by three matrix, this type of array is referred to as a “six pack.” When, as in this instance,
photographs are displayed together, this is referred to as a simultaneous procedure.
In 1999, Attorney General Janet Reno released the U.S. Department of Justice Eyewitness Evidence:
A Guide for Law Enforcement,7 one of the earliest efforts to establish standardized procedures for policearranged eyewitness identification. The guide set forth rigorous criteria and basic procedures to promote
accuracy in eyewitness evidence.8 However, after the guide was released most police departments in the
United States did not adopt these procedures.
Today, many police departments use computer systems to access image databases and assemble
photo arrays. Officers enter physical characteristics (e.g., race, gender, hair color) specific to the suspect
into a computer, and the system retrieves filler photographs with the desired attributes. If an officer
determines that a photograph in the array is suggestive or otherwise inappropriate, she or he can reject one
or more fillers and instruct the system to provide alternate photographs. Departments may conduct the
procedure without revealing to the witness how many photographs she or he will view.
In recent decades, many police agencies and prosecutors have adopted sequential presentation of
photographs, based on the belief that this approach improves the performance of an eyewitness.
Currently, however, there is no consensus among law enforcement professionals as to whether the
sequential presentation procedure is superior to the simultaneous procedure (see Chapter 5). The District
of Columbia Metropolitan Police Department, for example, does not endorse either simultaneous or
sequential procedures in its Procedures for Obtaining Pretrial Eyewitness Identification.9 The District
5

Police Executive Research Forum, p. 48.
Historically, the photographs were mug shots in the possession of a police department.
7
See generally, U.S. Department of Justice, Office of Justice Programs, Eyewitness Evidence: A Guide for Law
Enforcement (Washington, D.C.: 1999).
8
See U.S. Department of Justice, Office of Justice Programs, Eyewitness Evidence: A Guide for Law
Enforcement, p. 2.
9
See Metropolitan Police – District of Columbia, General Order – Procedures for Obtaining Pretrial
Eyewitness Identification, April 18, 2013.
6

16
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Attorneys Association of the State of New York in 2010 adopted recommended policies for New York
State and endorsed the simultaneous method.10 On the other hand, in North Carolina, legislation was
passed that requires that lineup photographs be presented sequentially,11 and in Massachusetts, the
Supreme Judicial Court Study Group on Eyewitness Identification recommended sequential procedures as
best practice for Massachusetts Police Departments.12.
The committee was presented with information regarding improvement efforts from states including
New Jersey, Oregon, Rhode Island, Texas, New York and Massachusetts. However, the committee is
unable to determine the percentage of police departments that have adopted policies for eyewitness
identification procedures and instituted training in these procedures.13 Some police departments require
that photo arrays be presented to the witness during a procedure that is either “double blind” or
“blinded.”14 (See Box 2-1 for a discussion of blinding as used in scientific practice and blinding as used in
an eyewitness identification procedures.) Blinding is used to prevent conscious and unconscious cues
from being given to the witness. In a double-blinding procedure, an individual who does not know the
identity of the suspect or the suspect’s position in the photo array shows a photo array to the eyewitness.
In cases where such a double-blind procedure is not feasible, a “blinded” procedure will approximate the
condition of double-blinding. For example, the photo array may be administered by an individual who
knows who the suspect is, but is unable to tell when the witness is looking at the suspect’s photo and so is
unable to provide even subconscious feedback to the witness. In one common “blinded” procedure, the
officer places each photo in a separate envelope or folder and then shuffles the envelopes/folders so only
the witness sees the images therein. Additional recommendations to minimize the possibility of biasing
feedback to the witness include requiring that the officer read instructions to the witness from a preprinted form.15
If the witness identifies someone from the photo array, some departments ask the witness for a
confidence statement. Based on information presented to the committee, it would appear that police
departments do not always document identification procedures in instances when an identification is not
made. Further, if a witness does make an identification, practices differ as to how such information is
documented and preserved. Some agencies, for example, require officers to document this information in
a written report. Others make audio or video recordings of the identification procedure.
LIVE LINEUP
A live lineup is a police-arranged identification procedure in which the physical suspect and
fillers/foils stand or sit in front of the witness (either individually, i.e., sequentially or en masse, i.e.,
simultaneously). The police generally use at least five fillers. Fillers are selected for their physical
similarities to the suspect (gender, race, hair length and color, facial hair, height, skin tone, and other
distinguishing features). The fillers are presumed to be unknown to the witness. Traditionally, the suspect
10

See New York State District Attorneys Association Best Practice Committee, New York State Photo
Identification Guidelines, October 2010.
11
N.C. Gen. Stat. § 15A-284.52 (West 2007).
12
Massachusetts Supreme Judicial Court Study Group on Eyewitness Identification, Report and
Recommendations to the Justices, July 24, 2013, available at: http://www.mass.gov/courts/docs/sjc/docs/eyewitnessevidence-report-2013.pdf.
13
The Police Executive Research Forum’s 2013 survey of eyewitness identification procedures in law
enforcement agencies [Police Executive Research Forum, A National Survey of Eyewitness Identification
Procedures in Law Enforcement Agencies, (2013)], notes that most agencies that completed the survey have no
written policy for eyewitness identification procedures and that more agencies provide training to their employees
than have written policies. See pp. 79–80.
14
Police Executive Research Forum, p. 64.
15
As discussed in Chapter 3, the courts have been sensitive to the potential for misidentification resulting from
“suggestive” identification procedures, and have set standards for admissibility of evidence.

17
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
and fillers are seated or standing in a row and the witness views the lineup from behind a two-way mirror.
Police use both simultaneous and sequential procedures for live lineups.

BOX 2-1 Blinding
Empirical evidencea has shown that the beliefs, desires and expectations of researchers can influence, often
subconsciously, how they observe and interpret the phenomena they study and thus the outcomes of experiments.
This evidence has influenced how scientists carry out their experiments, resulting in the use of blind or doubleblind procedures to control for this form of bias. Blind assessment b has been used since the late 18th century; an
early medical trial in 1835 used double-blind assessment, and psychologists started using it in the 20th century.c
By the 1950s, blind assessment in randomized controlled trials was considered standard procedure in both
psychological and medical research. Currently, virtually all of science uses some form of blinding.
In single-blind experiments, participants do not know which treatment they are receiving; this form of
blinding is used widely across scientific fields. In experiments involving humans, as in medical or psychological
research, double-blind procedures are used to guard against “expectancy effects” for both participants and
researchers. In a classic double-blind clinical trial, some patients will receive active medication and others are
given an alternative (either a “standard treatment” or a similar-looking placebo without the active ingredients), but
neither researchers nor participants know who is receiving which treatment.
In an eyewitness identification setting, double-blinding can be used to prevent a lineup administrator from
either intentionally or unintentionally influencing a witness. In these cases, neither the eyewitness nor the
administrator knows which persons in a photo array or live lineup are the suspected culprits and which are the
“fillers.” d, e In eyewitness identification procedures, as in science, the purpose of double-blinding is to prevent the
conscious or subconscious expectations of the administrator from influencing the witness or research outcomes.
In a double-blind photo array, the officer or detective conducting the investigation reads a set of standard
instructions to the witness. The instructions may include an advisory that the officer about to show the photos
does not know whether any of the photos are of the person who committed the crime. The officer then leaves the
room and a second officer—perhaps a patrol officer—displays the photos. It is the duty of this second officer (the
“blind administrator”) to show the photos and, if an identification is made, document what the witness said and
ask the witness how certain she or he is of their identification. Once all photos have been shown, the officer
reports the result of the procedure to the investigating officer (preferably out of earshot from the witness).
As an alternative to a “double-blind” array, some departments use “blinded” procedures. A blinded procedure
prevents an officer from knowing when the witness is viewing a photo of the suspect, but can be conducted by the
investigating officer. A common approach is the so-called “folder shuffle.” With a six-photo array, an officer uses
eight manila folders. A photograph of a filler is placed in the top folder and a photograph of the suspect and four
additional fillers are placed in the next five folders. The six folders are then shuffled so that the officer does not
know which folder contains the image of the suspect. Two folders with blank paper are placed on the bottom of
the stack so that the witness is led to believe that there are more than five images in the array (this is referred to as
back-loading, and it prevents the witness from knowing when she or he is about to view the last photograph). The
five folders are placed on top of the two folders without images, and the folder containing the image of the known
filler is placed on top of the stack. After reading instructions to the witness, the administering officer sits to the
witness’ left and hands him or her one folder at a time and instructs him/her to open each folder and look at the
enclosed photo. The cover of the folder blocks the officer from viewing the photo that the witness is viewing.
When an identification occurs, the officer notes the witness’ words and reaction and asks about the witness’
confidence in his/her identification.
a

R. Rosenthal, Experimenter Effects in Behavioral Research (New York: John Wiley, 1976).
M. Stolberg, “Inventing the randomized double-blind trial: The Nürnberg salt test of 1835,” James Lind Library
Bulletin (2006), available at: http://www.jameslindlibrary.org/illustrating/articles/inventing-the-randomizeddouble-blind-trial-the-nurnberg-salt.
c
T. J. Kaptchuk, ntentional ignorance: A history of blind assessment and placebo controls in medicine.” Bulletin of the
History of Medicine 72 (3): 389–433 (1998).
d
P. Kilmartin, Presentation to the committee, February 6, 2014.
e
K. Hamann, Presentation to the committee, December 2, 2013.
b

18
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Live lineups are used in some jurisdictions, but they are not the predominant method used by law
enforcement.16 The use of these police identification procedures is limited for a variety of reasons. First,
in certain circumstances, legal counsel may be required at a lineup, thereby making it less attractive to
police and prosecutors. Second, in smaller jurisdictions, it is difficult to obtain suitable fillers (e.g., those
with appropriate physical similarities to the suspect). Third, conducting a lineup requires a significant
amount of time and labor, thereby making photo arrays a more attractive alternative that may be
undertaken promptly and with less demand on department resources.17
SHOWUP
A showup is a police-arranged identification procedure in which the police show one person to a
witness and ask if she or he recognizes that person. This procedure typically is used when the police
locate the suspect shortly after the commission of the crime and within close proximity to the scene.
Case law limits the time and distance from a crime during which such a procedure will pass legal
standards. In response to such case law, police typically restrict showups to a two-hour time period after
the commission of a crime. Ideally, officials take the witness to the location where the suspect has been
detained and do not display the suspect in a suggestive manner (e.g., not in a police car, not handcuffed,
without drawn weapons). However, as chases, fights, or disarmaments frequently precede showups, the
apprehension of a suspect can raise safety issues that make it difficult to adhere to recommended
procedures. Further, the nature of a showup does not lend itself to the use of a blinded procedure. A
showup is used to promptly clear innocent suspects, thereby sparing them from a prolonged period of
detention as the investigation continues. Delaying the showup to locate an uninvolved officer may defeat
that purpose. While some law enforcement agencies use a standard procedure with written instructions
when conducting a showup, there is no indication that such procedures are used uniformly. Courts
consider showups highly suggestive, and prosecutors urge the police to exercise caution when conducting
them.
CONFIRMATORY PHOTOGRAPH
Police will, on occasion, display a single photograph to a witness in an effort to confirm the identity
of a perpetrator. Police typically limit this method to situations in which the perpetrator is previously
known to or acquainted with the witness.
FIELD VIEW
The police also use field views in attempts to identify perpetrators. The method, which involves
inviting a witness to view many people in a context where the perpetrator is thought likely to appear, is
used when the police do not have a suspect but believe that the offender frequents a particular location.
For example, police investigating a purse snatching may obtain information that the perpetrator frequents
a particular recreation site during the lunch hour. A plainclothes officer or investigator might take the
eyewitness to the site and walk around with him or her during the lunch hour without directing his or her
attention to any specific individual.
OTHER PROCEDURES – MUG BOOKS AND YEARBOOKS
At times, police use other means to identify perpetrators. In the past, police sometimes had
witnesses review mug shot books. Mug books have since been largely replaced by digitized images
16

Police Executive Research Forum, p. 48.
Live lineup construction may be further constrained by the inability to hold a suspect in custody without
probable cause. See Chapter 3.
17

19
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
displayed on computer screens. Nonetheless, there are situations in which the police will have a witness
review a large collection of photographs in an effort to identify a perpetrator. Witnesses who identify a
perpetrator as being a student at a specific school might be asked to review a yearbook for that school in
an effort to identify the perpetrator. When using this method, police typically attempt to mask the names
of the students. Similarly, if the offender is believed to be an individual from a certain profession, the
police might have the witness review photographs from the suspect’s professional society. Social media
sites also serve as the catalyst for novel police-arranged identification procedures. If a witness knows that
the perpetrator is a “friend” of Jane Doe through social media, the police might have the witness review
all friends of Jane Doe to see if she or he recognizes the individual.
All of these additional procedures (confirmatory photo, field view, mug books, yearbooks) have the
potential to introduce biases of the sort that blind lineup procedures are designed to avoid.
NON-POLICE IDENTIFICATION PROCEDURES
In some cases, the victims or witnesses, or both, identify suspects without involving the police. A
private citizen, organization, or corporation may conduct its own investigation before, during, or even
after reporting a crime to the police. The identification of suspects by entities other than law enforcement
has become increasingly common as more businesses and private citizens use security cameras to identify
criminal actors. High-resolution cameras coupled with high capacity hard drives allow for real-time
streaming of video with superior clarity. Such systems are relatively inexpensive and within financial
reach of many home and business owners. Additionally, the proliferation of smart phones puts the ability
to create a spontaneous, high quality video record of an event into the hands of more and more people.
The rise of social media has resulted in the rise of private investigations and identifications using
this resource. In one recent case,18 a stabbing victim drew a picture of her assailant and showed it to her
husband. Upon viewing the picture, the husband believed that the assailant looked familiar and might be
his ex-girlfriend. He obtained several pictures of the ex-girlfriend from her personal website and showed
them to the victim who, after looking at those and other photographs online, identified the suspect at a
lineup and at trial.
CONCLUSION
Many local, state, and federal law enforcement agencies have adopted policies and practices to
address the issue of misidentification. However, efforts are not uniform or systemic.19 Many agencies are
unfamiliar with the science that has emerged during the past few decades of research on eyewitness
identifications. Questions remain about the optimal design of photo array procedures, including the size of
the array, the contents of the photographs, and their relationship to the context of the crime scene. Similar
questions apply to the design of live lineups.20 Eyewitness identification is further complicated by the
increasing number of situations in which victims and witnesses seek to identify the perpetrator of a crime
without the aid of law enforcement. Such identifications raise new concerns about reliability and accuracy
of the identification of individuals. Inconsistent and nonstandard practices might easily add noise to the
eyewitness identification process, contaminate the witness, and bias the outcome of an identification
procedure.

18

New Jersey v. Chen, 27 A.3d 930 (N.J. 2011).
See Massachusetts Supreme Judicial Court Study Group on Eyewitness Evidence, Report and
Recommendations to the Justices, July 25, 2013, p. 2.
20
The design of a live lineup is subject to more practical constraints than a photo array.
19

20
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

3
The Legal Framework for Assessment of
Eyewitness Identification Evidence
The admissibility of eyewitness testimony at a criminal trial may be challenged on the basis of
procedures used by law enforcement officials in obtaining the eyewitness identification. The U.S.
Supreme Court, in its 1977 ruling in Manson v. Brathwaite, set out the modern test under the Due Process
Clause of the U.S. Constitution that regulates the fairness and the reliability of eyewitness identification
evidence.1 The Court also specified five reliability factors, discussed below, that a judge must consider
when deciding whether to exclude the identification evidence at trial.2
Although the constitutional standards for assessing eyewitness testimony have remained unchanged
in the decades since the Manson v. Brathwaite decision, a body of research has shed light on the extent to
which each of the five reliability factors supports a reliable eyewitness identification. Research has cast
doubt, for instance, on the belief that the apparent certainty displayed in the courtroom by an eyewitness
is an indicator of an accurate identification, and has found that a number of factors may enhance the
certainty of the eyewitness.
Recently state courts and lower federal courts have taken the lead in developing standards relating to
the admissibility of expert evidence, jury instructions, and judicial notice of scientific evidence. Some
states have adopted more stringent standards for regulating eyewitness identification evidence than the
U.S. Constitution requires, either by legislative statutes or by state court decisions, and have modified or
entirely supplanted the Manson v. Brathwaite test to take account of advances in the growing body of
scientific research. This chapter describes the changes in the legal standards for eyewitness identification
and explores the relationship between the state of the scientific research and the law regulating procedures
and evidence.
EYEWITNESS EVIDENCE AND DUE PROCESS UNDER THE U.S. CONSTITUTION
Beginning with rulings in 1967, the U.S. Supreme Court set out a standard under the Due Process
Clause of the Fourteenth Amendment for reviewing eyewitness identification evidence.3 In Manson v.
Brathwaite, the Court emphasized that “reliability is the linchpin in determining the admissibility of
identification testimony.”4 First, the Court instructed judges to examine whether the identification
procedures were unnecessarily suggestive. Second, to assess whether an identification is reliable, judges
were instructed to examine the following five factors: (1) the opportunity of the witness to view the
criminal at the time of the crime; (2) the witness’ degree of attention; (3) the accuracy of the witness’
prior description of the criminal; (4) the level of certainty demonstrated at the confrontation; and (5) the

1

Manson v. Brathwaite, 432 U.S. 98, 113–114 (1977).
Manson v. Brathwaite at 114.
3
In Stovall v. Denno, 388 U.S. 293, 302 (1967), the U.S. Supreme Court first set out a due process rule asking
whether identification procedures used were “so unnecessarily suggestive and conducive to irreparable mistaken
identification.” The Court elaborated that rule in decisions such as Simmons v. U.S., 390 U.S. 377, 384 (1968) and
Foster v. California, 394 U.S. 440, 442 (1969), and then adopted an approach setting out “reliability” considerations
in Neil v. Biggers, 409 U.S. 188 (1972). For a description of the development of this doctrine, see, e.g., B. L.
Garrett, “Eyewitnesses and Exclusion,” Vanderbilt Law Review 65 (2): 451, 463–67 (2012).
4
Brathwaite, 423 U.S. at 114.
2

21
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
time between the crime and the identification procedure.5 The five factors were drawn from earlier
judicial rulings and not from scientific research.6
Eyewitness identification evidence continues to be litigated primarily under the flexible two-part
Manson v. Brathwaite Due Process test.7 It is important to note, however, that the vast majority of
criminal cases are settled through plea bargaining. The role that evidence type and strength play in plea
bargaining is complex and necessarily difficult to study. Because eyewitness identification evidence may
never be tested at trial, it is doubly important for lawyers and judges to understand the credibility of the
proffered evidence.8
In the most recent U.S. Supreme Court ruling addressing a challenge to an eyewitness identification
(Perry v. New Hampshire),9 the Court ruled that a due process analysis was not triggered. In that case,
while the police were obtaining a description of the suspect, the eyewitness looked out of the apartment
window and recognized the suspect standing outside. The police had not intended to conduct an
identification procedure. In those circumstances, the Court ruled that the Due Process Clause does not
require a preliminary judicial review of the reliability of an eyewitness identification.10

5

Id. at 114.
Id. at 114. Justice Thurgood Marshall dissented, noting studies indicated that unnecessarily suggestive
eyewitness identifications had resulted in “repeated miscarriages of justice resulting from juries’ willingness to
credit inaccurate eyewitness testimony.” 432 U.S. at 125–27 (Marshall, J., dissenting).
7
Due process is the most important constitutional right that arises in challenges to eyewitness identification, but
rights under the Fourth and Sixth Amendments also may be implicated. The Fourth Amendment protects individuals
“against unreasonable searches and seizures,” and the probable cause typically required to seize and arrest a suspect
may arise from an eyewitness identification. U.S. Const. Amend. IV. The few lower courts to address the question
are divided on whether probable cause is needed to place individuals in a live lineup procedure. Biehunik v.
Felicetta, 441 F.2d 228, 230 (2d Cir. 1971); but see, e.g., Wise v. Murphy, 275 A.2d 205, 212–15 (D.C. 1971); State
v. Hall, 461 A.2d 1155 (N.J. 1983). In contrast, probable cause is not required to place a person’s photograph in an
array, since doing so does not involve a seizure. However, courts may also rule that an illegal stop or seizure
renders a subsequent identification inadmissible, absent an “independent” source for the courtroom identification.
U.S. v. Crews, 445 U.S. 463, 473 (1980).
In addition, the Sixth Amendment provides that in all criminal prosecutions, the accused has the right “to have
the assistance of counsel for his defense.” In United States v. Wade, the Supreme Court held that, once indicted, a
person has a right to have a lawyer present at a lineup, reasoning that the right to counsel applies at all “critical”
stages of the criminal process. 388 U.S. 218, 235–37 (1967). However, the Court subsequently held that a photo
array procedure, of the type now most commonly used by police agencies, does not implicate the Wade right to
counsel. U.S. v. Ash, 413 U.S. 300, 321 (1973).
8
As the current report demonstrates, a comparative consideration of evidence value is particularly important in
the case of eyewitness identification evidence. Similar consideration should be given when other adjudication
mechanisms are used (e.g., bench trials).
9
Perry v. New Hampshire, 132 S. Ct. 716, 718 (2012). In that case, the eyewitness happened to look out her
window and see the suspect standing at the crime scene where the police had told him to wait. The Court held that
the Due Process Clause did not regulate such a situation, since the police did not intend to conduct an identification
procedure. Id. at 729. The Court indicated that the reliability of the evidence could be addressed by federal and state
evidentiary standards, and added: “In appropriate cases, some States also permit defendants to present expert
testimony on the hazards of eyewitness identification evidence.” Id.
10
Justice Sotomayor dissented, arguing, “Our due process concern. . . arises not from the act of suggestion, but
rather from the corrosive effects of suggestion on the reliability of the resulting identification,” and the manner in
which “[a]t trial, an eyewitness’ artificially inflated confidence in an identification’s accuracy complicates the jury’s
task of assessing witness credibility and reliability.” Perry, 132 S. Ct. at 731–32 (Sotomayor, J., dissenting). Justice
Sotomayor also emphasized: “A vast body of scientific literature has reinforced every concern our precedents
articulated nearly a half-century ago.” Id. at 738.
6

22
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
STATE LAW REGULATION OF EYEWITNESS EVIDENCE
State Supreme Court Standards
Several state supreme courts have altered or supplemented the federal Manson v. Brathwaite due
process rule to focus more on the effects of suggestion, to emphasize certain factors in specific
circumstances,11 or to focus on showup identifications in particular.12 New Jersey and Oregon have now
supplemented the Manson v. Brathwaite test with separate state law standards regulating eyewitness
identification evidence.
In 2011, the New Jersey Supreme Court issued a unanimous decision in State v. Larry R.
Henderson13 that revised the legal framework for admitting eyewitness identification evidence and
directed that revised jury instructions be prepared to help jurors evaluate such evidence. The new
framework was based on the record of hearings before a Special Master that considered an extensive
review of scientific research regarding eyewitness identifications.14 The legal framework established by
the Henderson opinion relies on pretrial hearings to review eyewitness evidence and more comprehensive
jury instructions at trial.15 To obtain a pretrial hearing, a defendant must show some evidence of
suggestiveness related to either estimator or system variables that could lead to mistaken identification.16
At the pretrial hearing, the State must offer proof that the eyewitness identification is reliable. However,
the ultimate burden of proving a “very substantial likelihood of irreparable misidentification” is on the
defendant.17
In July 2012, the New Jersey Supreme Court released an expanded set of jury instructions and
related rules that govern the use of suggestive identifications.18 The jury instructions state that, “[r]esearch
has shown that there are risks of making mistaken identifications” and noted that eyewitness evidence

11

See State v. Ramirez, 817 P.2d 774, 780–81 (Utah 1991) (altering three of the reliability factors to focus on
effects of suggestion); State v. Marquez, 967 A.2d 56, 69–71 (Conn. 2009) (adopting criteria for assessing
suggestion); Brodes v. State, 614 S.E.2d 766, 771 & n.8 (Ga. 2005) (rejecting eyewitness certainty jury instruction);
State v. Hunt, 69 P.3d 571, 576 (Kan. 2003) (adopting Utah’s five factor “refinement” of the Biggers factors); State
v. Cromedy, 727 A.2d 457, 467 (N.J. 1999) (requiring, when applicable, instruction on cross-racial
misidentifications).
12
See, e.g., State v. Dubose, 285 Wis.2d 143, 166 (Wis. 2005); Commonwealth v. Johnson, 650 N.E.2d 1257,
1261 (Mass. 1995); People v. Adams, 423 N.E.2d 379, 383–84 (N.Y. 1981).
13
State v. Henderson, 27 A.3d 872 (N.J. 2011). The Henderson opinion described criticisms of the Manson v.
Brathwaite test, including that suggestion may itself affect the seeming “reliability” of the identification. Id. at 877–
78. For examples of scholarly criticism of the Manson v. Brathwaite test in light of scientific research, see, e.g., G.
L. Wells and D. S. Quinlivan, “Suggestive Eyewitness Identification Procedures and the Supreme Court’s
Reliability Test in Light of Eyewitness Science: 30 Years Later,” Law and Human Behavior 33 (1): 1, 16 (February
2009); T. P. O’Toole and G. Shay, “Manson v. Brathwaite Revisited: Towards a New Rule of Decision for Due
Process Challenges to Eyewitness Identification Procedures,” Valparaiso University Law Review 41 (1): 109 (2006).
14
See Report of the Special Master at 16–17, State v. Henderson, No. A-8-08 (N.J. June 18, 2011) (available at:
http://www.judiciary.state.nj.us/pressrel/HENDERSON%20FINAL%20BRIEF%20.PDF%20(00621142.pdf).
15
In the companion case, State v. Chen, 27 A.3d 930, 932 (N.J. 2011), the New Jersey Supreme Court took an
approach that departed from that of the U.S. Supreme Court in Perry, ruling that the defendant may be entitled to a
hearing in a case in which the eyewitness identified the defendant using social media, not a police-orchestrated
identification procedure.
16
Henderson, 27 A.3d. at 878.
17
Id.
18
New Jersey Criminal Model Jury Instructions, Identification (July 19, 2012), available at:
http://www.judiciary.state.nj.us/pressrel/2012/jury_instruction.pdf; New Jersey Court Rule 3:11, Record of an Outof-Court Identification Procedure (July 19, 2012), available at:
http://www.judiciary.state.nj.us/pressrel/2012/new_rule.pdf, New Jersey Court Rule 3:13-3, Discovery and
Inspection (July 19, 2012), available at: http://www.judiciary.state.nj.us/pressrel/2012/rev_rule.pdf.

23
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
“must be scrutinized carefully.”19 Human memory involves three stages—encoding, storage, and retrieval.
At “each of these stages, memory can be affected by a variety of factors.”20 The Court identified a set of
factors that jurors should consider when deciding whether eyewitness identification evidence is reliable,
including estimator variables (stress, duration, weapon focus, distance, lighting, intoxication, disguises or
changed appearance of the perpetrator, time since the incident, and cross-racial effects) and system
variables (lineup composition, fillers, use of multiple viewings, presence of feedback, use of double-blind
procedures, and use of showup identifications). The instructions also noted the possible influence of
outside opinions, descriptions or identifications by other witnesses, and photographs or media accounts.21
In 2012, in Oregon v. Lawson, the Oregon Supreme Court established a new procedure for
evaluating the admissibility of eyewitness identifications. In a unanimous decision, the Court found
“serious questions” about the reliability of eyewitness identification,22 citing research conducted over the
past thirty years. The Court determined that the Manson v. Brathwaite two-step process for weighing
eyewitness identification “does not accomplish its goal of ensuring that only sufficiently reliable
identifications are admitted into evidence,” because it relies on an eyewitness’ self-reports to determine
whether the threshold level of suggestiveness is reached, rendering the identification unreliable.23 The
Court set forth a process that requires the trial court to examine whether investigators used “suggestive”
identification procedures and whether other factors, such as estimator variables, may have affected the
reliability of the identification.24 The Court ruled that “intermediate remedies,” including the use of expert
testimony, should be available even if the trial judge concludes that the identification is admissible. The
Court also briefly noted that judges might use “case-specific jury instructions.”25
Other states continue to explore possible changes to the judicial review of eyewitness identification
evidence. In 2013, the Massachusetts Supreme Judicial Court Study Group on Eyewitness Identification
offered guidance on the adjudication of eyewitness identification evidence.26 The report adopted
Lawson’s approach of taking judicial notice of “certain scientifically-established facts about eyewitness
identification.”27 The report recommended that trial judges conduct pretrial hearings to determine whether
suggestive identification procedures were used, and if so, whether these procedures impaired the
reliability of identification evidence. Pretrial hearings would consider the effects of both estimator
variables (relating to viewing at the crime scene) and system variables (relating to the lineup or showup
procedures) on the identification. The report also recommended that the state adopt a set of recommended
practices for conducting identification procedures, create new model jury instructions on eyewitness
identifications, and set limitations on the admissibility of certainty statements and in-court
identifications.28

19

See New Jersey Criminal Model Jury Instructions, Identification, supra, at 2.
Id.
21
Id. at 9.
22
State v. Lawson, 352 Ore. 724 (Or. 2012).
23
Id. at 746–48.
24
Id. at 747–48, 755–56.
25
Id. at 759, 763.
26
See Massachusetts Supreme Judicial Court Study Group on Eyewitness Evidence, Report and
Recommendations to the Justices.
27
Id. at 48.
28
Id. at 28. In the courtroom, the eyewitness can easily see where the defendant is sitting. Thus, in-court
identifications do not reliably test an eyewitness’ memory. Nevertheless, courts have shown great tolerance of incourt identifications, deeming them based on “independent” memory, and even following suggestive out-of-court
procedures. Garrett, Eyewitnesses and Exclusion, supra. For example, the New York Court of Appeals ruled that
“[e]xcluding evidence of a suggestive showup does not deprive the prosecutor of reliable evidence of guilt. The
witness would still be permitted to identify the defendant in court if that identification is based on an independent
source.” People v. Adams, 423 N.E.2d 379, 384 (N.Y. 1981).
20

24
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
State Statutes Regulating Identification Procedures
Judicial rulings regulating admissibility of eyewitness evidence in the courtroom do not specify the
identification procedures to be used by law enforcement officials. However, fourteen states have adopted
legislation regarding eyewitness identification procedures. Of the fourteen, eleven states (Connecticut,
Illinois, Maryland, North Carolina, Ohio, Texas, Virginia, West Virginia, Wisconsin, Utah, and Vermont)
have enacted statutes directly requiring that law enforcement officials adopt written procedures for
eyewitness identifications and regulating the particular procedures to be used.29 Three more states
(Georgia, Nevada, and Rhode Island) have passed statutes recommending further study, tasking a group
with developing best practices, or requiring some form of written policy.30
State statutes typically assert that a trial judge may consider the failure to follow the prescribed
procedures as a factor in assessing admissibility and informing the jury. The statutes rarely require that a
trial judge exclude such identification evidence from consideration by the jury. However, some of the
more detailed statutes, such as those in Ohio, North Carolina, and West Virginia, require that law
enforcement officials use particular practices (e.g., eyewitness instructions, a blind administrator). Other
statutes require adherence to model policies or guidelines. Utah requires that lineup procedures be
recorded. Some jurisdictions and departments also have voluntarily adopted guidelines or policies
regulating eyewitness identifications.31 Several state courts have issued rulings regulating lineup practices
(e.g., New Jersey’s Supreme Court has required documentation of identification procedures).32
AIDING JURORS IN ASSESSMENT OF EYEWITNESS TESTIMONY
Expert Witness Testimony Regarding Eyewitness Identification
The standards for assessing the admissibility of testimony by expert witnesses have undergone great
changes in the past two decades. Before 1993, the Frye test allowed scientific expert testimony in federal
courts if it met the standard of “general acceptance” in the relevant scientific community. 33 In 1993, the
Supreme Court, in Daubert v. Merrell Dow Pharmaceuticals, Inc.,34 ruled that, under Federal Rule of
Evidence 702, a “trial judge must ensure that any and all scientific testimony or evidence admitted is not
only relevant, but reliable.”35 Judges determine reliability by assessing the scientific foundation of the
expert’s testimony prior to trial, so that “evidentiary reliability will be based upon scientific validity.”36
Many states have adopted Daubert, and many of those that have not formally adopted Daubert have
revised their Frye test to adopt much of the Daubert standard. In turn, Federal Rule of Evidence 702 has

29

See Conn. Gen. Stat. § 54-1p (West 2012); 725 Ill. Comp. Stat. § 5/107A-5 (West 2003); Md. Code Ann.,
Pub. Safety § 3-506 (West 2007); N.C. Gen. Stat. § 15A-284.52 (West 2007); Ohio Rev. Code Ann. § 2933.83
(West 2010); Tex. Code Crim. Proc. Ann. art. 38.20 (West 2011); Utah Code Ann. §77-8-4 (West 1980); Va. Code
Ann. §19.2-390.02 (West 2005); Va Code Ann. § 9.1-102.54; 13 V.S.A. § 5581; W. Va. Code Ann. § 62-1E-1 (West
2013); Wis. Stat. § 175.50 (West 2005).
30
GA. H.R. 352, 149th Gen. Assem., Reg. Sess. (April 20, 2007); Nev. Rev. Stat. § 171.1237 (West 2011); R.I.
Gen. Laws § 12-1-16 (West 2012); 2010 Leg. Reg. Sess. (Vt. 2010).
31
See, e.g., John J. Farmer, Jr., Attorney General of the State of New Jersey, “Letter to All County Prosecutors:
Attorney General Guidelines for Preparing and Conducting Photo and Live Lineup Identification Procedures” (April
18, 2001), available at: http://www.state.nj.us/lps/dcj/agguide/photoid.pdf; CALEA Standards for Law Enforcement
Agencies: 42.2.11 Lineups, available at: http://www.calea.org/content/standards-titles; International Association of
Chiefs of Police, Model Policy: Eyewitness Identification (2010).
32
State v. Delgado, 188 N.J. 48, 63–64, 902 A.2d 888 (2006).
33
Frye v. United States, 54 App. D.C. 46, 293 F. 1013 (1923).
34
509 U.S. 579 (1993).
35
Id. at 589.
36
Id. at 590 n.9.

25
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
been revised to incorporate the holding in Daubert.37 Federal and state courts remain divided on whether
expert testimony on eyewitness identifications is admissible under Daubert or Frye, and on the proper
exercise of trial court discretion when deciding whether to admit such expert testimony. Appellate rulings
emphasize that a trial judge should use discretion when deciding whether proffered expert evidence
satisfies the Daubert or Frye standards. An increasing number of rulings emphasize the value of
presenting expert testimony regarding eyewitness identification. Some courts have held that it can be an
abuse of discretion for a trial judge to bar the defense from admitting such testimony.38 Detailed
descriptions of the relevant scientific research findings accompany such decisions.39 There also are many
federal and state courts that continue to follow the traditional approach, emphasizing that credibility of
eyewitnesses is a matter within the “province of the jury” and insisting that information regarding valid
scientific research in this area will not assist the jury in its task.40
The trend is toward greater acceptance of expert testimony regarding the factors that may affect
eyewitness identification. In a 2012 decision, the Connecticut Supreme Court disavowed earlier rulings
restricting expert testimony and stated that such rulings are now “out of step with the widespread judicial
recognition that eyewitness identifications are potentially unreliable in a variety of ways unknown to the
average juror.”41 Similarly, the Pennsylvania Supreme Court recently held that expert testimony on
eyewitness identifications was no longer per se inadmissible, emphasizing that “courts in 44 states and
the District of Columbia have permitted such testimony at the discretion of the trial judge,” and that “all
federal circuits that have considered the issue, with the possible exception of the 11th Circuit, have
embraced this approach.”42 As the Seventh Circuit Court of Appeals recently explained:
It will not do to reply that jurors know from their daily lives that memory is fallible. The question
that social science can address is how fallible, and thus how deeply any given identification
should be discounted. That jurors have beliefs about this does not make expert evidence
irrelevant; to the contrary, it may make such evidence vital, for if jurors’ beliefs are mistaken then
they may reach incorrect conclusions. Expert evidence can help jurors evaluate whether their
beliefs about the reliability of eyewitness testimony are correct.43
37

Fed. R. Evid. 702. Rule 702 now provides:
A witness who is qualified as an expert by knowledge, skill, experience, training, or education may testify in the
form of an opinion or otherwise if: (a) the expert’s scientific, technical, or other specialized knowledge will help the
trier of fact to understand the evidence or to determine a fact in issue; (b) the testimony is based on sufficient facts
or data; (c) the testimony is the product of reliable principles and methods; and
(d) the expert has reliably applied the principles and methods to the facts of the case.
38
See, e.g., Tillman v. State, 354 S.W.3d 425, 441 (Tex. Crim. App. 2011); People v. LeGrand, 835 N.Y.S.2d
523, 524 (2007); State v. Clopten, 223 P.3d 1103, 1117 (Utah 2009); U.S. v. Smithers, 212 F.3d 306, 311–14 (6th
Cir.2000).
39
See, e.g., State v. Copeland, 226 S.W.3d 287, 299-300 (Tenn.2007); Tillman, 354 S.W.3d at 441; Clopten, 223
P.3d at 1108.
40
For scholarly examination of this case law, see, e.g., “The Province of the Jurist: Judicial Resistance to Expert
Testimony on Eyewitnesses as Institutional Rivalry,” Harvard Law Review 126 (8): 2381 (2013); R. Simmons,
“Conquering the Province of the Jury: Expert Testimony and the Professionalization of Fact-Finding,” University of
Cincinnati Law Review 74: 1013 (2006); G. Vallas, “A Survey of Federal and State Standards for the Admission of
Expert Testimony on the Reliability of Eyewitnesses,” American Journal of Criminal Law 39 (1): 97 (2011).
41
State v. Guilbert, 306 Conn. 218, 234 (Conn. 2012). Prior to that decision, the Connecticut Supreme Court had
long ruled that “the reliability of eyewitness identification is within the knowledge of jurors and expert testimony
generally would not assist them in determining the question” (State v. Kemp, supra, 199 Conn. at 473, 477), and that
factors affecting eyewitness memory are “nothing outside the common experience of mankind” (State v.
McClendon, supra, 248 Conn. at 572, 586).
42
Com. v. Walker, 2014 WL 2208139 *13 (Pa. 2014) (collecting authorities).
43
U.S. v. Bartless, 567 F.3d 901, 906 (7th Cir. 2009). Other federal courts have found it a proper exercise of
discretion to exclude expert testimony on eyewitness identifications. See, e.g., United States v. Lumpkin, 192 F.3d
280, 289 (2d Cir. 1999). Most federal courts treat the subject as one of considerable trial discretion; see, e.g. United

26
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Courts also have allowed expert witnesses to testify about particular issues concerning eyewitness
identifications, such as cross-race effects, stress, weapons focus, suggestive lineup procedures, and the
like.44 Rarely have experts conducted eyewitness identification research related to the specific case before
the court. However, in one such case, in which an experiment was conducted with the actual photo array
used in the case, the federal courts found expert testimony admissible where it was directed not just to
general research, but to the question of whether suggestive procedures affected the identification in that
case.45
Expert witnesses who explain the complications of eyewitness identification can be expensive. Most
criminal defendants are indigent and cannot afford such assistance.46 In Ake v. Oklahoma, the Supreme
Court held that an indigent defendant has a constitutional due process right to assistance by an expert
witness only if that expert assistance is so crucial to the defense (or such a “significant factor”) that its
denial would deprive the defendant of a fundamentally fair trial.47 In federal courts, funding for expert
witnesses is available, and requests by indigent defendants are common.48 In state courts, such assistance
is uncommon, especially in state courts that rarely find denial of expert assistance on eyewitness matters
to be a due process violation.
Expert testimony on eyewitness memory and identifications has many advantages over jury
instructions as a method to explain relevant scientific framework evidence to the jury: (1) Expert
witnesses can explain scientific research in a more flexible manner, by presenting only the relevant
research to the jury; (2) Expert witnesses are familiar with the research and can describe it in detail; (3)
Expert witnesses can convey the state of the research at the time of the trial; (4) Expert witnesses can be
cross-examined by the other side; and (5) Expert witnesses can more clearly describe the limitations of
the research. The benefits of expert testimony are offset somewhat by the expense. In addition, conflicting
testimony by opposing experts may lead to confusion among the jurors. Trial judges have discretion to
determine whether the potential benefits of expert testimony outweigh the cost.
Jury Instructions Regarding Eyewitness Identification
Some courts restricting expert testimony have found jury instructions regarding the fallible nature of
eyewitness identifications to be an acceptable substitute for expert testimony.49 At the conclusion of a
criminal trial, the trial judge can instruct jurors on the factors that may result in an erroneous
identification while also offering instructions on the legal principles jurors must apply when assessing the
factual record. Such instructions may be given when the witness testifies. Judges tend to rely on model or
pattern instructions, because any departure from these standard instructions may be a ground for appellate
reversal.

States v. Rodriguez-Berrios, 573 F.3d 55, 71–72 (1st Cir. 2006). For a survey of federal decisions, see Lauren
Tallent, Note, Through the Lens of Federal Evidence Rule 403: An Examination of Eyewitness Identification Expert
Testimony Admissibility in the Federal Circuit Courts, Washington & Lee Law Review 68 (2): 765 (2011); see also
Walker, 2014 2208139 *13.
44
See, e.g., Loftus, Doyle & Dysart at § 14-8[a]-[b] p.408 n. 41–42, 410, n. 53 (5th Edition, 2013) (collecting
cases).
45
Newsome v. McCabe, 319 F.3d 301 (7th Cir.2003).
46
See, e.g., Bureau of Justice Statistics, “Indigent Defense,” available at:
http://www.bjs.gov/index.cfm?ty=pbdetail&iid=995.
47
470 U.S. 68, 82–83 (1985). Even if an indigent defendant receives funding to retain an expert, the judge may
ultimately decide that the expert testimony is not admissible at trial.
48
18 U.S.C. § 3006A(e)(1).
49
See, e.g., U.S. v. Jones, 689 F.3d 12, 20 (1st Cir. 2012) (“The judge was fully entitled to conclude that this
general information could be more reliably and efficiently conveyed by instructions rather than through dueling
experts.”).

27
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
The New Jersey Supreme Court viewed jury instructions as preferable to expert testimony. 50 The
New Jersey instructions adopted, following the Henderson decision, are by far the most detailed set of
jury instructions regarding eyewitness identification evidence. Traditionally, instructions regarding
eyewitness identifications have been brief, and remind the jurors to consider the following: (1) the
credibility of an eyewitness is like that of any other witness, and (2) any eyewitness identification is part
of the prosecutor’s burden of proof in a criminal case.51 Many state courts have held that, although
general jury instructions regarding credibility and the burden of proof are appropriate, more specific
instructions on eyewitness identifications are considered an inappropriate judicial comment on the
evidence.52 Following the U.S. Supreme Court’s decision in Manson v. Brathwaite, some state courts
supplemented their jury instructions by including the five reliability factors named by the Supreme
Court.53
In 1972, in U.S. v. Telfaire, the D.C. Circuit Court of Appeals adopted a set of influential model jury
instructions to be used in appropriate federal cases involving eyewitness identifications.54 The instructions
emphasized the following:
You must consider the credibility of each identification witness in the same way as any other
witness, consider whether he is truthful, and consider whether he had the capacity and opportunity to
make a reliable observation on the matter covered in his testimony.55
The Telfaire instructions departed from the brief traditional instruction by adding that the jury
should consider factors related to the initial sighting, including “how long or short a time was available,
how far or close the witness was, how good were lighting conditions, [and] whether the witness had had
occasion to see or know the person in the past.” The decision also noted that an identification is more
reliable if the witness is able to pick the defendant out of a group, rather than at a showup, and that the
jury should consider the length of time between the crime and the identification.56
Some states have adopted cautionary instructions on specific issues related to eyewitness
identification evidence. In State v. Ledbetter, the Connecticut Supreme Court ordered lower courts to use
a special instruction in cases in which law enforcement failed to instruct the eyewitness that the
perpetrator may or may not be present in a lineup.57 The Georgia Supreme Court concluded in 2005 that
50

The New Jersey Supreme Court indicated: “Jury charges offer a number of advantages: they are focused and
concise, authoritative (in that juries hear them from the trial judge, not a witness called by one side), and cost-free;
they avoid possible confusion to jurors created by dueling experts; and they eliminate the risk of an expert invading
the jury’s role or opining on an eyewitness’ credibility.” Henderson, 27 A.3d at 925.
51
New Jersey courts used such instructions a decade before Henderson. See, e.g., State v. Robinson, 165 N.J.
32, 46–47 (N.J. 2000). Some states have also approved instructions informing the jury that there may be an
“independent source” for an in-court identification. See, e.g. State v. Cannon, 713 P.2d 273, 281 (Ariz. 1985).
52
Brodes v. State, 279 Ga. 435, 439 & n.6 (Ga. 2005) (surveying state caselaw).
53
State v. Tatum, 219 Conn. 721 (1991).
54
U.S. v. Telfaire, 469 F.2d 552, 558 (D.C. Cir. 1972). Some federal courts follow that approach, while others
adopt a “flexible approach.” See, e.g., United States v. Luis, 835 F.2d 37, 41 (2d Cir.1987). Some more recent
federal model instructions include added detail, reflecting variables such as stress and cross-race identifications.
See, e.g., Third Circuit Model Criminal Jury Instructions, 4.15 (Jan. 2014), at
http://www.ca3.uscourts.gov/sites/ca3/files/2013%20Chapter%204%20final%20revised.pdf.
55
U.S. v. Telfaire, 469 F.2d at 559.
56
Id. at 558.
57
State v. Ledbetter, 275 Conn. 534, 579–80 (2005) (The instruction reads, in part, “the individual conducting
the procedure either indicated to the witness that a suspect was present in the procedure or failed to warn the witness
that the perpetrator may or may not be in the procedure. Psychological studies have shown that indicating to a
witness that a suspect is present in an identification procedure or failing to warn the witness that the perpetrator may
or may not be in the procedure increases the likelihood that the witness will select one of the individuals in the
procedure, even when the perpetrator is not present. Thus, such behavior on the part of the procedure administrator
tends to increase the probability of a misidentification.”).

28
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
one particular use of the Manson v. Brathwaite factors must no longer be permitted: “we can no longer
endorse an instruction authorizing jurors to consider the witness’ certainty in his/her identification as a
factor to be used in deciding the reliability of that identification.”58 Other courts have done the same.59 In
1999, the New Jersey Supreme Court ruled in State v. Cromedy that instructions on cross-racial
identifications are required in certain cases.60
Expert testimony on eyewitness memory and identifications appears to have many advantages when
used as a method to explain relevant scientific framework evidence to the jury. However, when expert
testimony is not available to the defense, jury instructions may be a preferable alternative means to inform
the jury of the findings of scientific research in this area. Brief instructions may not, however, provide
sufficient guidance to explain the relevant scientific evidence to the jury, but lengthy instructions may be
cumbersome and complex.
More research is warranted to better understand how best to communicate to jurors the factors that
may affect the validity of eyewitness testimony and support a more sensitive discrimination of the
strengths and weaknesses of eyewitness testimony in individual cases. Indeed, research findings on the
effectiveness of jury instructions on assessment of eyewitness identification evidence have been mixed. In
general, such studies find that jury instructions cause jurors to become more suspicious of all eyewitness
identification evidence.61 A recent study of the effect of the New Jersey jury instructions used in
Henderson found that the instructions reduced juror reliance on both strong and weak eyewitness
identification evidence.62 Among the few studies finding that jury instructions succeed in increasing
jurors’ sensitivity to the strength of such evidence63 are those that study the effect of jury instructions
presented before the eyewitness testimony rather than at the end of the case before deliberation. Such
studies also have examined instructions that use visual aids rather than rely on a judge’s recitation of
written instructions.64 In addition, research studies might explore the use of videotape as an alternative
way to present such information65 and the effects of moving jury instructions to precede the introduction
of the testimony by the eyewitness.

58

Brodes, 279 Ga. at 442.
See supra; see, e.g., Commonwealth v. Payne, 426 Mass. 692 (1998); State v. Romero, 191 N.J. 59 (2007).
60
State v. Cromedy, 158 N.J.112 (1999); see also Innocence Project, Know the Cases: McKinley Cromedy, at
http://www.innocenceproject.org/Content/McKinley_Cromedy.php.
61
For a review of this research, see K. A. Martire and R. I. Kemp, “The Impact of Eyewitness Expert Evidence
and Judicial Instruction on Juror Ability to Evaluate Eyewitness Testimony,” Law and Human Behavior 33:225–36,
226 (reviewing studies of jury instructions on eyewitness identification and concluding that increased skepticism
and confusion is a common result); see also J. L. Devenport, C. D. Kimbrough, and B. L. Cutler, “Effectiveness of
traditional safeguards against erroneous conviction arising from mistaken eyewitness identification,” in Expert
testimony on the psychology of eyewitness identification, ed. B. L. Cutler (New York: Oxford University Press,
2009), 51–68 (summarizing research studying the Telfair jury instruction and concluding that “cautionary jury
instructions may be an ineffective safeguard against erroneous convictions resulting from mistaken eyewitness
identifications.”).
62
A. P. Papailiou, D. V. Yokum, C. T. Robertson, “The Novel New Jersey Eyewitness Instruction Induces
Skepticism But Not Sensitivity,” August 2014, available at:
http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2475217.
63
See, e.g., N. B. Pawlenko, M. A. Safer, R. A. Wise, and B. Holfeld, “A Teaching Aid for Improving Jurors’
Assessments of Eyewitness Accuracy,” Applied Cognitive Psychology 27(2): 190–97. Other studies are reviewed in
Martire and Kemp, supra note 105 at 226.
64
Pawlenko et al., supra note 107.
65
For an example of videotaped instructions, see Federal Judicial Center, The Patent Process: An Overview for
Jurors, available at: http://www.youtube.com/watch?v=ax7QHQTbKQE.
59

29
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
CONCLUSION
The Manson v. Brathwaite test under the Due Process Clause of the U.S. Constitution set out the
modern test that regulates the fairness and the reliability of eyewitness identification evidence. The test
evaluates the “reliability” of eyewitness identifications using factors derived from prior rulings and not
from empirically validated sources. It includes factors that are not diagnostic of reliability and treats
factors such as the confidence of a witness as independent markers of reliability when, in fact, it is now
well established that confidence judgments may vary over time and can be powerfully swayed by many
factors. The best guidance for legal regulation of eyewitness identification evidence comes not, however,
from constitutional rulings, but from the careful use and understanding of scientific evidence to guide
fact-finders and decision-makers.

30
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

4
Basic Research on Vision and Memory
Accurate eyewitness identification requires that a witness to a crime correctly sense, perceive, and
remember objects and events that occurred and recall them later. The veracity of the witness’
identification thus depends on the limits of sensation, perception, and memory. Recent scientific studies
have yielded great advances in our understanding of how vision and memory work. This chapter provides
a brief overview of current knowledge, identifies areas in which vision and memory are imperfect, and
describes implications for the accuracy of eyewitness identification. These implications, in turn, have
guided much of the applied research on this topic (see Chapter 5) and provide a general framework for the
recommendations made herein (see Chapter 6).
VISION AND MEMORY IN CONTEXT
We begin this chapter by offering a concrete example to place the body of basic scientific research
on vision and memory in context and so as to better communicate its relevance to eyewitness
identification. In the sections that follow the example, we dissect the different functional steps of the
sequence (highlighted in italics) in some detail, with special reference to its limitations and the ways in
which it may fail to deliver accurate eyewitness identification.
While returning home late, you hear a muffled scream from around the street corner. Seconds later,
you come face-to-face with a man turning the corner and moving swiftly past you. Instantaneously,
properties of the scene are conveyed to you through patterns of light cast on the backs of your eyes
and sensed by photoreceptors in your retina. Only a fraction of the information sensed is selected for
further processing; in this case you focus your attention on certain features of the man’s face. Those
features are integrated and interpreted to yield a coherent percept of the man. As you round the
corner, you perceive, through an identical process, the victim slumped lifelessly against the wall.
You quickly grasp the meaning of these perceptual experiences and they immediately elicit both
cognitive and visceral components (e.g., increased heart rate) of fear and anxiety. Your percepts are
initially encoded in short-term working memory, where content is limited and labile. Your elevated
level of arousal may cause interference and some loss of content, but with time and recognition of
the importance of the experience, your percepts are consolidated into long-term memory. Long-term
memories are maintained in storage but subject to ongoing updates and modifications resulting from
new experiences and perhaps distortions caused by sustained levels of stress. All of these processes
of perception and memory take place while you experience strong emotions which greatly influence
what you see and remember.
At a later date, you are asked to look at a police lineup that includes a suspect apprehended near the
crime scene. Visual features of the men in the lineup are sensed, selectively attended, and perceived,
using the same visual processes engaged on the night in question. Some of these features—the high
brow and sharp cheekbones of one man in the lineup—elicit retrieval of memories of your visual
experiences on the night of the crime. The simultaneously perceived and retrieved experiences are
implicitly compared, leading to a cycle of greater visual scrutiny of the man in front of you and
retrieval of additional details of the original percept. The context of the lineup procedure, the sight
of the man, and the retrieved memories trigger latent emotions and anxiety, which may interfere
with your comparison of percept and memory. Eventually, the comparison reaches your internal

31
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
criterion for identification: You decide, with an implicit level of certainty, that your current visual
percept and the percept from the night of the crime were caused by the same external source (the
man now in front of you), and you assert that you have identified the person you witnessed at the
crime scene.
VISION
Functional Processes of Vision
To understand the contributions and limitations of vision to eyewitness identification, it is useful to
consider the workings of three functional stages of visual processing—sensation, attention, and
perception—bearing in mind that they comprise highly interdependent elements of a continuous
operation. Sensation is the initial process of detecting light and extracting basic image features.
Sensations themselves are evanescent, and only a small fraction of what is sensed is actually perceived.
Attention is the process by which information sensed by the visual system is selected for further
processing. Perception is the process by which attended visual information is integrated, linked to
environmental cause, made coherent, and categorized through the assignment of meaning, utility, value,
and emotional valence. In addition, memories and emotions resulting from prior experiences with the
world can influence all stages of visual processing and thus define a thread that weaves throughout the
following discussions.
All of the functional processes of vision are beset by sources of noise, which affects the quality and
types of information accessible from the visual environment, and bears heavily on the validity of
eyewitness identification. Before considering the processes of sensation, attention, and perception in
greater detail, we first introduce the concept of noise in visual processing and ways of interpreting its
impact on visual experience.
The Fundamental Role of Noise
Vision is usefully understood as the process of detecting informative signals about the external
world and using those signals to recognize objects, make decisions, and guide behavior. As with any
signal detection, there are occasionally factors present that lead to uncertainty on the part of the observer
about whether a particular signal is present. These factors are generically termed noise, following the
definition used in electronic signal transmission, in which noise refers to random or irrelevant elements
that interfere with detection of coherent and informative signals. In vision, noise comes from a variety of
sources, some associated with the structure of the visual environment (e.g., occluding surfaces, glare,
shadows), some inherent to the optical and neuronal processes involved (e.g., scattering of light in the
eye), some reflecting sensory content not relevant to the observer’s goals (e.g., a distracting sign or a loud
noise), and some originating with incorrect expectations derived from memory. Consider, for example,
the seemingly simple problem of detecting a green light while waiting at a traffic signal. In this case, your
ability to “see” the green light may be compromised by glare or dust on your windshield, by poor visual
acuity, by your eyes having been aimed instead at the driver of the adjacent car, by the presence of other
(irrelevant) colored lights in your field of view (e.g., a traffic signal at a different intersection or the lights
of a nearby restaurant), or by a cell phone conversation or the news on the car radio. The significance of
this view for eyewitness identification is profound, as it helps us to realize that the accuracy of
information about the environment—the face of a criminal, for example—gained through vision is
necessarily, and often sharply, limited by noise.1
1

W. S. Geisler, “Sequential ideal-observer analysis of visual discriminations,” Psychological Review 96 (2):
267–314 (1989).

32
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
The fact that vision is noise-limited suggests a familiar statistical framework—signal detection
theory—for assessing and understanding the effects of noise on visual perception and recognition ability.2
Signal detection theory has long been successfully applied to analogous problems in electronic signal
reception.3 To illustrate these principles as applied to sensory processing, consider the problem of
detecting the vibrating cell phone in your pocket. Anyone who has operated a cell phone in vibrate mode
will be familiar with two types of signal detection errors: (1) the occasional sense that the phone is
vibrating in your pocket, only to discover that it is not, and, conversely, (2) the phone call that is
sometimes missed because you attribute the vibration to some other cause. Signal, in this example, is a
subtle tactile stimulus resulting from an incoming phone call. Noise, in this example, is all of the other
things in your environment that may also lead to subtle tactile stimulation, such as vibration of your car
seat, a shift of the keys in your pocket, or the touch of another person.
Signal detection theory posits that there are three main factors that determine whether a signal will
be detected: (1) the distribution of stimuli (e.g., the variety of stimulus magnitudes) that reflect noise only,
(2) the distribution of stimuli that reflect signal, and (3) the observer’s criterion for “deciding” that a
specific stimulus resulted from noise sources or signal. An important factor for the fidelity of signal
detection is the degree to which noise and signal distributions overlap with one another. In the case of the
vibrating cell phone, if the distributions of tactile stimuli resulting from noise and signal overlap, as is
often the case, then there will always be some cases in which you believe the phone is vibrating when it is
not (noise stimuli attributed to signal source), and there will be some cases in which the phone is vibrating
and you miss the call (signal stimuli attributed to noise source).
The third factor that influences signal detection in the presence of noise is the observer’s decision
criterion, which is simply the value (e.g., stimulus amplitude) above which a stimulus is attributed to
signal, and below which a stimulus is attributed to noise. In the same sense that your car radio is
programmed to “decide” (and allow you to hear) when informative patterns of electromagnetic radiation
(signal) are sufficiently different from random fluctuations (noise), an observer adopts a criterion for
deciding whether a stimulus is caused by a signal or is simply a manifestation of noise. This criterion
reflects the level of precision acceptable for the observer’s needs, given uncertainty about whether a given
stimulus reflects a real signal.
In practice, the criterion4 used is determined by a host of factors unique to the circumstances,
including psychological and social demands and behavioral goals. These factors collectively determine
the relative “costs” of incorrect attributions of signal as noise (“misses”) and of noise as signal (“false
alarms”).
If an individual places high value on not missing a phone call, then she or he will adopt a very
liberal criterion, in which all stimuli reflecting real incoming calls (signal) are successfully detected, but
many noise stimuli (e.g., shifting keys in pocket) are erroneously (and frustratingly) believed to be
incoming calls. By contrast, if an individual places little value on detecting incoming phone calls, she or
he will adopt a conservative criterion, in which many calls are missed and noise stimuli rarely elicit an
effort to answer the phone, which may be of value to the individual who wishes to avoid distraction.
The example of the signal detection logic used for the vibrating cell phone applies similarly to all
aspects of visual perceptual experience, including the conditions of witnessing criminal events. The
uncertainty about visual events caused by manifold sources of noise will inevitably lead to inaccurate
visual perceptual experiences, which result from conditions in which an observer fails to detect a critically
informative stimulus as “real” (attributing the stimulus instead to a source of noise), or confidently
perceives a noise stimulus to have originated from an informative source. The latter instance is
problematic because it increases the likelihood that observers will unwittingly “construct,” on the basis of
2

D. M. Green and J. A. Swets, Signal Detection Theory and Psychophysics (New York: Wiley,1966).
W. W. Peterson, T. G. Birdsall, and W. C. Fox, “The theory of signal detectability,” Proceedings of the IRE
Professional Group on Information Theory 4 (4): 171–212 (1954).
4
The criterion is sometimes referred to as bias.
3

33
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
expectations derived from memory and situational context, perceptual experiences to account for noise
erroneously interpreted as signal.
What follows from this consideration of uncertainty and decision criteria for visual perception is that
the actual impact of factors that limit the amount of visual information available to an eyewitness (factors
considered in more detail below) will depend on the criterion adopted. The criterion may reflect the
values and prejudices of the eyewitness, his or her motivational and emotional state, and a variety of
behavioral goals. In principle, the observer’s criterion can be altered by instruction or incentives, but it is
important to note that the criterion held by an observer witnessing a crime scene cannot be anticipated,
nor can it be altered after the fact. It is an “estimator variable,” which simply needs to be recognized and
understood when evaluating eyewitness reports. By contrast, the decision criterion held by an observer at
the time of identification can be controlled, and there may be valid reasons for doing so (see Chapter 5).5
In the following discussions of sensation, attention, and perception, we address the various means
and conditions under which many different types of noise introduce uncertainty in visual signal detection
and thus fundamentally limit the accuracy of eyewitness identification.
Visual Sensation
When an observer views an object of any sort (such as a person) or events involving the object (a
criminal act), patterns of light reflected from the environment are focused by the lens at the front of the
eye and projected onto the back surface of the eye (the retina) to form the retinal image. Light in the
image is initially “sensed” by activation of photoreceptors, and early stages of sensory processing
function to detect spatial and temporal contrast along a number of dimensions, including intensity and
wavelength of light.6 These contrast measurements are integrated by subsequent processing stages in the
brain to yield representations of basic image features, or primitives, such as oriented image contours.7
Several sources of noise, or factors that limit the ratio of signal to noise, can restrict the visual
information accessible to these early sensory processes. Some factors are inherent to the visual system
and largely uncontrollable, such as scattering of light by the fluid and tissues of the eye, and can be
exacerbated by common observer-specific visual deficits, such as myopia, poor contrast sensitivity, or
color blindness. Others factors are dependent on viewing conditions, such as the effects of viewing time
and level of illumination.8 Both of these factors predictably influence the quantity of information—the
visual signal strength—that a viewer gains from a visual scene, and thus the degree to which the
perceptual experience can accurately reflect the properties of the external world.9 At the extreme, short
viewing times and low levels of illumination simply reduce the number of correlated photons reaching the
5

L. Mickes, H. D. Flowe, and J. T. Wixted, “Receiver operating characteristic analysis of eyewitness memory:
Comparing the diagnostic accuracy of simultaneous and sequential lineups,” Journal of Experimental Psychology:
Applied, 18 (4): 361–76 (2012).
6
M. Meister and M. Tessier-Lavigne, “Low-level Visual Processing: The Retina,” in Principles of
Neuroscience, 5th Edition, ed. E. Kandel, J. H. Schwartz, T. M. Jessell, S. A. Siegelbaum, and A.J. Hudspeth (New
York: McGraw-Hill Professional, 2012), 577–601.
7
C. D. Gilbert, “Intermediate-level Visual Processing and Visual Primitives,” in Principles of Neuroscience, 5th
Edition, ed. E. Kandel, J. H. Schwartz, T. M. Jessell, S. A. Siegelbaum, and A. J. Hudspeth (New York: McGrawHill Professional, 2012), 602–620.
8
D. G. Pelli, “Uncertainty explains many aspects of visual contrast detection and discrimination” Journal of the
Optical Society of America A2 (9): 1508–32 (1985). D. G. Pelli, “The quantum efficiency of vision,” in Vision:
Coding and Efficiency, ed. C. Blakemore (Cambridge: Cambridge University Press, 1990), 3–24. G. Sperling, “The
information available in brief visual presentations,” Psychological Monographs: General and Applied, 74 (11,
Whole No. 498): 1–29 (1960).
9
G. Sperling, “A signal-to-noise theory of the effects of luminance on picture memory: Comment on Loftus,”
Journal of Experimental Psychology: General, 115 (2): 189–92 (1986).

34
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
retina to the point where they scarcely exceed photon noise,10 and uncertainty is very high. At slightly
longer viewing times and greater illumination levels, signal-to-noise levels improve, but there may remain
marked limits on visual sensitivity. Visual acuity, for example, which is a measure of the ability to
resolve the fine spatial details of a visual pattern, is known to decline significantly with decreases in
illumination.11
Signal-to-noise loss can depend on the direction of the observer’s gaze. Visual acuity is highest at
the observer’s center of gaze. The center is the part of your visual system that is used for fine sensing,
such as reading or scrutinizing faces in a social context. Acuity drops off markedly with angular distance
from this center, such that the quality and quantity of information sensed a mere 10 degrees from center
are far less than what is available at the center of gaze.12
Under unrestricted viewing conditions, the movements of the eyes largely overcome the effects of
gaze direction. However, under the viewing conditions associated with a typical crime, this source of
noise may place severe limitations on the ability of the observer to sense key pieces of information that
are not present at the center of gaze. To appreciate the impact of these limitations, consider that patients
with macular degeneration are effectively blinded in the region of the visual field possessing highest
acuity, and must rely instead on the much-reduced quality of visual information gained from the
peripheral visual field. To compensate for this clinical loss, images and text must be greatly magnified to
enable comprehension—a tactic that is clearly not available to an eyewitness.
Visual Attention
Light falling on all parts of the retina is available to be sensed—and must be sensed for it to be
available for further processing—but only a small fraction of the information sensed reaches awareness or
is used by the observer for recognition, action, or storage in memory. This limited access to visual sensory
information is a product of selective attention.13 Attention is an active process that can be directed by
external factors—visual attributes with high salience, such as a bright light or an unfamiliar object—or
internal control.14 If you are searching for the coffee cup, for example, you may explicitly direct your
attention to the table where it was last seen. Attention can be directed to different types of image content,
including specific locations in space15 and specific image features (such as a specific color),16 or to
specific objects (such as the coffee cup).17
Attended image content is transiently enhanced to increase the fidelity of visual experience.18
Attention interacts with sensory processing, for example, by selectively enhancing contrast19 and
10

S. Hecht, S. Schlaer, and M. H. Pirenne, “Energy, quanta, and vision,” Journal of General Physiology 25 (6):
819–840 (1942).
11
P. W. Cobb, “The influence of illumination of the eye on visual acuity,” American Journal of Physiology, 29:
76–99 (1911). S. Hecht, “A quantitative basis for the relation between visual acuity and illumination,” Proceedings
of the National Academy of Sciences, 13: 569–574 (1927). S. Shlaer, “The relation between visual acuity and
illumination,” Journal of General Physiology, 21 (2): 165–188 (1937).
12
H. Strasburger, I. Rentschler, and M. Jüttner, “Peripheral vision and pattern recognition: a review,” Journal of
Vision, 11(5):13, 1–82 (2011).
13
W. James, Principles of Psychology (New York: Henry Holt, 1890); H. Pashler, J. Johnston, and E. Ruthruff,
“Attention and Performance,” Annual Review of Psychology 52: 629–651 (2001).
14
M. I. Posner, “Orienting of attention,” Quarterly Journal of Experimental Psychology 32: 3–25 (1980).
15
Ibid.
16
A. F. Rossi and M. A. Paradiso, “Feature-specific effects of selective visual attention,” Vision Research 35 (5):
621–634 (1995).
17
J. Duncan, “Selective attention and the organization of visual information,” Journal of Experimental
Psychology: General 113 (4): 501–517 (1984).
18
H. Pashler, J. Johnston, and E. Ruthruff, “Attention and Performance,” Annual Review of Psychology 52: 629–
651 (2001).

35
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
potentially overcoming low signal-to-noise levels resulting from limited viewing time or illumination.20
The effects of attention on contrast enhancement can be potentiated further when attention is commanded
by emotionally laden stimuli.21 Image content not falling within the focus of attention is processed with
less fidelity.22 In some cases, unattended content is effectively invisible:23 it does not reach awareness, it
is not perceived, and it is not available for use in guiding decisions or actions, or for storage in memory.
Different pieces of visual information compete for selection,24 as their attributes of physical
salience, location in space, novelty, and relevance to the observer’s needs and behavioral goals are always
changing.25 The outcome of the competition is highly susceptible to noise (in this instance, noise is
defined as uncontrolled factors that bias the focus of attention and create uncertainty about the content of
a visual scene), because the informational content of the visual image vastly exceeds what can be attended
at any point in time. The implications of such noise for eyewitness identification are profound. An
observer must “select” what to attend to, often within a short window of time, without advance warning,
in the presence of many novel objects and events, and under the confounding influences of anxiety and
fear.
The signal detection framework is readily adaptable to the problem of noise in visual attention and
provides some insights into the limits of attentional selection in the presence of noise.26 In essence, this
signal detection approach quantifies the extent to which multiple items competing with one another for
attention affect attentional enhancement for any one of the items.27 Reductions in efficiency are common
under such noise conditions. Indeed, sensitivity to unattended items can be markedly reduced under
conditions of high “perceptual load,” in which there are many objects simultaneously competing for
19

M. Carrasco et al.,“Attention alters appearance” Nature Neuroscience 7: 308–313 (2004).
M. I. Posner, C. R. Snyder, B. J. Davidson, “Attention and the Detection of Signals,” Journal of Experimental
Psychology 109(2):160–74 (1980). M. Carrasco and B. McElree, “Covert attention accelerates the rate of visual
information processing,” Proceedings of the National Academies of Science U. S. A. 98 (9): 5363–5367 (2001). Y.
Yeshurun and M. Carrasco, “Attention improves or impairs visual performance by enhancing spatial resolution,”
Nature 396, 72–75 (1998). M. Carrasco et al., “Covert attention increases spatial resolution with or without masks:
support for signal enhancement,” Journal of Vision 2 (6): 467–79 (2002). E. Blaser et al., “Measuring the
amplification of attention,” Proceedings of the National Academies of Science U.S.A. 96 (20): 11681–11686 (1999).
K. Anton-Erxleben and M. Carrasco, “Attentional enhancement of spatial resolution: linking behavioural and
neurophysiological evidence,” Nature Reviews Neuroscience 14(3):188–200 (2013). J. W. Couperus and G. R.
Mangun, “Signal enhancement and suppression during visual-spatial selective attention,” Brain Research
1359:155–177 (2010).
21
E. A. Phelps, S. Ling, and M. Carrasco, “Emotion facilitates perception and potentiates: the perceptual
benefits of attention,” Psychological Science 17(4): 292 (2006).
22
M. I. Posner, C. R. Snyder, and B. J. Davidson, “Attention and the Detection of Signals,” Journal of
Experimental Psychology 109(2):160–74 (1980). Y. Yeshurun and M Carrasco, “Attention improves or impairs
visual performance by enhancing spatial resolution,” Nature 396: 72–75 (November 1998).
23
A. Mack and I. Rock, Inattentional blindness (Cambridge, MA: MIT Press, 1998).
24
R. Desimone and J. Duncan, “Neural Mechanism of Selective Visual Attention,” Annual Review of
Neuroscience 18: 193–222 (March 1995).
25
J. M. Wolfe and T. S. Horowitz, “What attributes guide the deployment of visual attention and how do they do
it?” Nature Reviews Neuroscience 5: 495–501 (June 2004). H. E. Egeth and S.Yantis, “Visual attention: Control,
representation, and time course,” Annual Review of Psychology 48 (1): 269–297 (February 1997). M. I. Posner,
“Orienting in attention,” Quarterly Journal of Experimental Psychology 32 (1): 3–25 (1980). A. Treisman and G.
Gelade, “A feature integration theory of attention,” Cognitive Psychology 12 (1):97–136 (January 1980). L. Itti and
C. Koch, “A saliency-based search mechanism for overt and covert shifts of visual attention,” Vision Research
40(10–12):1489–1506 (June 2000).
26
G. Sperling and M. J. Melchner, “The attention operating characteristic: examples from visual search,”
Science 202(4365):315–18 (October 1978). G. Sperling and B. A. Dosher, “Strategy and optimization in human
information processing,” in Handbook of perception and human performance, ed. K. Boff, L. Kaufman, and J.
Thomas (New York: Wiley, 1986).
27
Ibid.
20

36
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
attention.28 The spacing of items in the visual field also impacts visual sensitivity.29 When objects are
closely spaced, their discriminability is reduced. One explanation offered for this “crowding effect” is that
the spacing of visual items is smaller than the resolution of visual attention.30 The visual phenomenon of
crowding suggests that a crime committed in a visually complex scene, such as a sporting event, could
easily place limits on the ability of a witness to accurately perceive the facial features of a perpetrator.
A related consequence of attentional noise is that competing interests can readily hijack the
attentional focus. The technique of misdirection—one of the original mainstays of performance magic—
directs attention to uninformative image content and exploits the invisibility of unattended features.31 The
well-studied inattentional blindness effect32 is another example of this phenomenon, in which attention
that is pre-directed to one behaviorally significant property of a visual scene precludes awareness of other
features that also may be important. (For a dramatic demonstration of this effect, produced by Simons and
Chabris,33 see http://tinyurl.com/inattentional-blindness.)
Inattentional blindness effects translate well to real-world interactions between people. An
individual can be surprisingly unaware of surreptitious changes to the physical appearance of another
person while engaged in conversation.34 One demonstration of this phenomenon involved two strangers
(experimenter and pedestrian) in a brief face-to-face conversation on a sidewalk. At some point in the
conversation an opaque door was carried between the two individuals, and another person with different
appearance, clothing, and voice quickly replaced the experimenter. More than half of the participants
(pedestrians) failed to notice that their conversation partner had changed. This finding suggests that
naturally occurring events that briefly divert attention have the potential to markedly impair the accuracy
of eyewitness identifications.
Attentional hijacking is particularly characteristic of stimuli that elicit strong emotional responses,35
such as fear and arousal. Visual stimuli that trigger fear responses act as powerful external cues that

28

N. Lavie, “Perceptual load as a necessary condition for selective attention,” Journal of Experimental
Psychology: Human Perception and Performance, 21 (3): 451–468 (June 1995). J. W. Couperus, “Perceptual load
influences selective attention across development,” Developmental Psychology 47(5):1431–1439 (September 2011).
29
D. M. Levi, “Crowding—An essential bottleneck for object recognition: A mini-review,” Vision Research 48:
635–654 (2008).
30
J. Intriligator and P. Cavanagh, “The spatial resolution of visual attention,” Cognitive Psychology 43: 171–216
(2001).
31
G. Kuhn, B. W. Tatler, J. M. Findlay, and G. G. Cole, “Misdirection in magic: Implications for the
relationship between eye gaze and attention,” Visual Cognition, 16(2–3): 391–405 (2008). S. L. Macknik, S.
Martinez-Conde, and S. Blakeslee, Sleights of mind: what the neuroscience of magic reveals about our everyday
deceptions (New York: Henry Holt and Co., 2010).
32
A. Mack and I. Rock, Inattentional blindness (Cambridge, MA: MIT Press, 1998). U. Neisser and R.
Becklen, “Selective looking: Attending to visually specified events,” Cognitive Psychology, 7(4): 480–494 (October
1975). D. Simons, “Attentional capture and inattentional blindness,” Trends in Cognitive Sciences, 4(4): 147–155
(April 2000).
33
D. J. Simons and C. F. Chabris, “Gorillas in our midst: Sustained inattentional blindness for dynamic events,”
Perception 28: 1059–1074 (1999).
34
D. J. Simons and D. T. Levin, “Failure to detect changes to people during a real-world interaction,”
Psychonomic Bulletin and Review 5 (4): 644–649 (1998).
35
C. H. Hansen and R. D. Hansen, “Finding the Face in the Crowd: An Anger Superiority Effect,” Journal of
Personality and Social Psychology 54: 917–924 (1988). E. Fox, V. Lester, R. Russo, R. J. Bowles, A. Pichler, and
K. Dutton, “Facial expressions of emotion: Are angry faces detected more efficiently?” Cognition and Emotion 14
(1): 61–92 (2000). R. Compton, “The interface between emotion and attention: A review of evidence from
psychology and neuroscience,” Behavioral & Cognitive Neuroscience Reviews 2(2): 115–129 (2003). R. L.
Bannerman, E. V. Temminck, and A. Sahraie, “Emotional stimuli capture spatial attention but do not modulate
spatial memory,” Vision Research 65: 12–20 (15 July 2012).

37
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
command attention.36 While this potentiates sensitivity to those stimuli, at the considerable expense of
sensitivity to others,37 it is often the case that the attended emotional stimuli are not the ones with relevant
informational content. The so-called weapon focus is a real-world case in point for eyewitness
identification,38 in which attention is compellingly drawn to emotionally laden stimuli, such as a gun or a
knife, at the expense of acquiring greater visual information about the face of the perpetrator (see also
discussion of weapon focus in Chapter 5). (One might argue that this is an adaptation that benefits
immediate action or engagement with a threatening stimulus, but is surely detrimental to one’s efforts to
bear witness.)
Visual Perception
Visual perception is the conscious functional result of efforts to identify the environmental causes of
the pattern of light cast onto the back of the eye.39 Perception does not reflect the sensory world passively,
as camera film detects patterns of light. On the contrary, visual perception is constructive40 and entails (1)
integrating and segmenting attended attributes of the visual image into objects, (2) complementing and
interpreting the product with expectations derived from memory of prior experiences with the world, and
(3) assigning meaning and emotional valence by reference to prior knowledge of function and value. All
of the perceptual processes are affected by noise. Because the things perceived are the things we place
into memory, perceptual noise can dramatically limit the accuracy of eyewitness identification.
The process of feature integration and interpretation may be distorted by images of an object unique
to a specific angle of view.41 The retinal pattern generated by a face viewed directly from the front differs
considerably—with changes in aspect ratio and relative placement of facial features—from that generated
by a face viewed from an oblique side angle. Viewing a face from an angle above or below center (as
might be the case if the criminal were standing over you, or below you on the stairs) also yields retinal
distortions of facial features. In this case, the distortions prominently mimick facial gestures of smiling
versus frowning, and perhaps cause incorrect inferences about the emotional state of the person observed
and his or her intentions and motivations. (This distortion is the basis for the Japanese Noh Theatre mask
effect,42 in which a rigid mask tilted forward leads to the appearance of a smile and backward leads to the
appearance of a frown—an effect you can simulate by simply looking into the mirror and tilting your face
up or down.)
36

J. A. Easterbrook, “The effects of emotion on cue utilization and the organization of behavior,” Psychological
Review 66 (3):183–201 (1959).
37
E. Ferneyhough, M. K. Kim, E. A. Phelps, M. Carrasco, “Anxiety modulates the effects of emotion and
attention on early vision,” Cognition & Emotion 27(1):166–176 (2013). G. Pourtois and P. Vuilleumier, “Dynamics
of emotional effects on spatial attention in the human visual cortex,” Progress in Brain Reserach 156: 67–91 (2006).
38
T. Kramer, R. Buckhout, and P. Eugenio, “Weapon focus, arousal, and eyewitness memory: Attention must be
paid,” Law and Human Behavior 14 (2): 167–184 (1990). R. S. Truelove, “Do weapons automatically capture
attention,” Applied Cognitive Psychology 20 (7): 871–893 (2006). E. F. Loftus, G. R. Loftus, and J. Messo, “Some
facts about ‘weapon focus’,” Law and Human Behavior 11 (1): 55–62 (1987).
39
W. James, Principles of Psychology (New York: Henry Holt, 1890). S. Harnad, ed., Categorical Perception:
The Groundwork of Cognition (New York: Cambridge University Press, 1987). T. D. Albright, “Perceiving,”
Daedalus (in press).
40
C. D. Gilbert, “The Constructive Nature of Visual Processing,” in Principles of Neuroscience, 5th Edition, ed.
E. Kandel, J. H. Schwartz, T. M. Jessell, S. A. Siegelbaum, and A. J. Hudspeth (New York: McGraw-Hill
Professional, 2012). T. D. Albright, “On the Perception of Probable Things: Neural Substrates of Associative
Memory, Imagery, and Perception,” Neuron 74 (2): 227–245 (2012).
41
W. G. Hayward and P. Williams, “Viewpoint dependence and object discriminability,” Psychological Science
11(1): 7–12 (2000).
42
M. J. Lyons et al., “The Noh mask effect: vertical viewpoint dependence on facial expression perception,”
Proceedings of the Royal Society B: Biological Sciences 267 (1459): 2239–2245 (2000).

38
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Viewing conditions can also affect the perception of face, gender, and age.43 Investigators found that
faces that were physically identical—and particularly those bordering on androgyny—were perceived as
unambiguously male or female depending on where they appeared in the observer’s visual field. The
spatial patterning of these effects was distinctive and stable for each observer. Perceptual distortions of
this sort are a source of noise that may have important implications for the accuracy of eyewitness
identification.
Perceptual distortions also may be introduced through memory recall. The way an observer
experiences a visual scene—the setting, the people, and the actions associated with a crime —is
commonly influenced as much by expectations from prior experience with the world as it is by the precise
patterns of light cast upon the retina. There are good reasons why this is true. As noted above, the sensory
input (the pattern of light received) is often noisy, incomplete, and ambiguous, and our memories of what
is likely to be out there, given the context, are called on to fill in the blanks, reconcile ambiguities, and
leave us with clear and coherent percepts.44 This perceptual completion is probabilistic.45 It is an
hypothesis, and the accuracy naturally depends on the degree to which the observer’s expectations match
the noisy sensory data.
What is implied is that the same mechanism that grants the certainty of perceptual experience in the
face of noise and ambiguity is also capable of implicitly fabricating content that does not correspond to
external reality and yet is experienced with no less certainty. Performance magic relies on this
constructive nature of perceptual experience, and that nature is also the foundation for many visual
illusions and forms of visual art.46 In a classic experiment that drives home the point, Bruner and Postman
looked at the ability of observers to recognize ‘‘trick’’ playing cards.47 The trick cards were created by
altering the color of a given suit (e.g., a red seven of spades). Observers were shown a series of cards with
brief presentations. Some cards were trick, and the remainder normal. With astonishing frequency,
observers reported that the trick cards were normal. When questioned, observers defended their reports,
even after being allowed to scrutinize the trick cards, thus demonstrating that learned properties of the
world are capable of sharply altering our experience and, moreover, reinforcing our convictions about
what we have seen, even in the face of countermanding sensory evidence. In view of this inherent
dependence of perception on prior experiences and context—and, importantly, the fact that the viewer is
commonly none the wiser when perception differs from the “ground truth” of the external world—it
appears that accurate eyewitness identification may be difficult to achieve.
Additional noise (in this case defined as uncertainty resulting from loss of perceptual resolution)
may result from the fact that visual perception is categorical.48 Although the objects of our experience
43

A. Afraz, M. Vaziri-Pashkam, and P. Cavanagh, “Spatial heterogeneity in the perception of face and form
attributes,” Current Biology 20(23): 2112–2116 (2010).
44
T. D. Albright, “On the Perception of Probable Things: Neural Substrates of Associative Memory, Imagery,
and Perception,” Neuron, 74 (2): 227–245 (2012). U. Neisser, Cognitive Psychology (New York: AppletonCentury-Crofts, 1967).
45
D. C. Knill and W. Richards, Perception as Bayesian Inference, ed. D. C. Knill and W. Richards (Cambridge:
Cambridge University Press, 1996). D. Kersten “High-level vision as statistical inference,” in The New Cognitive
Neurosciences 2nd Edition, ed. M. S. Gazzaniga (Cambridge: MIT Press,1999), 353–363. D. Kersten, P.
Mamassian, and A. Yuille, “Object perception as Bayesian inference,” Annual Review of Psychology 55: 271–304
(February 2004).
46
E.H. Gombrich, Art and Illusion. A Study in the Psychology of Pictorial Representation (London: Phaidon
1960). T. D. Albright, “The Veiled Christ of Cappella Sansevero: On art, vision and reality,” Leonardo 46(1): 19–23
(2013). S. L. Macknik, S. Martinez-Conde and S. Blakeslee, Sleights of mind: what the neuroscience of magic
reveals about our everyday deceptions (New York: Henry Holt and Co., 2010).
47
J. S. Bruner and L. Postman, “On the perception of incongruity; a paradigm,” Journal of Personality 18 (2):
206–223 (1949).
48
W. James, Principles of Psychology (New York: Henry Holt, 1980). S. Harnad, ed., Categorical Perception:
The Groundwork of Cognition (New York: Cambridge University Press, 1987).

39
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
vary broadly along multiple sensory dimensions, we lump them into categories using prior associations,
many of which stem from common functions, physical properties, meanings, or emotional valence. The
apples in a basket, or the many typographic fonts for the letter “A,” are visually distinct, yet we readily
perceive them as categorically identical. For most behavioral and cognitive goals, perceptual processing is
greatly simplified by treating all members of a category as the same, despite their differences. It rarely
matters, for example, whether the apple we choose is dappled on one side or irregular in shape, nor does
the font used bear greatly on our ability to read. One of the functional corollaries of categorical perception
is that observers are far better at discriminating between objects from different categories than objects
from the same category.49 Evidence indicates that the structure of object memory is also categorical,50
suggesting that perceived objects are encoded in memory as a category type, often without specific detail.
Perceptual categorization naturally applies to faces.51 We readily categorize faces by distinctions
along the obvious dimensions of gender, age, and race, but we also draw distinctions along dimensions
such as skin tone, hair color and style, presence and type of facial hair, such subtler factors as shape of
cheeks and jaw, and subjective qualities such as attractiveness. The practical consequence of this for
eyewitness identification is that the precision of a perceptual experience may be reduced within any of
these categories, particularly because we typically witness criminal events for such a brief period of time.
The ensuing memory of the experience will likely reflect that reduced precision, and the memory
retrieved may regress to a category prototype, or to other exemplars of the perceived category.52 The
witness may categorically perceive a square jawed man with a moustache, but the fine details needed for
individuation of a suspect are neither perceived nor encoded in memory. For example, although you may
have seen the iconic Marlboro Man countless times on billboards and in magazines, it is unlikely that you
could distinguish him in a lineup from other square jawed mustachioed men.
MEMORY
Functional Processes of Memory
Conscious visual perceptual experiences, rendered by the processes described in the previous section
on vision, are commonly stored as declarative memories, meaning that they can be consciously accessed
49

R. Goldstone, “Influences of categorization on perceptual discrimination,” Journal of Experimental
Psychology General 123 (2):178–200 (1994). R. Goldstone, Y. Lippa, and R. M. Shiffrin, “Altering object
representations through category learning,” Cognition 78 (1): 27–43 (2001).
50
E. Tulving, “Episodic and semantic memory,” in Organization of Memory, ed. E. Tulving and W. Donaldson
(New York: Academic Press, 1972), 381–403. L. K. Tyler et al., “Processing objects at different levels of
specificity,” Journal of Cognitive Neuroscience, 16 (3): 351–362 (2004). M. J. Farah and J. L. McClelland, “A
computational model of semantic memory impairment: modality specificity and emergent category specificity,”
Journal of Experimental Psychology: General, 120 (4): 339–357 (1991). C. Gerlach et al., “Categorization and
category effects in normal object recognition: A PET study,” Neuropsychologia, 38 (13): 1693–1703 (2000). G. W.
Humphreys and E. M. Forde, “Hierarchies, similarity, and interactivity in object recognition: ‘Category-specific’
neuropsychological deficits,” Behavioral and Brain Sciences, 24 (3): 453–476 (2001).
51
J. M. Beale and F. C. Keil, “Categorical effects in the perception of faces,” Cognition 57(3): 217–239 (1995).
D. T. Levin, “Classifying Faces by Race: The Structure of Face Categories,” Journal of Experimental Psychology:
Learning, Memory, and Cognition 22(6):1364–1382 (1996). D. T. Levin and J. Beale, “Categorical perception
occurs in newly learned faces, cross-race faces, and inverted faces,”Perception and Psychophysics, 62: 386–401
(2000). M. A. Webster, D. Kaping, Y. Mizokami, and P. Duhamel, “Adaptation to natural facial categories,” Nature
428(6982): 5 57–61 (2004). Y. Lee, C. R. Smith, C. L. Grady, N. Hoang, M. Moscovitch, “Broadly tuned face
representation in older adults assessed by categorical perception,”Journal of Experimental Psychology: Human
Perception and Performance 40(3): 1060–1071 (2014).
52
J. Huttenlocher, L. V. Hedges, and J. L. Vevea, “Why do categories affect stimulus judgement?” Journal of
Experimental Psychology: General, 129 (2): 220–241 (2000). R. Goldstone, Y. Lippa, and R. M. Shiffrin, “Altering
object representations through category learning,” Cognition 78 (1): 27–43 (2001).

40
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
and expressed as knowledge about the world (as distinct from procedural memories, such as motor
skills).53 Declarative memories are of two types, semantic and episodic,54 reflecting a distinction between
memories of meanings, facts, and concepts versus memories of events (such as those witnessed during a
crime). Declarative memories are conceptualized as involving three core processes55—encoding, storage
and retrieval—which refer to the placement of items in memory, their maintenance therein, and
subsequent access to the stored information.
Like vision, memory is also beset by noise. Encoding, storage, and remembering are not passive,
static processes that record, retain, and divulge their contents in an informational vacuum, unaffected by
outside influences. The contents cannot be treated as a veridical permanent record, like photographs
stored in a safe. On the contrary, the fidelity of our memories for real events may be compromised by
many factors at all stages of processing, from encoding through storage, to the final stages of retrieval.
Without awareness, we regularly encode events in a biased manner and subsequently forget, reconstruct,
update and distort the things we believe to be true.56
In the following sections, we discuss memory encoding, storage and retrieval, with emphasis on the
limits of these processes as they pertain to eyewitness identification. Emotions can strongly influence
these processes of memory; some specific actions are highlighted. We also discuss the phenomenon of
“recognition memory,” which refers to the specific type of memory retrieval in which a stimulus (e.g., a
face) is used to probe memory and the rememberer (e.g., an eyewitness) must decide whether the strength
of the elicited memory evidence is sufficient to declare that the stimulus was previously encountered or is
novel. Recognition memory underlies eyewitness identification, as the witness must make a recognition
decision.
Memory Encoding
Memory encoding refers to the process whereby perceived objects and events are initially placed
into storage. The encoding process involves two stages, which are commonly distinguished by the
quantity of information stored, the duration of storage, and the susceptibility to interference.57 Short-term
53

W. James, Principles of Psychology (New York: Henry Holt, 1890). S. Harnad, ed., Categorical Perception:
The Groundwork of Cognition (New York: Cambridge University Press,1987). B. Milner, Physiologie de
l’hippocampe, ed. P. Passouant (Paris: Centre National de la Recherche Scientifique, 1962), 257–272. L. R. Squire
and J. Wixted, “The Cognitive Neuroscience of Human Memory since H.M.,” Annual Review of Neuroscience 34:
259–288 (2011).
54
E. Tulving, “Episodic and semantic memory,” in Organization of Memory, ed. E. Tulving & W. Donaldson
(New York: Academic Press, 1972), 381–403.
55
E. Tulving, “Organization of memory: Quo vadis?” in The Cognitive Neurosciences, ed. M. S. Gazzaniga
(Cambridge, MA: MIT Press, 1995), 839–847.
56
J. T. Wixted, “The psychology and neuroscience of forgetting,” Annual Review of Psychology 55: 235–269
(2004). E. Tulving and D. M. Thomson, “Encoding specificity and retrieval processes in episodic memory,”
Psychological Review 80 (5): 352–373 (1973). Y. Dudai, “Reconsolidation: The advantage of being refocused,”
Current Opinion in Neurobiology 16 (2): 174–178 (2006). E. F. Loftus, “Planting misinformation in the human
mind: A 30-year investigation of the malleability of memory,” Learning & Memory 12(4): 361–366 (2005). R. A.
Bjork, “Interference and memory,” in Encyclopedia of learning and memory, ed. L. R. Squire (New York:
Macmillan, 1992), 283–288. J. A. McGeoch, “Forgetting and the law of disuse,” Psychological Review 39(4): 352–
370 (1932). J. G. Jenkins and K. M. Dallenbach, “Obliviscence during Sleep and Waking,” The American Journal
of Psychology 35 (4): 605–612 (1924). B. J. Underwood and L. Postman, “Extra-experimental sources of
interference in forgetting,” Psychological Review 67 (2): 73–95 (1960).
57
R. C. Atkinson and R. M. Shiffrin, “Human memory: A proposed system and its control processes,” in The
psychology of learning and motivation (Volume 2), ed. K. W. Spence and J. T. Spence (New York: Academic
Press,1968), 89–195. W. James, Principles of Psychology (New York: Henry Holt, 1890). A. Baddeley, “Working
memory: looking back and looking forward,” Nature Reviews Neuroscience 4 (10): 829–839 (2003). A. Baddley,
Working Memory (New York: Oxford University Press,1986).

41
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
or working memory is the conscious content of recent perceptual experiences, or information recently
recalled from long-term storage. Information that remains at the focus of attention persists in and forms
the contents of short-term memory. This form of memory is of limited duration and capacity,58 and is
labile, decaying quickly with time and easily disrupted by other perceptual or cognitive processes.59
Through cellular and molecular events that play out over time, the contents of short-term memories may
be encoded and consolidated into long-term memory,60 which is more enduring (albeit evolving with
ongoing experience), and of greater capacity.
The structure of an individual’s full library of long-term declarative memories can be thought of as a
collection of associations between items of specific semantic (e.g., the fact that that person X is a 34-yearold female) or episodic content61 (e.g., the fact that person X was at location Y on the night of the
witnessed crime). As the individual gains new experiences, long-term declarative memories may be
updated by adding new content to the existing library, or by forming new associations between existing
content.62
Memories are particularly labile during the encoding process. The contents of short-term memory
are limited and highly subject to interference by subsequent sensory, cognitive, emotional, or behavioral
events; the contents can also be biased by prior knowledge, expectations, or beliefs, resulting in a
distorted representation of experience. Short-term memories of events that happened early in a witnessed
proceeding may simply be forgotten with the passage of time, or badly compromised by attention directed
to subsequent emotional events, or by cognitive and behavioral demands (e.g., anxiety, fear, or the need to
escape). In such cases, the compromised information may never be consolidated fully into long-term
storage or that storage may contain distorted content.63 At the same time, the quality of encoding of
stimuli that are attended is commonly enhanced by highly emotional content.64
Memory Storage
Memory storage refers to the long-term retention of information after encoding. The stability of
stored information is continuously challenged and subject to modification. We forget, qualify, or distort
existing memories as we acquire new perceptual experiences and encode new content and associations
into memory.65
58

G. A. Miller, “The magical number seven,” The Psychological Review 63 (2): 81–97 (1956).
J. Jonides et al., “The mind and brain of short-term memory,” Annual Review of Psychology 59: 193–224
(2008).
60
E. Kandel and L. Squire, Memory: From Mind to Molecules (New York: Scientific American Library, 2008).
61
J. R. Anderson, The Architecture of Cognition (Cambridge: Harvard University Press, 1983). J. R. Anderson
and C. Lebiere, The Atomic Components of Thought (Mahwah: Lawrence Erlbaum Associates, 1998).
62
M. P. Walker et al., “Dissociable stages of human memory consolidation and reconsolidation,” Nature 425:
616 (2003).
63
J. L. McGaugh, “Memory—a Century of Consolidation,” Science 287 (5451): 248–251 (2000). J. L.
McGaugh and B. Roozendaal, “Role of adrenal stress hormones in forming lasting memories in the brain,” Current
Opinion in Neurobiology 12 (2): 205–210 (2002).
64
K. N. Ochsner, “Are affective events richly recollected or simply familiar? The experience and process of
recognizing feelings past,” Journal of Experimental Psychology: General 129 (2): 242–261 (2000). D. Talmi, A. K.
Anderson, L. Riggs, J. B. Caplan, and M. Moscovitch, “Immediate memory consequences of the effect of emotion
on attention to pictures,” Learning and Memory 15 (2008):172–182. E. A. Kensinger and D. L. Schacter, “Neural
processes supporting young and older adults’ emotional memories,” Journal of Cognitive Neuroscience 7 (2008):1–
13. E. A. Phelps. “Emotion and cognition: insights from studies of the human amygdala,” Annual Review of
Psychology 57: 27–53 (2006).
65
J. T. Wixted, “The psychology and neuroscience of forgetting,” Annual Review of Psychology 55: 235–269
(2004). E. Tulving and D. M. Thomson, “Encoding specificity and retrieval processes in episodic memory,”
Psychological Review 80 (5): 352–373 (1973). Y. Dudai, “Reconsolidation: The advantage of being refocused,”
Current Opinion in Neurobiology 16 (2):174–178 (2006). E. F. Loftus, “Planting misinformation in the human
59

42
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Forgetting can be partially mitigated, and memories stabilized, by habits of retrieval (or reactivation)
and reconsolidation,66 which happen whenever we tell the story of our experiences. Reactivation is not
perfect. With each implicit retrieval or explicit telling of the story, we may smooth over inconsistencies
unconsciously or modify content based on our prior beliefs, the accounts of others, or through the lens of
new information. We may add embellishments that reflect opinions, emotions, or prejudices67 rather than
observed facts; or we may simply omit disturbing content and pass over fine details.68
A second threat to the stability of long-term memories is, ironically, our life-long ability to learn
new things. Because memory mechanisms are inherently plastic throughout life, content stored for the
long term is surprisingly labile in the face of new information. Our memories are thus an ever-evolving
account of our experiences. A memory that reflects witnessing person X at location Y on a particular
evening might be readily and notably updated by subsequent learning that location Y is the home of a
business associate of person X. Our memories of the witnessed actions of person X may be qualified by
new knowledge of her life history. Moreover, because new content can be added and the source of that
content forgotten, we may attribute our updated memories to the originally witnessed events—in some
cases substantially changing what we believe we have seen.69 It is thus not surprising that newly
incorporated information need not be true to fact. Research on false memories shows that it is possible to
plant fabricated content in memory, which leads us to recall things we never experienced.70
The emotional content of stored memories is a factor that appears to promote long-term retention;
memories of highly arousing emotional stimuli, such as those associated with a witnessed crime, tend to
be more enduring than memories of non-arousing stimuli.71 Highly salient, unexpected, or arousing
mind: A 30-year investigation of the malleability of memory,” Learning and Memory, 12(4): 361–366 (2005). R. A.
Bjork, “Interference and memory,” in Encyclopedia of learning and memory, ed. L. R. Squire (New York:
Macmillan, 1992), 283–288. J. A. McGeoch, “Forgetting and the law of disuse,” Psychological Review 39(4): 352–
370 (1932). J. G. Jenkins and K. M. Dallenbach, “Obliviscence during Sleep and Waking,” The American Journal
of Psychology 35 (1924): 605–612. B. J. Underwood and Postman, “Extra-experimental sources of interference in
forgetting,” Psychological Review 67 (2): 73–95 (1960). E. F. Loftus, “The malleability of human memory,”
American Scientist 67 (3): 312–320 (1979). D. J. Yi et al., “When a thought equals a look: Refreshing enhances
perceptual memory,” Journal of Cognitive Neuroscience 20 (8): 1371–1380 (2008).
66
C. M. Alberini, Memory Reconsolidation (Waltham: Academic Press, 2013).
67
D. L. Schacter, Psychology, Second Edition (New York: Worth Publishers, 2011), 253–254. E. F. Loftus and
H. G. Hoffman, “Misinformation and memory, the creation of new memories,” Journal of Experimental Psychology
118(1): 100–104 (1989). G. Mazzoni and A. Memon, “Imagination can create false autobiographical memories,”
Psychological Science 14(2): 186–188 (2003).
68
F. C. Bartlett, Remembering: A Study in Experimental and Social Psychology (London: Cambridge University
Press, 1932).
69
D. S. Lindsay and M. K. Johnson, “Recognition memory and source monitoring,” Bulletin of the Psychonomic
Society 29(3): 203–205 (1991). D. L. Schacter and C. S. Dodson, “Misattribution, false recognition and the sins of
memory,” Philosophical Transactions of the Royal Society: Biological Sciences 356 (1413): 1385–1393 (2001). L.
A. Henkel, N. Franklin, and M. K. Johnson, “Cross-Modal Source Monitoring Confusions Between Perceived and
Imagined Events,” Journal of Experimental Psychology: Learning, Memory, and Cognition 26 (2): 321–335 (2000).
D. L. Schacter, ed., Memory distortion: How minds, brains, and societies reconstruct the past (Cambridge, MA:
Harvard University Press, 1995). K. J. Mitchell and M. K. Johnson, “Source monitoring: Attributing mental
experiences,” in The Oxford handbook of memory, ed. E. Tulving and F. I. M. Craik (New York: Oxford University
Press, 2000), 179–195. H. L. Roediger III, and K. B. McDermott, “Creating false memories: Remembering words
not presented in lists,” Journal of Experimental Psychology: Learning, Memory, & Cognition 21 (4): 803–814
(1985).
70
E. F. Loftus, “Planting misinformation in the human mind: A 30-year investigation of the malleability of
memory,” Learning & Memory 12(4): 361–366 (2005). E. F. Loftus and J. E. Pickrell, “The formation of false
memories,” Psychiatric Annals, 25 (12): 720–725 (1995). M. K. Johnson and C. L. Raye, “False memories and
confabulation,” Trends in Cognitive Sciences 2 (4): 137–145 (1998).
71
L. J. Kleinsmith and S. Kaplan, “Paired-associate learning as a function of arousal and interpolated interval”
Journal of Experimental Psychology 65 (2): 190–193 (1963). M. W. Eysenck, “Arousal, learning, and memory,”

43
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
events—such as the Kennedy assassination or the Space Shuttle disaster—are commonly more strongly
stored in memory and their later retrieval is often associated with the subjective experience of high
vividness and a sense of reliving72 (although not necessarily with greater accuracy, as detailed below).
The stronger encoding and storage of emotional memories results from the engagement of a specialized
system of stress hormones (glucocorticoids) which is triggered by arousing content and has potentiating
effects on the neuronal processes underlying memory consolidation and storage.73 Despite the vividness
and the sense of reliving that characterizes retrieval of emotional memories, there are many indications
that such memories are just as prone to errors.74 This may reflect, in part, memory enhancements, of the
sort described above, which accompany frequent re-consolidation or re-telling of the story of the
emotional experience,75 and often include details (some true to fact, some not) learned after the
experience. Although emotional memories are often inaccurate in detail, one important corollary of their
vividness is that they are frequently held with high confidence.76 This breakdown of the relationship
between accuracy and confidence can obviously undermine eyewitness accounts.77
Psychological Bulletin 83 (3): 389–404 (1976). F. Heuer and D. Reisberg, “Vivid memories of emotional events:
The accuracy of remembered minutiae,” Memory & Cognition 18 (5): 496–450 (1990). T. Sharot and E. A. Phelps,
“How arousal modulates memory: Disentangling the effects of attention and retention,” Cognitive, Affective, &
Behavioral Neuroscience 4 (3): 294–306 (2004). E. A. Kensinger, R. J. Garoff-Eaton, and D. L. Schacter, “Memory
for specific visual details can be enhanced by negative arousing content,” Journal of Memory and Language 54 (1):
99–112 (2006). E. Kensinger, “Remembering emotional experiences: The contribution of valence and arousal,”
Reviews in the Neurosciences 15 (4): 241–251 (2004).
72
G. Wolters and J. J. Goudsmit, “Flashbulb and event memory of September 11, 2001: Consistency, confidence
and age effect,” Psychological Report 96: 605–619 (2005). E. A. Kensinger, A. C. Krendl, and S. Corkin S.
“Memories of an emotional and a nonemotional event: Effects of aging and delay interval,” Experimental Aging
Research 32: 23–45 (2006). U. Neisser and N. Harsch, “Phantom flashbulbs: false recollections of hearing the news
about Challenger,” in Affect and accuracy in recall: Studies of “flashbulb” memories, ed. E. Winograd and U.
Neisser (New York: Cambridge University Press, 1992): 9–31. K. S. LaBar and E. A. Phelps, “Arousal-mediated
memory consolidation: Role of the medial temporal lobe in humans,” Psychological Science 9 (6): 490–493 (1998).
73
J. L. McGaugh, “Memory: A Century of Consolidation,” Science 287 (5451): 248–251 (2000). J. L.
McGaugh and B. Roozendaal, “Role of adrenal stress hormones in forming lasting memories in the brain,” Current
Opinion in Neurobiology 12 (2): 205–210 (2002).
74
E. A. Kensinger, “Remembering the details: Effects of emotion,” Emotion Review 1(2): 99–113 (2009). T.
Sharot, M. R. Delgado, and E. A. Phelps, “How emotion enhances the feeling of remembering,” Nature
Neuroscience 7 (12):1376–1380 (2004). H. Schmolck, E. A. Buffalo, L. R. Squire, “Memory distortions develop
over time: Recollections of the O.J. Simpson trial verdict after 15 and 32 months,” Psychological Science 11 (1):
39–45 (2000). S. R. Schmidt, “Autobiographical memories for the September 11th attacks: Reconstructive errors
and emotional impairment of memory,” Memory and Cognition 32 (3): 443–454 (2004). T. W. Buchanan and R.
Adolphs, “The role of the human amygdala in emotional modulation of long-term declarative memory,” in
Emotional Cognition: From Brain to Behavior, ed. S. Moore and M. Oaksford. (Amsterdam: John Benjamins
Publishing, 2002), 9–34.
75
E. Soleti, A. Curci, A. Bianco, and T. Lanciano, “Does Talking About Emotions Influence Eyewitness
Memory? The Role of Emotional vs. Factual Retelling on Memory Accuracy,” Europe’s Journal of Psychology
8(4), 632–640 (2012).
76
U. Rimmele, L. Davachi, R. Petrov, S. Dougal, and E. A. Phelps, “Emotion enhances the subjective feeling of
remembering, despite lower accuracy for contextual details,” Emotion 11(3): 553–562 (2011). E. A. Kensinger,
“Remembering the detail: Effects of emotion,” Emotion Review 1(2): 99–113 (2009). U. Neisser and N. Harsch,
Affect and Accuracy in Recall: Studies of Flashbulb Memories (New York: Cambridge University Press, 1992). R.
A. Poldrack, A. D. Wagner, E. A. Phelps, and T. Sharot, “How (and Why) Emotion Enhances the Subjective Sense
of Recollection,” Current Directions in Psychological Science 17 (2): 147–152 (2008).
77
K. A. Houston, B. R. Clifford, L. H. Phillips, and A. Memon, “The emotional eyewitness: The effects of
emotion on specific aspects of eyewitness recall and recognition performance,” Emotion 13(1): 118–128 (2013). R.
B. Edelstein, K. W. Alexander, G. S. Goodman and J. W. Newton, “Emotion and eyewitness memory,” in Memory
and emotion, ed. D. Reisberg and P. Hertel (New York: Oxford University Press, 2004): 308–346. S-A.

44
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
The enduring plasticity of stored memories is a serious concern for the validity of eyewitness
identification. A witness’ inevitable interactions with law enforcement and legal counsel, not to mention
communications from journalists, family, and friends, have the potential to significantly modify the
witness’ memory of faces encountered and of other event details at the scene of the crime.78 Thus, the
fidelity of retrieved events—and the accuracy of identification—is likely to be greater when retrieval
occurs closer to the time of the witnessed events. The conclusion above has important implications for
law enforcement and the legal process and calls into question the validity of in-court identifications and
their appropriateness as statements of fact.
Memory Retrieval
Memory retrieval refers to the process by which stored information is accessed and brought into
consciousness, where it can be used to make decisions and guide actions. Retrieval of long-term
declarative memories is often triggered through association with an external stimulus (i.e., a retrieval
cue).79 For example, the slight stubble on a lineup participant’s face may be enough to elicit retrieval of a
suspect’s entire face. These same retrieval processes can also be engaged internally – a verbally triggered
stream of thought related to a witnessed crime may readily bring to mind visual features of the
perpetrator. A corollary of this association-based phenomenon is that memory retrieval is often context
dependent; a memory may be more readily retrieved if the observer is in physical surroundings that are
the same as or similar to those in which the original experiences took place (because the surroundings
provide additional cues to trigger memory retrieval).80
Memory retrieval is heavily affected by various sources of noise. Similarities of meaning or
appearance between retrieval cues and items in memory can easily lead to retrieval of the wrong item,
producing a false memory.81 This is particularly a problem given the categorical nature of memory. 82 The
Christianson, “Emotional stress and eyewitness memory: A critical review,” Psychological Bulletin 112(2): 284–
309 (1992).
78
M. S. Zaragoza and S. M. Lane, “Sources of misattribution and suggestibility of eye witness testimony,”
Journal of Experimental Psychology: Learning, Memory, and Cognition 20 (4): 934–945 (1994). W. C. Thompson,
K. A. Clarke-Stewart, and S. J. Lepore, “What Did the Janitor Do? Suggestive Interviewing and the Accuracy of
Children’s Accounts,” Law and Human Behaviour 21(4): 405–426 (1997). D. S. Lindsay and M. K. Johnson, “The
eyewitness suggestibility effect and memory for source,” Memory & Cognition 17(3): 349–358 (1989).
79
E. Tulving and Z. Pearlstone, “Availability versus accessibility of information in memory for words,” Journal
of Verbal Learning and Verbal Behaviour 5: 381–391 (1966).
80
D. Godden and A. Baddeley, “Context dependent memory in two natural environments,” British Journal of
Psychology 66 (3): 325–331 (1975). S. M. Smith and E. Vela, “Environmental context-dependent eyewitness
recognition,” Applied Cognitive Psychology 6: 125–139 (1992). S. M. Smith and E. Vela, “Environmental contextdependent memory: A review and meta-analysis,” Psychonomic Bulletin Review 8 (2): 203–220 (2001). E. Tulving
and M. Thomson, “Encoding specificity and retrieval processes in episodic memory,” Psychological Review 80(5):
352–373 (1973).
81
J. R. Anderson, “A spreading activation theory of memory,” Journal of Verbal Learning and Verbal Behavior
22 (3): 261–295 (1983). A. M. Collins and E. F. Loftus, “A spreading-activation theory of semantic processing,”
Psychological Review 82 (6):407–428 (1975). H. L. Roediger III, D. A. Balota, J. M. Watson, “Spreading activation
and arousal of false memories,” in The nature of remembering: Essays in honor of Robert G. Crowder, ed. H. L.
Roediger III, J. Nairne, I. Neath, and A. Surprenant (Washington, DC: American Psychological Association, 2001):
95–115. C. J. Brainerd and V. F. Reyna, The science of false memory (New York: Oxford University Press, 2005).
82
E. Tulving, “Episodic and semantic memory” in Organization of Memory, ed. E. Tulving and W. Donaldson
(New York: Academic Press, 1972): 381–403. M. J. Farah and J. L. McClelland, “A computational model of
semantic memory impairment: modality specificity and emergent category specificity,” Journal of Experimental
Psychology: General, 120 (4): 339–357 (1991). G. W. Humphreys and E. M. Forde, “Hierarchies, similarity, and
interactivity in object recognition: ‘Category-specific’ neuropsychological deficits,” Behavioral and Brain Sciences,
24 (3): 453–476 (2001).

45
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
rugged mustachioed man in the lineup may lead to retrieval of the familiar categorical prototype—the
Marlboro Man—rather than the specific person perceived at the scene of the crime, which in turn could
interfere with or lead to errors in recognition (i.e., identification). Another type of memory retrieval
failure is caused by “intrusion errors,” in which information known to be commonly associated with
events of a general type becomes incorporated into the retrieved content of a specific memory (and
subsequently incorporated into the reconsolidated memory). For example, because guns are often
associated with robbery, an observer may readily and unwittingly incorporate a gun into the retrieved
version of his or her memory of a witnessed robbery.
Intrusion errors are one manifestation of a larger retrieval problem in which there is loss of
information about the source of a memory. In cases of “source memory failure,” we effectively forget
how we know things, forget when and where we learned the content of our memories. What this means
practically is that we may attribute later acquisition of information to earlier experiences. An eyewitness
might learn from the police or some other source that a potential suspect has a moustache and then
attribute that knowledge to the witnessed events, which will, in turn, have disastrous consequences for the
ability of the eyewitness to accurately report what she or he has seen.
As for the processes of memory encoding and storage, the emotional content of memory also affects
memory retrieval. As noted above, memory retrieval is commonly context dependent. A related and welldocumented phenomenon that bears on emotional memories is state dependent memory, in which retrieval
accuracy is best if the individual’s cognitive state at the time of retrieval matches cognitive state at the
time of encoding.83 When memories have an emotional component, retrieval may be best when the
individual is induced to a corresponding emotional state (mood dependent memory),84 which is
accomplished by verbally or physically placing him or her in the same context, and may offer a valuable
investigative tool for probing eyewitness accounts.85
Recognition Memory
Recognition memory is the specific type of declarative memory retrieval in which a sensory
stimulus (a “cue” stimulus) elicits a memory of the stimulus stored following a prior encounter, and often
the sequence of events involving the stimulus, the spatial context in which the stimulus was experienced,
and the presence of other objects, people, or thoughts that had appeared with the stimulus during the
event.86 Recognition memory decisions are based on the retrieved memory evidence, which can be
triggered by the stimulus and can also emerge from an active search of items in memory. One factor
83

D. W. Goodwin, B. Powell, D. Bremer, H. Hoine, and J. Stern, “Alcohol and recall: State-dependent effects in
man,” Science 163(3873): 1358–1360 (1969). E. Tulving and D. Thomson, “Encoding specificity and retrieval
processes in episodic memory,” Psychological Review 80(5): 352–373 (1973). E. Girden and E. Culler,
“Conditioned responses in curarized striate muscle in dogs,” Journal of Comparative Psychology 23(2): 261–74
(1937). D. A. Overton, “State-dependent or ‘dissociated’ learning produced with pentobarbital,” Journal of
Comparative and Physiological Psychology 57(1): 3–12 (1964).
84
P. M. Kenealy, “Mood state-Dependent Retrieval: The Effects of Induced Mood on Memory Reconsidered,”
The Quarterly Journal of Experimental Psychology Section A: Human Experimental Psychology 50 (2): 290–317
(1997). P. A. Lewis and H. D. Critchley, “Mood-dependent memory,” Trends in Cognitive Sciences 7(10): 431–433
(2003). G. H. Bower, “Mood and memory,” American Psychologist 36 (2): 129–48 (1981). F. I. M. Craik and R. S.
Lockhart, “Levels of processing: A framework for memory research,” Journal of Verbal Learning and Verbal
Behavior 11 (6):671–684 (1972). E. A. Kensinger, “Remembering the detail: Effects of emotion,” Emotion Review
1(2): 99–113 (2009). K.A. Leight and H. C. Ellis “Emotional mood states, strategies, and state-dependency in
memory,” Journal of Verbal Learning and Verbal Behavior 20 (3): 251–266 (1981).
85
S. M. Smith and E. Vela, “Environmental context-dependent eyewitness recognition,” Applied Cognitive
Psychology 6: 125–139 (1992).
86
G. Mandler, “Recognizing: The judgment of previous occurrence,” Psychological Review 87 (3): 252–271
(1980).

46
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
affecting the strength of the evidence retrieved is the similarity between the cue stimulus and the stimulus
or stimuli that was/were previously encountered during the event. An observer engaged in this process
holds an implicit criterion for the strength of evidence required to reach a positive decision. In the case of
eyewitness identification, this process is routinely elicited by viewing faces in a lineup. When the
evidence retrieved is insufficient to reach a decision, this can lead to a cycle of ever-greater scrutiny of
the cue stimulus and efforts to recollect additional details of the original event. Ultimately a decision must
be made about whether the retrieved evidence is sufficient to declare that the stimulus was previously
experienced (or previously experienced in the particular event of interest) or whether the stimulus is novel
(or not from the event of interest). If a recognition event occurs – that is, if the memory search triggered
by one of the faces in the lineup leads to a strong enough subjective experience that the face is familiar
and/or the recollection of sufficient event details– then the witness may declare that they recognize the
face as having been previously encountered. Recognition memory decisions can thus be thought of as the
final stage in the process of eyewitness identification.
Because it is a form of memory retrieval, recognition memory is susceptible to all of the factors
summarized above that are known to interfere with retrieval. Recognition memory differs from other
forms of retrieval (such as recalling a phone number or a cake recipe), however, in that a comparison
must be made between the retrieved evidence and a decision threshold. That is, as noted above,
recognition judgments require a decision criterion, an understanding of which presents a unique set of
challenges for eyewitness identification (and recognition memory, generally). In particular, an observer’s
report of recognition (or, in a lineup setting, of identification) is influenced not simply by the strength or
quality of the recalled memory evidence. The report of recognition (identification of a lineup member) is
also influenced by the level of evidence that the observer finds acceptable to reach such a decision, i.e., by
his/her decision criterion, or bias. An observer who holds a liberal criterion will likely recognize many
true targets (i.e., the guilty), but will frequently err by reporting recognition of many false targets (i.e.,
innocents). Conversely, an observer who holds a conservative criterion will avoid the problem of
erroneous recognition (identification), but will fail to identify some true targets. Estimating (or
controlling) the observer’s decision criterion is thus a critical step in efforts to judge the validity of an
identification (see also Chapter 5).
Recognition memory for faces differs greatly between familiar and unfamiliar faces.87 Because we
often identify familiar individuals with ease, we tend to think we are generally very good at face
recognition. However, we are not as good with unfamiliar faces.88 All the sources of noise during
perception and memory contribute to these difficulties, and they are exacerbated by criminals’ attempts to
conceal their identity (even a change in hairstyle and clothing can have a major effect on recognition).
The ability to recognize unfamiliar faces differs widely across individuals. At one extreme are those
people, referred to as “super recognizers,”89 who hardly forget a face. At the other end of the spectrum are
“face-blind people (prosopagnosics),” who have great difficulty recognizing even highly familiar faces.90
Current estimates of the fraction of the general population afflicted by prosopagnosia are as high as
~2%.91 The ability of an eyewitness to identify a suspect may thus differ greatly from individual to
individual simply as a consequence of general variations in face recognition ability.
87

P. J. B. Hancock et al., “Recognition of unfamiliar faces,” Trends in Cognitive Sciences 4 (9): 330–337
(2000).
88
V. Bruce, “Changing faces: visual and non-visual coding processes in face recognition,” British Journal of
Psychology 73: 105–116 (1982).
89
R. Russell, B. Duchaine, and K. Nakayama, “Super-recognisers: people with extraordinary face recognition
ability,” Psychonomic Bulletin and Review 16 (2): 252–272 (2009).
90
T. Susilo and B. Ducahine, “Advances in developmental prosopagnosia research,” Current Opinion in
Neurobiology 23 (3):423–429 (2013).
91
I. Kennerknecht, T. Grueter, B. Welling, S. Wentzek, J. Horst, S. Edwards and M. Grueter, “First report of
prevalence of nonsyndromic hereditary prosopagnosia (HPA),” American Journal of Medical Genetics Part A 140
(15):1617–1622 (2006).

47
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
CONCLUSION
The shortcomings of eyewitness identification present a societal problem that has profound
implications for our systems of law and justice. Ultimately, a solution to this problem must be informed
by a thorough understanding of human vision and memory. The processes of vision and memory, which
are fundamental to human experience, have been frequent targets of scientific investigation since the 19th
century. The past few decades have seen an explosion of additional research that has led to important
insights into how vision and memory work, what we see and remember best, and what causes these
processes to fail. The committee has reviewed much of this research, as it pertains to eyewitness
identification and has identified restrictions on what can be seen under specific environmental and
behavioral conditions (such as poor illumination, limited viewing duration or viewing angle), factors that
impede the ability to attend to critically informative features of a visual scene (e.g., the deleterious effect
of an attention-grabbing element, such as a weapon, on the ability to correctly perceive the features of the
assailant’s face), distortions of perceptual experience derived from expectations, and ways in which
emotion and stress enhance or suppress specific perceptual experiences. Memory is often far from a
faithful record of what was perceived through the sense of sight: its contents can be forgotten or
contaminated at multiple stages, it can be biased by the very practices designed to elicit recall, and it is
also heavily swayed by emotional states associated with witnessed events and their recall. From this
analysis, the committee must conclude that there are insurmountable limits on vision and memory
imposed by our biological nature and the properties of the world we inhabit. With this knowledge, we can
more fully appreciate the value and risks associated with eyewitness reports and accordingly advise those
who collect, handle, defend, consider, and adjudicate such reports.

48
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

5
Applied Eyewitness Identification Research
The committee was tasked with (1) critically assessing the existing body of scientific research on
eyewitness identification; (2) identifying gaps in the literature; and (3) suggesting other research that
would further the understanding of eyewitness identification and improve law enforcement and courtroom
practice. Eyewitness identification research resides in both the scientific literature and the law and justicerelated scholarly literature. Although experiential, anecdotal, and some administrative records from law
enforcement and the judiciary could contribute to a better understanding of eyewitness identification, the
committee did not comprehensively review this more qualitative material. The committee did, however,
examine select examples of law enforcement policies and influential judicial rulings.
In late 2013, the committee compiled an extensive and comprehensive bibliography from the
following nine electronic databases, with the search limited to publications over the past two decades (i.e.,
since 1993): Academic Search Premier (EBSCO), Embase (Elsevier), MEDLINE (National Library of
Medicine), NCJRS Abstracts Database (U.S. Department of Justice), PsycINFO (American Psychological
Association), PubMed (National Institutes of Health), Scopus (Elsevier), Web of Science (Thomson
Reuters), and LexisNexis.1 Papers were drawn from such fields as social science, cognitive science,
behavioral science, neuroscience, criminology, and law using Boolean-logic-based search strategies
designed to identify empirical research reports, review articles, systematic reviews, meta-analyses, and
articles in law reviews and legal journals.
The committee concentrated its review on the subset of the bibliography deemed most important to
its task, focusing more on the scientific literature than on the law review literature. These materials
included meta-analyses and systematic reviews and primary research in neuroscience, statistics, and
eyewitness identification. This report also was informed by several early foundational papers and written
comments from, and presentations to, the committee by representatives from science, law enforcement,
state courts and government, private organizations, and other interested parties. The comments and
presentations revealed additional highly relevant new findings, some recently published or in press and
others in submission. The agenda for each committee meeting is available in Appendix B. All materials
submitted to the committee are retained in the Academies’ public access file and are available upon
request.
COMMITTEE ASSESSMENT
Many factors affect eyewitness accuracy. Some factors are related to protocols within the law
enforcement and legal systems, while others are related to characteristics associated with the crime scene,
perpetrator, and witness.
System variables are those that the criminal justice system can influence through the enforcement of
standards and through education and training of law enforcement personnel in the use of best practices2
and procedures (e.g., by specifying the content and nature of instructions given to witnesses prior to a
1

The law review literature was represented by the citations from the LexisNexis search. While all these
materials were not reviewed in detail, several of the documents informed Chapter 3 of this report (The Legal
Framework for Assessment of Eyewitness Identification).
2
As noted in Chapter 1, for the purposes of this report, the committee characterizes best practice as the adoption
of standardized procedures based on scientific principles. The committee does not make any endorsement of
practices designated as best practices by other bodies.

49
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
lineup identification). Estimator variables include factors operating either at the time of the criminal event
(relating to visual experience or memory encoding) or during the retention interval (the time between
witnessing an event and the identification process). Specific examples include the eyewitness’ level of
stress or trauma at the time of the incident, the light level and nature of the visual conditions that affect
visibility and clarity of a perpetrator’s features, similarity of age and race of the witness and perpetrator,
presence or absence of a weapon during the incident, and the physical distance separating the witness
from the perpetrator.
A scientific consensus about the effects of some factors has emerged, but no such consensus exists
for many other factors. One method of assessing scientific consensus is by surveys of experts. A 2001
survey collected responses from sixty-four psychologists about their courtroom experiences and their
opinions on thirty eyewitness-related phenomena to determine the “general acceptance” of these
phenomena within the eyewitness identification research community.3 General acceptance is relevant to
whether scientific testimony is admissible as evidence in court (see Chapter 3). The survey revealed
substantial agreement about which findings these experts felt were sufficiently reliable to present in
court.4
The committee examined the scientific literature on eyewitness identification, focusing first on
quantitative syntheses, largely systematic reviews and meta-analyses, which were identified in a
comprehensive search of electronic databases designed to locate research on both estimator and system
variables. In addition, primary research studies were identified in this database search, many of which
were also highlighted in the relevant systematic reviews and meta-analyses. Finally, some researchers
forwarded manuscripts to the committee that have been submitted for peer-review or are in press. In their
examination of this body of literature, the committee examined the quality of the identified research and,
where possible, worked to derive summary empirical generalizations related to variables of interest.
Quantitative Syntheses of Eyewitness Identification Research
The committee first evaluated the consistency of research findings across studies for system and
estimator variables by studying published quantitative reviews of empirical research. Systematic reviews,
which collect and appraise available research on specific hypotheses or research questions, are efforts to
synthesize the effects of variables across studies. Within systematic reviews, meta-analysis is often, but
not always, used to compute the effects of variables as well as to identify factoss that explain differences
across studies. When assumptions about consistency of data collected across studies are met, metaanalysis provides a quantitative summary of empirical findings by statistically averaging effect sizes
across individual studies, thereby increasing the precision of the effect size estimate as well as the
statistical power to detect effects. Done well, systematic reviews with or without meta-analysis provide
evidence for practice and policy for such fields as health care,5 crime and justice, social welfare, and
education.6 The utility of systematic reviews for informing practice and policy is predicated on the
included studies being transparently reported, conducted so as to minimize risk of bias, and representing
3

S. M. Kassin et al., “On the ‘General Acceptance’ of Eyewitness Testimony Research: A New Survey of the
Experts,” American Psychologist 56(5): 405–416 (2001).
4
Kassin et al. also compared the reliability assessments of the 2001 survey to assessments from a similar 1989
survey and noted that, for the seventeen propositions retested, there was a remarkable degree of consistency: “most
experts saw as sufficiently reliable expert testimony on the wording of questions,” lineup instructions, attitudes and
expectations, the accuracy-confidence correlation, the forgetting curve, exposure time, and unconscious
transference. “There was less, if any, consensus on the effects of color perception in monochromatic light,”
“observer training, high levels of stress, the accuracy of hypnotically refreshed testimony, and event violence.”
Kassin et al. observed that two phenomena were seen as significantly more reliable than had been the case when the
initial survey was conducted: weapon focus effect and hypnotic suggestibility effects. Ibid, 410.
5
See the Cochrane Collaboration, available at: http://www.cochrane.org.
6
See the Campbell Collaboration, available at: http://www.campbellcollaboration.org.

50
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
as complete a sample as possible of research conducted on the central question, including both published
and unpublished studies. In turn, systematic reviews should specify inclusion criteria and data extraction
procedures a priori, use independent and duplicate procedures for study selection and data extraction,
rigorously evaluate potential biases in included studies, and interpret results of meta-analyses in terms
that are useful to decision makers. Further, meta-analyses should not be conducted outside the context of
systematic reviews. In short, both systematic reviews and the studies they include need to be transparent
and reproducible in order to best inform practice and policy decisions about eyewitness identification.
The committee examined quantitative reviews that covered decades of research on both estimator
variables (exposure time,7 retention interval,8 stress,9 weapon focus,10 own-race bias,11 and own-age
bias12) and system variables (identification test medium, i.e., live lineup versus photo array ,13 biased and
unbiased lineup instructions,14 post-identification feedback,15 simultaneous versus sequential lineup
presentation,16 target absent versus target present lineups,17 foil similarity,18 blinding,19 showup versus
7

B. H. Bornstein et al., “Effects of exposure time and cognitive operations on facial identification accuracy: A
meta-analysis of two variables associated with initial memory strength,” Psychology, Crime and Law 18(5): 473–
490 (2012).
8
K. A. Deffenbacher et al., “Forgetting the once-seen face: estimating the strength of an eyewitness’s memory
representation,” Journal of Experimental Psychology: Applied 14(2): 139–150 (2008).
9
K. A. Deffenbacher et al., “A meta-analytic review of the effects of high stress on eyewitness memory,” Law
and Human Behavior 28(6): 687–706 (2004).
10
J. M. Fawcett et al., “Of guns and geese: a meta-analytic review of the ‘weapon focus’ literature,” Psychology
Crime & Law 19(1): 35–66 (2013).
11
C. A. Meissner and J. C. Brigham, “Thirty years of investigating the own-race bias in memory for faces—A
meta-analytic review,” Psychology, Public Policy, and Law 7(1): 3–35 (2001).
12
M. G. Rhodes and J. S. Anastasi, “The Own-Age Bias in Face Recognition: A Meta-Analytic and Theoretical
Review,” Psychological Bulletin 138(1): 146–174 (2012).
13
B. L. Cutler et al., “Conceptual, practical, and empirical issues associated with eyewitness identification test
media,” in Adult Eyewitness Testimony: Current Trends and Developments, ed. D. F. Ross (New York: Press
Syndicate of the University of Cambridge, 1994): 163–181.
14
S. E. Clark, “A re-examination of the effects of biased lineup instructions in eyewitness identification,” Law
and Human Behavior 29(4): 395–424 (2005). S. E. Clark, “Costs and Benefits of Eyewitness Identification Reform:
Psychological Science and Public Policy,” Perspectives on Psychological Science 7(3): 238–259 (2012). N. K.
Steblay, “Social influence in eyewitness recall: A meta-analytic review of lineup instruction effects,” Law and
Human Behavior 21(3): 283–97 (1997). N. K. Steblay et al., “The Eyewitness Post Identification Feedback Effect 15
Years Later: Theoretical and Policy Implications,” Psychology, Public Policy, & Law 20(1): 1–18 (2014).
15
S. E. Clark and R. D. Godfrey, “Eyewitness identification evidence and innocence risk,” Psychonomic
Bulletin & Review 16(1): 22–42 (2009). A. B. Douglass and N. K. Steblay, “Memory distortion in eyewitnesses: A
meta-analysis of the post-identification feedback effect,” Applied Cognitive Psychology 20(7): 859–869 (2006).
16
S. E. Clark, “Costs and Benefits of Eyewitness Identification Reform: Psychological Science and Public
Policy,” Perspectives on Psychological Science 7(3): 238–259 (2012). S. E. Clark et al., “Regularities in eyewitness
identification,” Law and Human Behavior 32(3): 187–218 (2008). N. K. Steblay et al., “Eyewitness accuracy rates
in sequential and simultaneous lineup presentations: A meta-analytic comparison,” Law and Human Behavior 25(5):
459–473 (2001). N. K. Steblay et al., “Seventy-two Tests of the Sequential Lineup Superiority Effect: A MetaAnalysis and Policy Discussion,” Psychology, Public Policy, and Law 17(1): 99–139 (2011).
17
S. E. Clark, “A re-examination of the effects of biased lineup instructions in eyewitness identification,” Law
and Human Behavior 29(4): 395–424 (2005). S. E. Clark et al., “Regularities in eyewitness identification,” Law and
Human Behavior 32(3): 187–218 (2008). S. E. Clark and R. D. Godfrey, “Eyewitness identification evidence and
innocence risk,” Psychonomic Bulletin & Review 16(1): 22–42 (2009).
18
S. E. Clark, “Costs and Benefits of Eyewitness Identification Reform: Psychological Science and Public
Policy,” Perspectives on Psychological Science 7(3): 238–259 (2012). S. E. Clark and R. D. Godfrey, “Eyewitness
identification evidence and innocence risk,” Psychonomic Bulletin & Review 16(1): 22–42 (2009). S. E. Clark et al.,
“Regularities in eyewitness identification,” Law and Human Behavior 32(3): 187–218 (2008). R. J. Fitzgerald et al.,
“The Effect of Suspect-Filler Similarity on Eyewitness Identification Decisions: A Meta-Analysis,” Psychology,

51
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
lineup,20 prior mug shot exposure,21 verbal description and identification,22 and the Cognitive Interview23).
Many of these quantitative reviews were published recently, with more than one-third of them published
since 2010. However, none of the reviews met all current standards for conducting and reporting
systematic reviews,24 and few met even a majority of these standards, making assessment of the
credibility of their findings problematic.
After examining the reviews, the committee concluded that the findings may be subject to
unintended biases and the conclusions are less credible than was hoped. In many cases, the data from the
studies being cited were not readily available or were not clearly presented. Nevertheless, these reviews
were helpful in highlighting some of the issues associated with specific research questions and in
identifying primary studies that might be both credible and important.
RESEARCH STUDIES ON SYSTEM VARIABLES
After its assessment of the systematic reviews and meta-analytic studies, the committee’s review
focused on the most-studied system variables. We note that key system variables, such as lineup
procedures (e.g. simultaneous vs. sequential lineups, blinded vs. non-blinded lineup administration) and
the collection/use of witness confidence statements, can have a marked influence over the validity of
eyewitness identifications. In the following section, we address one of the most important practical issues
raised by this influence: What is the best way to evaluate the effects of system variables on the diagnostic
accuracy of eyewitness reports and how might we use the results of such an evaluation to optimize the
states of key system variables and thus maximize performance of an eyewitness? This question is, in
principle, relevant to all system variables, but we address it first in the timely and controversial context of
simultaneous vs. sequential lineup presentations and in the role of eyewitness confidence judgments in
evaluation of identification performance. This examination of lineup procedures and confidence reports is
followed by a brief discussion of the effects on eyewitness performance of another important system
variable: the extent and content of communications between the witness and the larger community (law
enforcement, legal defense, the press, family and friends).
Evaluating Eyewitness Performance
Perhaps the most important empirical question that can be asked about eyewitness identification is:
How well do witnesses perform as a function of different system and estimator variables? For example,
Public Policy, and Law 19(2): 151–164 (2013). S. L. Sporer et al., “Choosing, confidence, and accuracy: A metaanalysis of the confidence-accuracy relation in eyewitness identification studies,” Psychological Bulletin 118(3):
315–327 (1995).
19
S. E. Clark, “Costs and Benefits of Eyewitness Identification Reform: Psychological Science and Public
Policy,” Perspectives on Psychological Science 7(3): 238–259 (2012).
20
S. E. Clark, “Costs and Benefits of Eyewitness Identification Reform: Psychological Science and Public
Policy,” Perspectives on Psychological Science 7(3): 238–259 (2012). N. K. Steblay et al., “Eyewitness accuracy
rates in police showup and lineup presentations: A meta-analytic comparison,” Law and Human Behavior 27(5):
523–540 (2003).
21
K. A. Deffenbacher et al., “Mugshot exposure effects: Retroactive interference, mugshot commitment, source
confusion, and unconscious transference,” Law and Human Behavior 30(3): 287–307 (2006).
22
C. A. Meissner et al., “A theoretical review and meta-analysis of the description-identification relationship in
memory for faces,” European Journal of Cognitive Psychology 20(3): 414–455 (2008).
23
A. Memon et al., “The Cognitive Interview: A meta-analytic review and study space analysis of the past 25
years,” Psychology, Public Policy, and Law 16 (4): 340–372 (2010).
24
See, e.g., Institute of Medicine, Finding what works in health care: Standards for systematic reviews
(Washington, D.C.: The National Academies Press, 2011); B. J. Shea et al., Development of AMSTAR: a
measurement tool to assess the methodological quality of systematic reviews, BMC Medical Research Methodology
2007, 7:10 doi:10.1186/1471-2288-7-10.

52
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
do factors such as the structure of a lineup, stress, or weapon focus affect the ability of a witness to
provide reliable information? If so, what practices will yield the best performance? The issues are
multifaceted, and the answers likely depend upon many factors. Given the complexity of these issues, the
experimental literature to date has focused largely on one of the more tractable problems: How do
different lineup identification procedures affect witness identifications. We will use this focus (and its
eminent practical relevance) to illustrate how one might go about evaluating eyewitness performance
generally.
Most lineup identification procedures take one of two forms: simultaneous or sequential. In a
simultaneous procedure, the witness views all individuals in the lineup at the same time, and either
identifies one (or more) as the perpetrator, or reports that the person (s)he saw at the crime scene was not
in the lineup. In a sequential procedure, the witness views individuals one at a time and reports whether or
not each one is the person from the crime scene. Rigorous evaluation of eyewitness identification
performance as a function of these two procedures requires a formal understanding of the task that the
witness confronts, and it requires criteria for assessing the outcome.
The task of a witness viewing a lineup is an example of what is known as a binary classification
problem.25 Each eyewitness faces two possible (binary) states associated with each person in the lineup
(guilt or innocence) and the witness must assign each person to one of two classes (guilty or innocent).
For each decision, the witness can be correct or incorrect, yielding four possible outcomes: a correct
classification as guilty (“hit”), an incorrect classification as guilty (“false alarm”), a correct classification
as innocent (“correct rejection”), and an incorrect classification as innocent (“miss”). These outcomes are
commonly presented in a contingency table26 (Table 5-1), and the frequencies in each part of that table are
the raw data used to evaluate performance on a binary classification task, such as eyewitness
identification.27
TABLE 5-1 Contingency Table for Possible Eyewitness Identification Outcomes

SOURCE: Courtesy of Thomas D. Albright.

25

The binary classifier in this context is defined as the witness operating under a specific set of conditions, such
as lineup procedures.
26
Also termed “confusion matrix.”
27
The prevalence or “base-rate” – the fraction of individuals in each category (guilty or innocent, in the
eyewitness problem) in the population is also a factor that may come into play when evaluating binary classification
performance.

53
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
There are many different performance measures that can be derived from data of this sort – indeed,
the fields of statistical classification and machine learning are replete with tools for the evaluation of
binary classifiers.28 The preferred measure will depend to a large degree upon the criteria one adopts for
performance evaluation.
Perhaps the simplest measure of binary classification performance is the ratio of hit rates (HR)29 to
false alarm rates (FAR), i.e., HR/FAR. The magnitude of this measure, which is known in the eyewitness
identification literature as the “diagnosticity ratio,”30 is proportional to the likelihood that a classification
is correct, i.e., that the person identified as guilty is actually guilty. The diagnosticity ratio is appealing if
the criterion that matters most is avoiding erroneous identifications.
Not surprisingly, the diagnosticity ratio was adopted in early pioneering efforts to identify lineup
conditions that would yield better witness identification performance.31 Most laboratory-based studies and
meta-analyses of the effects of lineup procedures on eyewitness identification performance show that,
with standard lineup instructions informing the witness that the perpetrator may or may not be present, the
sequential procedure produces a higher diagnosticity ratio.32 That is, when considering only those cases in
which a witness actually selects someone from a lineup, the ratio of correct to false identifications is
commonly higher with the sequential than with the simultaneous procedure.33
A higher diagnosticity ratio could result from a higher hit rate, a lower false alarm rate, or some
combination of the two. Some early reports suggested that sequential procedures (relative to
simultaneous) lead to fewer false alarms without changing the frequency of hits, which would result in a
higher diagnosticity ratio.34 More recent laboratory-based studies and meta-analyses typically show that
28

See, e.g., T. Hastie, R. Tibshirani, and J.H. Friedman, The Elements of Statistical Learning: Data Mining,
Inference, and Prediction (New York: Springer, 2009); A. Smola and S.V.N. Vishwanathan, Introduction to
Machine Learning (Cambridge: Cambridge University Press, 2008).
29
The “rate” associated with each cell of the contingency table is computed as the number of counts within that
cell (e.g. number of people correctly classified as guilty) divided by the number of instances that are truly in that
class (e.g. total number of guilty people being classified). Thus, hit rate (HR) = number of hits / (number of
hits+number of misses), and false alarm rate (FAR) = number of false alarms / (number of false alarms+number of
correct rejections).
30
The “diagnosticity ratio” is also known in other disciplines by other names; e.g., “positive likeli-hood ratio” or
“LR+ = Likelihood Ratio of a Positive Call;” see Peter Lee, Bayesian Statistics: An Introduction (Chichester: Wiley,
2012), Sec 4.1.
31
R. C. L. Lindsay and G. L. Wells, “Improving eyewitness identifications from lineups: Simultaneous versus
sequential lineup presentation,” Journal of Applied Psychology, 70(3), 556–564 (1985).
32
N. K. Steblay et al. “Eyewitness accuracy rates in sequential and simultaneous lineup pre-sentations: A metaanalytic comparison,” Law and Human Behavior 25(5): 459-473 (2001). N.K. Steblay, et al., “Seventy-two Tests of
the Sequential Lineup Superiority Effect: A Meta-Analysis and Policy Discussion,” Psychology, Public Policy, and
Law 17(1): 99–139 (2011). S.D. Gronlund et al., “Robustness of the sequential lineup advantage,” Journal of
Experimental Psychology: Applied, 15 (2): 140–152 (2009). S. D. Gronlund, J. T. Wixted, and L. Mickes,
“Evaluating eyewitness identification procedures using ROC analysis,” Current Directions in Psychological Science
23 (1): 3–10 (2014).
33
But see C.A. Carlson, S.D. Gronlund, and S.E. Clark, “Lineup composition, suspect position, and the
sequential lineup advantage,” Journal of Experimental Psychology-Applied 14(2): 118-128 (2008), for a
counterexample. Also, Clark, Moreland and Gronlund have demonstrated that the accuracy advantage of sequential
lineups as measured by diagnosticity ratios has decreased over time since the original report. Reanalysis of
diagnosticity data for sequential studies showed slight, non-significant decreases in correct identification effects and
increases in false identification effects, which together combine to produce a significant decrease in the advantage of
sequential over simultaneous lineup methods. See S.E. Clark, M.B. Moreland, and S.D. Gronlund, “Evolution of the
empirical and theoretical foundations of eyewitness identification reform,” Psychonomic Bulletin and Review 21(2):
251–267 (2014).
34
R. C. L. Lindsay, “Applying applied research: Selling the sequential lineup,” Applied Cognitive Psychology
13 (3): 219–225 (1999). G. L. Wells, S. M. Rydell, and E. P. Seelau, “The selection of distractors for eyewitness
lineups,” Journal of Applied Psychology 78 (5): 835–844 (1993).

54
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
sequential procedures (relative to simultaneous) are associated with a somewhat reduced hit rate
accompanied by a larger reduction in the false alarm rate, thereby resulting in diagnosticity ratios higher
than those yielded by simultaneous procedures.35 In other words, when using a single diagnosticity ratio
as a measure of eyewitness performance, the sequential procedure (relative to simultaneous) comes closer
to satisfying the popular criterion that those identified as guilty are actually guilty. In light of these
findings, many policy makers have advocated sequential procedures, and those procedures have been
adopted by law enforcement in many jurisdictions.
While policy decisions and practice have been influenced by the aforementioned studies, there are
other criteria worthy of consideration when evaluating eyewitness performance. One alternative is
revealed by asking why the diagnosticity ratio changes across lineup conditions. This question can be
addressed given a plausible model of the mechanisms underlying human recognition memory. Most
models of recognition memory are based on the idea that a cue (e.g., a face in a lineup) results in the
retrieval of information stored in memory (see Chapter 4). When the retrieved information provides
enough evidence to satisfy the observer, they make an identification – that is, they decide that the
stimulus is “recognized.” Explicit in this model are two important parameters: the observer’s memory
sensitivity (that is, the “discriminability” between the strength of memory evidence elicited by a
previously encountered stimulus and that elicited by novel stimuli), and the degree of evidence that the
observer requires to make an identification (“response criterion” or “bias”) (see Box 5-1).
The first of these two parameters—discriminability—is important for evaluating eyewitness
performance. It tells you whether a difference in performance under different task conditions reflects a
true improvement in memory-based discrimination, i.e., an improvement in the strength of the observer’s
retrieved memory evidence of the perpetrator.
The fact that these two measures (the likelihood that an identified person is guilty vs. discriminability)
do not assess the same thing is counterintuitive—a fact that has generated controversy in the field of
eyewitness identification research.36 Intuitively, if sequential lineups yield a higher likelihood that an
identified person is guilty (as quantified by a higher diagnosticity ratio), then it seems as if that procedure
yields objectively better performance. The problem with this intuition is that it fails to take into account the
second of the two parameters of recognition memory models—the response bias or degree of evidence that
35

A recent field-based study comparing sequential to simultaneous procedures in a limited number of
jurisdictions computed the diagnosticity ratio using filler identifications as the false alarm rate (because the
innocence or guilt of the suspect is unknown in such situations) [G. L. Wells, N. K. Steblay, J. E. Dysart, “DoubleBlind Photo-Lineups Using Actual Eyewitnesses: An Experimental Test of a Sequential versus Simultaneous Lineup
Procedure,” Law and Human Behavior, 15 June 2014, doi: 10.1037/lhb0000096.] When computed in this manner,
the data revealed a modest diagnosticity ratio advantage for the sequential procedure. However, Amendola and
Wixted re-analyzed a subset of the data for which proxy measures of ground truth were available [K. Amendola and
J.T. Wixted, “Comparing the Diagnostic Accuracy of Suspect Identifications Made by Actual Eyewitnesses from
Simultaneous and Sequential Lineups,” accepted by Journal of Experimental Criminology (2014).] Their analyses
suggested that identification of innocent suspects is less likely and identification of guilty suspects is more likely
when using the simultaneous procedures. While future field studies are needed, these latter findings raise the
possibility that diagnosticity is higher for the simultaneous procedure. See also Clark, Moreland and Gronlund ,who
report that published diagnosticity ratios have changed over time, reflecting a significant decrease in the advantage
of sequential over simultaneous lineup procedures. [S.E. Clark, M.B. Moreland, and S.D. Gronlund, “Evolution of
the empirical and theoretical foundations of eyewitness identification reform,” Psychonomic Bulletin and Review
21(2): 251–267 (2014)].
36
See, e.g., J. T. Wixted and L. Mickes, “The Field of Eyewitness Memory Should Abandon Probative Value
and Embrace Receiver Operating Characteristic Analysis,” Perspectives on Psychological Science 7(3): 275-278
(2012); S. E. Clark, “Costs and Benefits of Eyewitness Identification Reform: Psychological Science and Public
Policy,” Perspectives on Psychological Science 7(3): 238–259 (2012); G. L. Wells, “Eyewitness Identification
Probative Value, Criterion Shifts, and Policy Regarding the Sequential Lineup,” Current Directions in
Psychological Science 23(1): 11–16 (2014); N. K. Steblay, et al. “Seventy-two Tests of the Sequential Lineup
Superiority Effect: A Meta-Analysis and Policy Discussion,” Psychology Public Policy and Law 17(1): 99-139

55
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
the observer finds acceptable to make an identification. This parameter, which is distinct from
discriminability, reflects the witness’ tendency to pick or not to pick someone from the lineup. If a witness
sets a high bar for acceptable evidence—a conservative bias—then he/she will be unlikely to select anyone
from the lineup (low pick frequency), meaning that they will have more misses (they’ll be more likely to fail
to select the suspect because they are less likely to make a selection at all) and fewer false alarms.
Conversely, if a witness sets a low bar for acceptable evidence—a liberal bias—then he/she will be
more likely to make a selection from the lineup (a high pick frequency), meaning they will have more hits
and will make more false identifications. Differences in pick frequency can, and generally do, lead to
differences in the ratio of hit rates to false alarm rates; all else being equal, the diagnosticity ratio will be
higher for a conservative bias than for a liberal bias.37 In other words, simply by inducing a witness to adopt
a more conservative bias, it is possible to increase the likelihood that an identified person is actually guilty.
Importantly, this may be true even if the procedure yields no better, or potentially worse, discriminability.38
BOX 5-1 The Influences of Discriminability and Response Bias on Human Binary Classification Decisions
All human decisions about the classification of objects based on memory—including a witness’
classifications of guilt or innocence for faces in a lineup, an individual’s decision as to whether a piece of luggage
is his/her own, a botanist’s recognition of a specific type of fern, a radiologist’s detection of a tumor in a
mammogram, or the determination of the sex of a newly-hatched chicken—can be distilled down to the influence
of two factors that are rooted in causal models of recognition memory a : the degree to which the relevant objects
are discriminable by the decider (the decider’s sensitivity to the difference between them), and the decider’s
criterion for making a decision (response bias, or the decider’s degree of specificity in making choices). There are,
of course, many other variables that will affect the outcome, including levels of stress, attentional focus, potential
rewards or expectations, but all of these are believed to exert their influence over memory-based classification
decisions by affecting discriminability and/or response bias.
To illustrate the distinction between discrimination and response bias as applied to a real-world decision
problem, consider how an audiologist conducts a hearing test. In a hearing test, an individual might be asked to
detect sounds along a continuum of loudness and to indicate when a sound is present. The audiologist wants to
know how well someone can discriminate presence versus absence of a sound, but that assessment is complicated
by the criterion people use when deciding to say that they heard a sound (response bias). Some people are hesitant
to respond positively, saying “I hear it” only when they are absolutely certain (“conservative” responders). Others
are more willing to respond positively, saying “I hear it” with less information and greater uncertainty (“liberal”
responders). Those with a conservative bias are less likely to report hearing a sound in general, so they will have
both fewer correct detections (“hits”) and fewer overt mistakes (“false alarms”). By contrast, those with a liberal
bias are more likely to say that they heard a sound, so they will have more hits but also more false alarms.
Importantly, this can occur even if the conservative and liberal responders do not differ in their ability to
discriminate the presence from absence of sound.
a

See, e.g., W.P. Banks, “Signal detection theory and human memory,” Psychological Bulletin 74(2): 81-99 (1970);
J.P. Egan, Recognition memory and the operating characteristic (Bloomington: Indiana University Hearing and
Communication Laboratory, 1958); D. M. Green and J. A. Swets, Signal Detection Theory and Psychophysics
(New York: Wiley,1966).

37

The sole exception to this rule is the case in which classifications are made at chance level of performance,
i.e., when the observer exhibits no ability to discriminate.
38
L. Mickes, H.D. Flowe, and J. T. Wixted, “Receiver operating characteristic analysis of eyewitness memory:
Comparing the diagnostic accuracy of simultaneous vs. sequential lineups,” Journal of Experimental Psychology:
Applied 18 (4): 361–376 (2012). C. A. Meissner et al., “Eyewitness decisions in simultaneous and sequential
lineups: A dual process signal detection theory analysis,” Memory and Cognition 33 (5): 783–792 (2005). M.A.
Palmer and N. Brewer, “Sequential lineup presentation promotes less-biased criterion setting but does not improve
discriminability,” Law and Human Behavior 36 (3): 247–255 (2012).

56
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Despite its merits, a single diagnosticy ratio thus conflates the influences of discriminability and
response bias on binary classification, which muddies the determination of which procedure, if any, yields
objectively better discriminability in eyewitness performance. To overcome this problem, some
investigators have recently adopted a technique from signal detection theory,39 which distinguishes the
relative influences of discriminability and bias on binary classification. This technique involves analysis
of Receiver Operating Characteristics (see Box 5-2). ROC analysis has been used extensively in multiple
contexts of human decision-making, notably in basic research on visual perception and memory and
applied studies of medical diagnostic procedures.40 In essence, ROC analysis examines diagnosticity
ratios integrated over different response biases. This approach to eyewitness research has been promoted
based on the claim that it can enable lineup procedures to be evaluated by their effect on discrimination,
separate from response bias, and—importantly—because the dimensions of analysis (discriminability and
response bias) correspond to the mechanistic parameters of causal models of human recognition memory.
Use of ROC analysis to evaluate eyewitness performance requires calculating the diagnosticity ratio
for different response bias conditions (see Box 5-2). Using expressed confidence level (ECL) as a proxy
for response bias (see below), a small set of recent studies using ROC analysis has reported that
discriminability (area under the ROC curve) for simultaneous lineups is as high, or higher, than that for
sequential lineups.41 In other words, when eyewitness identification performance is evaluated based on a
criterion of bias-free discriminability, the results differ from those based on a single diagnosticity ratio,
and they do so because the latter fails to account for response bias.
Looking broadly at the many empirical studies that have used a single diagnosticity ratio to evaluate
eyewitness performance, as well as the more recent findings using ROC analysis, it appears that the
practical advantage of one lineup procedure over another depends to a large degree upon the performance
criterion that one adopts. From the perspective of many, the ideal lineup procedure would elicit a
conservative bias (thus reducing false identifications) and high discriminability (that is, optimizing
memory sensitivity). If there exists no discriminability advantage for one lineup procedure over another,42
then eyewitness performance may benefit from any procedure (such as sequential) that elicits a more
conservative response bias.43 But one can only make that judgment after having applied an empirical test
39

D. M. Green and J. A. Swets, Signal Detection Theory and Psychophysics (New York: Wiley,1966); D.
McNicol, A Primer of Signal Detection Theory (London: George Allen & Unwin, 1972).
40
J. A.Swets, “ROC Analysis Applied to the Evaluation of Medical Imaging Techniques,” Investigative
Radiology 14 (2): 109–121 (1979).
41
L. Mickes, H. D. Flowe, and J. T. Wixted, “Receiver operating characteristic analysis of eyewitness memory:
Comparing the diagnostic accuracy of simultaneous and sequential lineups,” Journal of Experimental Psychology:
Applied 18 (4): 361–376 (2012). C. A. Carlson and M. A. Carlson, “An evaluation of lineup presentation, weapon
presence, and a distinctive feature using ROC analysis,” Journal of Applied Research in Memory and Cognition
3(2): 45–53 (2014). D. G. Dobolyi and C. S. Dodson, “Eyewitness confidence in simultaneous and sequential
lineups: a criterion shift account for sequential mistaken identification overconfidence,” Journal of Experimental
Psychology: Applied 19 (4): 345–357 (2013). S. D. Gronlund et al., “Showups versus lineups: An evaluation using
ROC analysis,” Journal of Applied Research in Memory and Cognition 1 (4): 221–228 (2012).
42
The committee notes that some of the few recent reports using ROC analysis indeed claim improved
discriminability for simultaneous lineup conditions, but the reported discriminability improvements are small.
43
In reality, a more conservative bias may not always be beneficial, and whether it is or not depends upon a
number of factors that have an impact distinct from diagnostic accuracy and are difficult to quantify. All else being
equal, the “best” response bias (HR/FAR) will be one that maximizes the “expected value” of the outcome [D. M.
Green and J. A. Swets, Signal Detection Theory and Psychophysics (New York: Wiley, 1966). J.A.Swets, “ROC
Analysis Applied to the Evaluation of Medical Imaging Techniques,” Investigative Radiology 14 (2): 109–112
(1979)]. For the problem of eyewitness identification, the response bias that maximizes expected value can be
computed from the prevalence of guilty suspects in lineups and from societal values or costs associated with each of
the possible eyewitness decisions (errors and correct assignments). Reliable data on prevalence are difficult to come
by, and value/cost quantities are difficult to assign and likely to vary significantly across crimes and cultures. One
can nonetheless gain an intuition for how these factors might define the best response bias conditions. Consider, for

57
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
to determine whether one or another procedure offers a discriminability advantage. Future research should
also explore the possibility that other methods of inducing a conservative response bias (such as verbal
instructions to the witness to be cautious in making an identification) might be combined with procedures
that improve discriminability, in order to optimize eyewitness identification performance.

BOX 5-2 Analysis of Receiver Operating Characteristics (ROC)
Binary classification decisions by human observers are affected by both discriminability (the observer’s
sensitivity to the difference between target and non-targets) and response bias (the observer’s degree of specificity
in making a response). Analysis of Receiver Operating Characteristics (ROC) is a method from signal detection
theory that enables one to distinguish the relative influences of discriminability and response bias on binary
classification decisions. ROC analysis is performed by plotting the frequency of decisions that are hits (correctly
detecting a target) vs. the frequency of decisions that are false alarms (incorrectly classifying a non-target as a
target).

FIGURE 5-1 Receiver Operating Characteristic (ROC) Curve
SOURCE: Courtesy of Thomas D. Albright.

The positive diagonal in an ROC plot corresponds to response bias, moving from high specificity at the
lower left corner [no detection of targets (hit rate = 0) and no incorrect attribution of non-targets as targets (false
alarm rate = 0)], to low specificity at the upper right corner [all targets detected (hit rate = 1.0) and all non-targets
attributed as targets (false alarm rate = 1.0)]. Because all points along this positive diagonal reflect equal ratios of
hits to false alarms, they vary in response bias (i.e., the frequency of lineup picks, or “pick frequency”), but they
(Continued)

example, the consequences of decreasing the prevalence of guilty suspects in lineups. In this case, expected value
can be maximized by inducing a conservative bias – i.e., if innocence is a priori likely, then there is value gained by
being more selective in your response. Similarly, the optimal response bias will depend upon normative costs
associated with different types of eyewitness errors. Generally speaking, if a society places greater emphasis on not
identifying the innocent, relative to failing to identify the guilty, then expected value can be increased by inducing a
more conservative response bias. But the opposite would be true if there were greater societal pressures for
identifying the guilty, relative to protecting the innocent. Importantly, although an understanding of the relationship
between response bias and expected value is important, expected value in this case has little to do with the
diagnostic accuracy of an eyewitness report. But it does nonetheless bear on decisions about which lineup procedure
should be employed.

58
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

BOX 5-2 Continued
do not manifest differences in discriminability. The negative diagonal in an ROC plot corresponds, by contrast, to
discriminability, moving from chance discriminability at the intersection with the positive diagonal, where hits
and false alarms are equally likely, to the highest discriminability in the upper left corner, where all targets are
detected (hit rate = 1.0), but no non-targets are attributed as targets (false alarm rate = 0).
To see how measured hit and false alarm rates vary over different conditions of discriminability and response
bias in laboratory experiments, one can manipulate or estimate these conditions and record a diagnosticity ratio
(HR/FAR) for each condition. The typical result is a set of diagnosticity ratios that, when plotted in the ROC
space (represented by the dots in above figure), form a curve spanning from lower left to upper right. The extent
to which that curve deviates (bows above and away) from the positive diagonal is a quantitative measure of
discriminability (assessed as the area under the curve) for which response bias has been factored out.
ROC analysis has been used extensively in basic and applied research on recognition memory. In these
experiments, response bias is sometimes manipulated explicitly by encouraging observers to be more or less
selective in their responses. Frequently, however, “expressed confidence level” (ECL)—the confidence that an
observer holds in his/her classification—is used as a proxy for response bias, based on the assumption that more
confident observers are likely to be more specific (conservative) in their responses, whereas less confident
observers are likely to be less specific (liberal) in their responses.

Perhaps the greatest practical benefit of recent debate over the utility of different lineup procedures
is that it has opened the door to a broader consideration of methods for evaluating and enhancing
eyewitness identification performance. ROC analysis is a positive and promising step, with numerous
advantages. For example, the area under the ROC curve is a single-number index of discriminability.
Moreover, this index reflects a parameter-free approach to binary classification performance; the outcome
is entirely data-dependent and thus identical across all users drawing from the same data set.44 Most
importantly for its application to the problem of evaluating eyewitness performance, the ROC approach
possesses a distinct advantage because the dimensions of analysis – discriminability and response bias –
map directly onto the mechanistic parameters of causal models of human recognition memory (see
Chapter 4). In other words, the approach affords insight into and quantification of the sensory and
cognitive processes that are believed to underlie memory-based classification decisions (see Box 5-1),
such as eyewitness identifications.
Despite these merits, as a general statistical procedure for evaluation of binary classification
performance, and as a tool for evaluation of eyewitness performance, the ROC approach has some welldocumented quantitative shortcomings. For example, ROC analysis depends on the ability to manipulate
response bias or to estimate it from some other variable, and in the case of eyewitness identification that
ability has been the subject of some debate. Recent studies have used expressed confidence level (ECL) –
a measure of a witness’ confidence in his/her selection – as a proxy for response bias,45 based on the
common-sense logic that a witness who has high confidence in their lineup selection should manifest a
more conservative response bias than a witness who selected someone from the lineup despite lacking
confidence in that selection (i.e., someone who made a selection even though they were not certain – a
44

D. M. Green and J. A. Swets, Signal Detection Theory and Psychophysics (New York: Wiley,1966). D.J.
Hand, “Measuring classifier performance: A coherent alternative to the area under the ROC curve,” Machine
Learning 77, 103–123 (2009).
45
See, e.g., N. Brewer and G. L. Wells, “The confidence-accuracy relationship in eyewitness identification:
Effects of lineup instructions, foil similarity, and target-absent base rates,” Journal of Experimental Psychology:
Applied 12(1): 11–30 (2012); L. Mickes, H. D. Flowe, and J. T. Wixted, “Receiver operating characteristic analysis
of eyewitness memory: Comparing the diagnostic accuracy of simultaneous and sequential lineups,” Journal of
Experimental Psychology: Applied 18 (4): 361–376 (2012); C.A. Carlson and M.A. Carlson, “An evaluation of
lineup presentation, weapon presence, and a distinctive feature using ROC analysis,” Journal of Applied Research in
Memory and Cognition 3(2): 45–53 (2014).

59
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
liberal response bias). This proxy relationship is inherently noisy within individuals, and the noisy
relationship is exacerbated by the fact that the eyewitness identification ROC is population-based;
individual data points are obtained from different people who may scale their confidence reports
differently.46 On the other hand, it is empirically clear that, when scaled appropriately (within and across
individuals), different levels of expressed confidence do, in fact, correspond to different pick frequencies
and response biases.47
An additional prerequisite for the use of ECL as a measure of response bias is that an orderly
relationship exists between confidence and accuracy – that witnesses expressing greater confidence are
more likely to be accurate in their identifications. Although this hypothesis conforms to intuition,48 the
existence of a significant confidence-accuracy relationship has been challenged repeatedly over the
years49 Recent evidence, however, suggests ways of improving the confidence-accuracy relationship (and
obtaining more reliable measurements of it).50 While the ECL measure thus has potential, more research
46

ECL is affected by over-confidence and under-confidence at the individual level, and the current
implementation of the ROC approach, combining results across subjects, does not build this measurement error into
the analysis or the comparison of empirical ROC curves. See Appendix C.
47
See, e.g., Table 1 of L. Mickes, H. D. Flowe, and J. T. Wixted, “Receiver operating characteristic analysis of
eyewitness memory: Comparing the diagnostic accuracy of simultaneous and sequential lineups,” Journal of
Experimental Psychology: Applied 18 (4): 361–376 (2012),which summarizes confidence ratings, hit rates, false
alarm rates, and diagnosticity ratios (HR/FAR) derived from data published by N. Brewer and G. L. Wells, “The
confidence-accuracy relationship in eyewitness identification: Effects of lineup instructions, foil similarity, and
target-absent base rates,” Journal of Experimental Psychology: Applied 12(1): 11–30 (2006), who employed a
“confidence calibration” technique to normalize scaling of expressed confidence across witnesses. Both hit rates and
false alarm rates declined steeply—implying an increasingly conservative response bias—as confidence levels
increased. Diagnosticity ratios increased monotonically with increasing confidence. An identical pattern can be seen
in Table 3 of L. Mickes, H. D. Flowe, and J. T. Wixted, “Receiver operating characteristic analysis of eyewitness
memory: Comparing the diagnostic accuracy of simultaneous and sequential lineups,” Journal of Experimental
Psychology: Applied 18 (4): 361–376 (2012), which summarizes data from their own experiments. See also H. L.
Roediger III, J. T. Wixted, and K. A. DeSoto, “The curious complexity between confidence and accuracy in reports
from memory,” in Memory and Law, ed. L. Nadel and W. Sinnott-Armstrong (Oxford: Oxford University Press,
2012), 97.
48
K. A. Deffenbacher and E. F. Loftus, “Do jurors share a common understanding concerning eyewitness
behavior?,” Law and Human Behavior 6: 15-30 (1982). G. L. Wells, T. J. Ferguson, and R. C. L. Lindsay, “The
tractability of eyewitness confidence and its implication for triers of fact,” Journal of Applied Psychology 66: 688696 (1981).
49
G. L. Wells and D. M. Murray, “Eyewitness confidence,” in Eyewitness Testimony: Psychological
Perspectives, ed. G. L. Wells and E. F. Loftus (New York: Cambridge University Press, 1984). B. L. Cutler and S.
D. Penrod, Mistaken Identification: The Eyewitness, Psychology, and the Law (Cambridge: Cambridge University
Press, 1995). R.K. Bothwell, K.A. Deffenbacher, and J.C. Brigham,“Correlation of eyewitness accuracy and
confidence: Optimality hypothesis revisited,” Journal of Applied Psychology, 72, 691-695 (1987). S. L. Sporer et al.,
“Choosing, confidence, and accuracy: A meta-analysis of the confidence-accuracy relation in eyewitness
identification studies.” Psychological Bulletin 118(3): 315-327 (1995). T. A. Busey et al., “Accounts of the
confidence-accuracy relation in recognition memory,” Psychonomic Bulletin and Review 7(1): 26-48 (2000).
50
N. Brewer and G. L. Wells, “The confidence-accuracy relationship in eyewitness identification: Effects of
lineup instructions, foil similarity, and target-absent base rates,” Journal of Experimental Psychology: Applied
12(1): 11–30 (2006). Juslin et al.,“Calibration and Diagnosticity of Confidence in Eyewitness Identification:
Comments on What Can Be Inferred From the Low Confidence-Accuracy Correlation,” Journal of Experimental
Psychology: Learning, Memory, and Cognition 22(5): 1304-1316 (September 1996). H. L. Roediger III, J. T.
Wixted, and K.A. DeSoto. “The Curious Complexity between Confidence and Accuracy in Reports from Memory”
in Memory and Law, ed. L. Nadel and W. P. Sinnott-Armstrong (Oxford: Oxford University Press, 2012). L.
Mickes, H. D. Flowe, and J. T. Wixted, “Receiver operating characteristic analysis of eyewitness memory:
Comparing the diagnostic accuracy of simultaneous and sequential lineups,” Journal of Experimental Psychology:
Applied 18 (4): 361–376 (2012).

60
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
on this and other possible methods of estimating or controlling response bias is warranted to support
efforts to extract a bias-free measure of discriminability.
Another technical concern raised by the use of ROC analysis to evaluate eyewitness identification
performance is that it relies on a partial, rather than full, area under the ROC curve measure (see Box 5.2)
as an index of discriminability that is separate from response bias. This is necessitated by the fact that the
highest false alarm rates in eyewitness identification data are commonly well below 1.0, even under the
most liberal response bias conditions.51 In practice, partial area under the curve is computed by truncating
the ROC curve at the highest false alarm rate obtained. Because the standard error of the partial area
under the curve measure depends upon the degree of truncation, accuracy of this discriminability measure
can easily vary across conditions and across studies, making the interpretation difficult.52
While ROC analysis has many recognized merits for the evaluation of binary classification, the
residual concerns associated with its typical use for evaluating eyewitness performance merit
consideration of other statistical approaches to this problem. As noted above, many methods have been
proposed – and adopted in specific applications – for evaluation of binary classification performance.53
The committee knows of no instance in which any of these alternative methods has been applied to the
problem of eyewitness identification. Moreover, because they have not been vetted, the committee is not
in a position to endorse any specific statistical tool, but the committee encourages a general exploration of
these alternatives. These alternatives may have their own share of unforeseen problems, and/or the
performance criteria employed by them may bear no meaningful relationship to the sensory and cognitive
processes involved in eyewitness identification. Nonetheless, some of these methods may provide greater
insight into the factors that affect eyewitness identification performance and may, in turn, suggest ways of
improving performance. To illustrate this opportunity by example, we consider the following
possibilities:
It has been argued that a basic weakness of the existing ROC approach to binary classification
performance results from the fact that, in principle (and in practice under certain commonly unrecognized
conditions), the area under the ROC curve is dependent on imprecise assumptions about the costs of
classification errors across different classification conditions.54 One might suppose, for example, that the
cost of a miss for a crime of murder is greater than the cost of a miss for a stolen car. But without a
51

C. A. Carlson and M. A. Carlson, “An evaluation of lineup presentation, weapon presence, and a distinctive
feature using ROC analysis,” Journal of Applied Research in Memory and Cognition 3(2): 45–53 (2014). L. Mickes,
H. D. Flowe, and J. T. Wixted, “Receiver operating characteristic analysis of eyewitness memory: Comparing the
diagnostic accuracy of simultaneous and sequential lineups,” Journal of Experimental Psychology: Applied 18 (4):
361–376 (2012).
52
Along the same lines, accuracy of discriminability measures derived from ROC studies may be called into
question when those studies do not take into account uncertainty in the data used to construct the ROC curves; see
Appendix C. An argument has also been made that the area under the ROC curve can be a flawed metric for
comparing binary classification conditions when the costs of classification errors are not precisely known and are
different for different conditions (D. J. Hand, “Measuring classifier performance: A coherent alternative to the area
under the ROC curve,” Machine Learning 77, 103–123 [2009]). The costs of classification errors may be similar
across some lineup comparisons and across some conditions of other systems variables, and for others they may be
different. But for the most part they not precisely known and this is thus a topic that deserves greater attention given
the growing use of ROC-based evaluation of eyewitness identification performance.
53
Numerous methods for the evaluation of binary classifiers have been developed and applied in the field of
machine learning, which seeks to optimize autonomous classification devices (such as, for example, the fingerprint
lock access control on your smart phone, which must quickly and reliably distinguish your finger from that of
another). This field has a long and rich history and candidate methods are summarized in several texts on statistical
classification and machine learning, such as Hastie, Tibshirani, and Friedman, The Elements of Statistical Learning,
Springer, 2nd ed; A. Smola and S.V.N. Vishwanathan, Introduction to Machine Learning (Cambridge: Cambridge
University Press, 2009).
54
See D. J. Hand, “Measuring classifier performance: A coherent alternative to the area under the ROC curve,”
Machine Learning 77, 103–123 (2009).

61
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
precise understanding of these relative decision costs, the area under the ROC curve measure can be
incoherent, in that it depends as much on the classification conditions as it does on the sensitivity of the
classifier. An alternative method has been proposed to address this problem—derivation of the “H
measure”55—that enables the performance of binary classifiers to be compared using a common metric
that is independent of the cost distributions for different types of classification errors. The committee
supports exploration of this alternative.
Another avenue for exploration emerges from the fact that the literature evaluating eyewitness
identification performance has focused exclusively on the positive predictive value (PPV) of a witness’
classification as guilty. For a given response bias, PPV is related to the diagnosticity ratio, in that, given
equal prevalence of the culprit in two conditions (e.g. lineup procedures) being compared, a higher
diagnosticity ratio leads to a higher PPV. As discussed above, the diagnosticity ratio is a critical piece of
information in efforts to evaluate eyewitness performance. As for any binary classification, however,
there is also information associated with a negative response, which is the predictive value of a
classifier’s assertion that a target is not present (in the eyewitness case, the witness’ assertion of
innocence). This negative predictive value (NPV) is related to a different ratio of decisions, namely
(1-HR)/(1-FAR)56, in that, given equal prevalence of the target in the two procedures being compared,
higher values of this ratio correspond to higher values of NPV.
While NPV is commonly used to evaluate the accuracy of human classification decisions, such as in
medical diagnosis, and is a source of information that may similarly be of additional value in efforts to
evaluate lineup procedures, it has been largely neglected in the field of eyewitness identification.57 One
might hold the intuition that PPV and NPV are monotonically related to one another – believing that the
likelihood that the witness will correctly identify the culprit is proportional to the likelihood that the
witness will correctly identify lineup candidates as innocent – and thus conclude that evaluation of PPV
alone is sufficient. Contrary to that intuition, however, evidence from studies of analogous binary
classification problems reveals that these two predictive probabilities can vary with respect to one another
in complex ways.58
In practice, NPV-related measures (quantified as negative likelihood ratios) can be subjected to
ROC analysis, to account for the effects of response bias, in the same manner as PPV-related measures
(quantified as positive likelihood ratios, i.e., diagnosticity ratios) – the ROC axes in the NPV case
corresponding to 1-HR and 1-FAR. Consideration of NPV and its relationship to PPV, by this and other
means,59 may provide additional insight into the ways in which estimator and system variables, such as
lineup procedures, influence eyewitness identification performance.
In sum, a formal understanding of the task facing an eyewitness, in conjunction with an appreciation
of causal models of human recognition memory, has led to a potentially more comprehensive method—
55

Ibid.
The reciprocal of this ratio is called the “negative likelihood ratio.” See, e.g., T. Hoffmann, S. Bennett, and
C. del Mar, Evidence-Based Practice Across the Health Professionals (Chatswood: Elsevier Australia, 2009).
57
It seems likely that this neglect stems from the fact that the primary concern in eyewitness identification has
been on incorrect assertions of guilt (i.e., false identifications) rather than incorrect assertions of innocence. There
are normative values in society that reinforce this concern – by William Blackstone’s maxim, “Better that ten guilty
persons escape than that one innocent suffer.”
58
S-Y Shiu and C. Gatsonis, “The predictive receiver operating characteristic curve for the joint assessment of
the positive and negative predictive values,” Philosophical Transactions, Series A, Mathematical, Physical and
Engineering Sciences 366 (1874): 2313–2333 (2008).
59
Another potentially informative analysis that combines PPV and NPV measures is known as a PROC
(predictive ROC), which affords the opportunity to see how a given system or estimator variable may have
interacting – synergistic or antagonistic – effects on assertions of guilt and innocence. See S-Y Shiu and C.
Gatsonis, “The predictive receiver operating characteristic curve for the joint assessment of the positive and
negative predictive values,” Philosophical Transactions, Series A, Mathematical, Physical and Engineering
Sciences (1874): 2313–2333 (2008).
56

62
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
ROC analysis—for evaluating eyewitness identification performance. Despite these advances, it is
important that practitioners in this field broadly explore the large and rich field of statistical tools for
evaluation of binary classifiers. While the committee recognizes that these tools are uninvestigated for
this application and may possess their own share of unforeseen problems or disadvantages, a move in this
direction may be of great value for improving the validity of eyewitness identification.
Interactions with Eyewitnesses (Feedback)
The nature of law enforcement interactions with the eyewitness before, during, and after the
identification plays a role in the accuracy of eyewitness identifications and in the confidence expressed in
the accuracy of those identifications by witnesses.60 Law enforcement’s maintenance of neutral preidentification communications—relative to the identification of a suspect—is seen as vital to ensuring that
the eyewitness is not subjected to conscious or unconscious verbal or behavioral cues that could influence
the eyewitness’ identification61 (see Box 2-1). If a witness happened to overhear an officer say, “We’ve
got him, but before we finalize the arrest, let’s have the witness confirm it,” the witness might be biased
to confirm the suspect’s identity in a showup. Furthermore, some types of law enforcement
communication with a witness after the witness has made an identification (e.g., “Good work! You picked
the right guy…”) can increase confidence in an identification, regardless of whether the identification is
correct.62
As discussed in Chapter 2, use of “blinded” or “doubl-blind” lineup identification procedures is an
effective strategy for reducing the likelihood that a witness will be exposed to cues from interactions with
law enforcement (such as feedback) that could influence identifications and/or confidence in those
identifications. More generally, efforts to maintain objectivity and eliminate potentially informative
communication will help ensure that eyewitness reports are not contaminated by knowledge or opinions
held by others.
RESEARCH STUDIES ON ESTIMATOR VARIABLES
The impact of estimator variables on eyewitness accuracy is harder to measure in the field than the
impact of system variables.63 Consequently, estimator variables have been studied nearly exclusively in
laboratory settings. The committee’s review revealed the need for further empirical research in individual
studies and systematic reviews of research on these factors.
The committee’s review focused on the most-studied estimator variables: weapon focus, stress and
fear, own-race bias, exposure, and retention interval. It is important to emphasize, however, that
numerous other estimator variables may affect both the reliability and the accuracy of eyewitness
identifications. Research has shown that the physical distance between the witness and the perpetrator is
an important estimator variable, as it directly affects the ability of the eyewitness to discern visual

60

S. E. Clark, T. E. Marshall, and R. Rosenthal,“Lineup Administrator Influences on Eyewitness Identification
Decisions,” Journal of Experimental Psychology: Applied 15(1): 63 (2009).
61
S.E. Clark, M.B. Moreland, and S.D. Gronlund, “Evolution of the empirical and theoretical foundations of
eyewitness identification reform,” Psychonomic Bulletin and Review 21(2): 251–267 (2014). “…the performance
advantage for unbiased instructions has decreased only slightly over the past 32 years. However, none of the
correlations approached statistical significance.” p. 258.
62
A. B. Douglas and N. K. Steblay, “Memory Distortion in Eyewitnesses: A Meta-analysis of the Postidentification Feedback Effect,” Applied Cognitive Psychology 20: 859–869 (2006).
63
G. L. Wells, “What Do We Know About Eyewitness Identification?” American Psychologist (May 1993):
553, 555.

63
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
details,64 including features of the perpetrator65 (see discussion of vision in Chapter 4). Research has also
shown that an appearance change can greatly diminish the ability of the eyewitness’ ability to recognize
the perpetrator; the eyewitness’ ability to recognize faces of one’s own age group is often superior to
memory for faces of another age group (own-age bias); and if an eyewitness hears information or
misinformation from another person before law enforcement involvement, his or her recollection of the
event and confidence in the identification can be altered (co-witness contamination).66 Interactions
between and among these variables have not been addressed systematically by researchers.
Weapon Focus
The presence of an unusual object at the scene of a crime can impair visual perception and memory
of key features of the crime event. Research suggests that the presence of a weapon at the scene of a
crime captures the visual attention of the witness and impedes the ability of the witness to attend to other
important features of the visual scene, such as the face of the perpetrator (see also discussion of visual
attention in Chapter 4). The ensuing lack of memory of these other key features may impair recognition of
a perpetrator in a subsequent lineup.
A 1992 analysis of weapon focus studies found that the presence of a weapon reduced both
identification accuracy and feature accuracy (e.g., the eyewitness’ ability to recall clothing, facial
features, and more).67 A more recent analysis of the weapon focus literature concluded that the presence
of a weapon has an inconsistent effect on identification accuracy, in that larger effect sizes were observed
in threatening scenarios than in non-threatening ones.68 As the retention interval increased, the weapon
focus effect size decreased. The analysis further shows that the effect of a weapon on accuracy is slight in
actual crimes, slightly larger in laboratory studies, and largest for simulations.
One possible cause of the inconsistent effects of the presence of a weapon is suggested by a recent
laboratory-based study69 that exposed participants to crime videos. These investigators used ROC
analysis to investigate discriminability as a function of (1) sequential versus simultaneous lineups; (2) the
presence of a weapon; and (3) the presence of a distinctive facial feature. The findings regarding lineup
procedure (sequential versus simultaneous) are reported above. Importantly for the present discussion,
discriminability was reduced when the perpetrator possessed a weapon, but only when no distinctive
facial feature was present. This interaction between weapon focus and distinctive feature highlights the
importance of exploring the effects of interactions between different estimator variables on eyewitness
identification performance.
Additional questions remain as to what is the cause of reduced eyewitness performance in cases
where a weapon is present. Is the effect caused by a diversion of selective attention, as is suggested by
basic research on the phenomenon of inattentional blindness (see Chapter 4)? Is stress a significant
factor, i.e., does anxiety cause the witness to focus less on the features of a person’s face? To what extent
64

B. Uttl, P. Graf, and A.L. Siegenthaler, “Influence of Object Size on Baseline Identification, Priming, and
Explicit Memory: Cognition and Neurosciences,” Scandinavian Journal of Psychology 48(4): 281–288 (2007).
65
C. L. Maclean, C. A. E. Brimacombe, M. Allison, L. C. Dahl, and H. Kadlec, “Post-Identification Feedback
Effects: Investigators and Evaluators,” Applied Cognitive Psychology 25(5): 739–752 (2011).
66
R. Zajac and N. Henderson, “Don’t It Make My Brown Eyes Blue: Co-witness Misinformation About a
Target’s Appearance Can Impair Target-absent Lineup Performance,” Memory 17(3): 266–278 (2009).
67
N. K. Steblay, “A Meta-analytic Review of the Weapon Focus Effect,” Law and Human Behavior 16 (4): 413,
415–17 (1992).
68
J. M. Fawcett, E. J. Russell, K. A. Peace, and J. Christie, “Of Guns and Geese: A Meta-analytic Review of the
‘Weapon Focus’ Literature,” Psychology, Crime & Law 19(1): 35–66 (2013).
69 C. Carlson and M. Carlson, “An Evaluation of Lineup Presentation, Weapon Presence, and a Distinctive
Feature using ROC Analysis,” Journal of Applied Research in Memory and Cognition 3(2): 45–53 (2014).

64
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
is the prominence of the issue an artifact of the particular studies included in the meta-analysis? Is it
possible, for example, that the magnitude of the weapon effect depends upon the whether the data are
collected in a laboratory setting versus the real world. To this latter point, some analyses of weapon focus
have been conducted using archival records of actual crimes involving weapons. Unfortunately, such
efforts often encounter serious methodological difficulties, which include a lack of information about the
crime (e.g., exposure duration) and the general lack of “ground truth” regarding accuracy of any
identification.
Some analyses of weapon focus have been conducted using archival records of crimes involving
weapons.70 Such efforts often encounter serious methodological difficulties that include a lack of
information about the crime (e.g., exposure duration) and the general lack of “ground truth” regarding
accuracy of any identification, among other problems.
Stress and Fear
High levels of stress or fear can affect eyewitness identification.71,72,73 This finding is not surprising,
given the known effects of fear and stress on vision and memory (see Chapter 4). Under conditions of
high stress, a witness’ ability to identify key characteristics of an individual’s face (e.g., hair length, hair
color, eye color, shape of face, presence of facial hair) may be significantly impaired.74
In the particular case of weapon focus, it may not be possible to sufficiently test the effects of stress
and heightened stress in the laboratory because of limitations on human participant research that uses
realistic and heightened threats. A meta-analysis of the effect of high stress on eyewitness memory
nonetheless found some support for the notion that stress impairs both eyewitness recall and identification
accuracy.75 The study authors noted that lineup type “moderated the effect of heightened stress on the
false alarm rate.”76 They also suggested that the modest effect of stress may be caused by the fact that the
analysis included many studies that involved modest stress-induction.77
Earlier studies were more mixed but with clearer results at “high levels of cognitive anxiety.”78 The
findings of an earlier study79 “provide a concrete illustration of catastrophic decline” of eyewitness
70

See, e.g., J. M. Fawcett, E. J. Russell, K. A. Peace, and J. Christie, “Of Guns and Geese: A meta-analytic
Review of the ‘Weapon Focus’ Literature,” Psychology Crime & Law 19(1): 43–44 (2013).
71
K. A. Deffenbacher, B. H. Bornstein, S. D. Penrod, and E. K. McGorty, “A Meta-analytic Review of the
Effects of High Stress on Eyewitness Memory,” Law and Human Behavior 28(6): 687–706 (2004).
72
C. A. Morgan III et al., “Accuracy of Eyewitness Memory for Persons Encountered During Exposure to
Highly Intense Stress,” International Journal of Law and Psychiatry 27(3): 265–79 (2004).
73
C. A. Morgan III, G. Hazlett, M. Baranoski, A. Doran, S. M. Southwick, A. Loftus, “Accuracy of Eyewitness
Identification is Significantly Associated with Performance on a Standardized Test of Recognition,” International
Journal of Law and Psychiatry 30 (3): 213–223 (2007).
74
C. A. Morgan III, S. M. Southwick, G. Steffian, G. A. Hazlett, E. F. Loftus, “Misinformation can influence
memory for recently experienced, highly stressful events,” International Journal of Law and Psychiatry 36(1): 11–
17 (2013).
75
K. A. Deffenbacher, B. H. Bornstein, S. D. Penrod, and E. K. McGorty, “A meta-analytic Review of the
Effects of High Stress on Eyewitness Memory,” Law and Human Behavior 28(6): 687–706 (2004). It should be
noted that the effect sizes for stress-induced support were small with wide confidence intervals, indicating
considerable heterogeneity across studies. Although the authors assert that 300 studies with null findings would be
required to negate the small effects found in this meta-analysis, fewer studies might be needed if they resulted in
opposite effects.
76
Ibid, 700.
77
Ibid, 704.
78
Ibid, 689.
79
T. Valentine and J. Mesout, “Eyewitness Identification Under Stress in the London Dungeon,” Applied
Cognitive Psychology 23(2): 151–161 (2009).

65
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
identification performance at high anxiety levels. The correct identification rate went from 75% for those
with low-state anxiety to 18% rate for those with high-state anxiety.80
The effects of suggestion and estimator variables may be particularly important when the original
memory is of a highly stressful event. A recent study looked at more than 850 active-duty military
personnel participating in a mock POW camp phase of U.S. military survival school training, which
included aggressive interrogation and physical isolation-related stress.81 The study found that
misinformative details of the interrogation event (e.g., regarding the identity of the interrogator), which
were introduced after the event had been encoded into long-term memory, affected identification
accuracy. The study also found that memories acquired during stressful events are highly vulnerable to
modification by exposure to post-event misinformation, even in individuals whose level of training and
experience might be considered relatively immune to such influences.
Another recent study comparing the eyewitness accuracy of officers and citizens, concentrated on
the effects of stress and weapon focus.82 The results of this study showed that officers were less stressed
and aroused than citizens, but that both police and citizens made more errors when a weapon was inferred
or present.
Own-Race Bias
The race and ethnicity of a witness as it relates to that of the perpetrator is another important
estimator variable. In eyewitness identification, own-race bias describes the phenomenon in which faces
of people of races different from that of the eyewitness are harder to discriminate (and thus harder to
identify accurately) than are faces of people of the same race as the eyewitness.83 In the laboratory, this
effect is manifested by higher hit rates and lower false alarm rates (higher diagnosticity ratio) in the
recognition of an observer’s own race relative to hits and false-alarms for recognition of other-races.84
Own-race bias occurs in both visual discrimination and memory tasks, in laboratory and field studies, and
across a range of races, ethnicities, and ages. Recent analyses revealed that cross-racial (mis)identification
was present in 42 percent of the cases in which an erroneous eyewitness identification was made.85
A recent meta-analysis of own-race bias found an interaction between own-race bias and the
duration of viewing exposure: reducing the amount of time allowed for viewing of each face significantly
increased the magnitude of the bias, largely manifested as an increase in the proportion of false alarm
responses to other-race faces.86 Own-race bias also interacts with the memory retention interval; crossrace errors of identification were greater when there were longer periods of time between the initial
exposure and the memory retrieval.87 A recent study found that “context reinstatement,” wherein a
80

K. A. Deffenbacher, “Estimating the Impact of Estimator Variables on Eyewitness Identification: A Fruitful
Marriage of Practical Problem Solving and Psychological Theorizing,” Applied Cognitive Psychology 22(6): 822
(2008).
81
C. A. Morgan, S. Southwick, G. Steffian, G. A. Hazlett, and E. F. Loftus, “Misinformation Can Influence
Memory for Recently Experienced, Highly Stressful Events,” International Journal of Law and Psychiatry 36(1):
11–17 (2013).
82
J. C. DeCarlo, “A Study Comparing the Eyewitness Accuracy of Police Officers and Citizens,” (PhD Diss,
City University of New York, 2010).
83
R. S. Malpass and J. Kravitz, “Recognition for Faces of Own and Other Race,” Journal of Personality and
Social Psychology 13(4): 330–334 (1969).
84
C. A. Meissner and J. C. Brigham, “Thirty Years of Investigating the Own-race Bias in Memory for Faces: A
Meta-analytic Review,” Psychology Public Policy and Law 7(1): 3–35 (2001).
85
The Innocence Project, “What Wrongful Convictions Teach Us About Racial Inequality,” available at:
http://www.innocenceproject.org/Content/What_Wrongful_Convictions_Teach_Us_About_Racial_Inequality.php.
86
C. A. Meissner and J. C. Brigham, “Thirty Years of Investigating the Own-Race Bias in Memory for Faces: A
Meta-analytic Review,” Psychology Public Policy and Law 7(1): 3–35 (2001).
87
Ibid.

66
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
researcher asks an individual to mentally re-create the context in which an incident occurred, failed to
influence the identification of other-race faces.88
Although the existence of own-race bias is generally accepted, the causes for this effect are not fully
understood. Some possible explanations are rooted in in-group/out-group models of human behavior (e.g.,
favoritism in which decisions regarding members of one’s own “group” are regarded as having greater
importance than decisions regarding members of a different “group”), and differential perceptual
expertise that results from different degrees of exposure to and familiarity with same versus other races.
Recent work has examined the role that stereotyping might play.89 One study suggests that, in
general, cross-race identification is further impaired when faces are presented in a group (as opposed to
one at a time).90 Additional research is needed to identify procedures that may help estimate the degree of
own-race biases in individual eyewitnesses following an identification procedure.Until the scientific basis
for these effects is better understood, great care may be warranted when constructing lineups in instances
where the race of the suspect differs from that of the eyewitness.
Exposure Duration
Eyewitness identification researchers have long believed that exposure duration (e.g., time spent
observing a perpetrator’s face during a crime) is correlated with greater accuracy of eyewitness
identification. The courts also have assumed that exposure duration has an effect on identification
accuracy.91 Meta-analyses on the effects of exposure time have found that relatively long exposure
durations produce greater accuracy92 and a larger and more stable effect size for exposure duration on
eyewitness identification accuracy.93 Longer exposures were associated with higher rates of correct
identifications and lower false alarm rates. Exposure duration may affect, or interact with, other variables,
including own-race bias and the confidence–accuracy relationship assessed immediately after the lineup
decision.94
The findings and conclusions from eyewitness identification studies of exposure duration are in
keeping with much of the basic research on visual system function (reviewed in Chapter 4). This basic
88

J. R. Evans, J. L. Marcon, and C.A. Meissner, “Cross-racial Lineup Identification: Assessing the Potential
Benefits of Context Reinstatement,” Psychology, Crime, and Law 15 (1): 19–28 (2009).
89
H. M. Kleider, S. E. Cavrak, and L. R. Knuycky, “Looking Like a Criminal: Stereotypical Black Facial
Features Promote Face Source Memory Error,” Memory & Cognition 40(8): 1200–1213 (2012).
90
K. Pezdek, M. O’Brien, and C. Wasson, “Cross-race (but Not Same-race) Face Identification Is Impaired by
Presenting Faces in a Group Rather Than Individually,” Law and Human Behavior 36(6): 488–495 (2012).
91
Manson v. Brathwaite, 432 U.S. 98, 114 (1977), for example, included as a factor for assessing the reliability
and admissibility of an identification, “the opportunity of the witness to view the criminal at the time of the crime”
and explained that this factor includes both the length of time and the viewing conditions.
92
B. H. Bornstein et al., “Effects of Exposure Time and Cognitive Operations on Facial Identification Accuracy:
A Meta-analysis of Two variables Associated with Initial Memory Strength,” Psychology, Crime, and Law 18 (5):
473–490 (2012). The authors state, “We used z as the primary effect size measure for differences between
proportions correct, but we also converted z to Pearson’s r for comparability to other meta-analyses (see Tables 1
and 2). The rs were then normalized and averaged to obtain the overall mean effect sizes. We also report the value
of Cohen’s d associated with each mean effect size” [B. H. Bornstein, K. A. Deffenbacher, S. D. Penrod, and E. K.
McGorty, “Effects of Exposure Time and Cognitive Operations on Facial Identification Accuracy: A Meta-analysis
of Two variables Associated with Initial Memory Strength,” Psychology, Crime, and Law 18 (5):479 (2012)].
Although not defined, presumably z refers to the usual difference in means divided by its standard error, and, from
their tables, their r was calculated as z divided by the square root of the report sample size.
93
B. H. Bornstein, K. A. Deffenbacher, E. K. McGorty, and S. D. Penrod, The Effect of Cognitive Processing on
Facial Identification Accuracy: A Meta-analysis (Unpublished manuscript, University of Nebraska-Lincoln, 2007).
94
M. A. Palmer, N. Brewer, N. Weber, and A. Nagesh, “The Confidence–Accuracy Relationship for Eyewitness
Identification Decisions: Effects of Exposure Duration, Retention Interval, and Divided Attention,” Journal of
Experimental Psychology: Applied 19(1): 55–71 (2013).

67
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
research indicates that the additional information available from longer viewing times reduces uncertainty
and enables better detection and discrimination of visual stimuli.
Retention Interval
Retention interval, or the amount of time that passes from the initial observation and encoding of a
memory to a future time when the initial observation must be recalled from memory, can affect
identification accuracy. Laboratory studies have demonstrated that stored memories are more likely to be
forgotten with the increasing passage of time and can easily become “enhanced” or distorted by events that
take place during this retention interval (see discussion of memory in Chapter 4). The amount of time
between viewing a crime and the subsequent identification procedure can be expected to similarly affect the
accuracy of the eyewitness identification, either independently or in combination with other variables.95
It is difficult to specify the precise relationship between retention interval and the accuracy of
eyewitness identification testimony, and to estimate when a lengthy retention interval will significantly
impair the accuracy of identification. Although, in general, it appears that longer retention intervals are
associated with poorer eyewitness identification performance, the strength of this association appears to
vary greatly across the circumstances of the initial encounter, identification procedures, and research
methodologies.96 A meta-analysis of published facial recognition and eyewitness identification studies
found, for example, that an increase in the retention interval was associated with a decreased probability
of an accurate identification of a previously seen but otherwise unfamiliar face.97 This same study also
found that the rate of forgetting for an unfamiliar face is greatest soon after the initial observation and
tends to level off over time, but was unable to specify the shape of this function.
The effect of the retention interval also is influenced by the strength and quality of the initial
memory that is encoded, which, in turn, may be influenced by other estimator variables associated
with witnessing the crime (such as the degree of visual attention) and viewing factors (such as distance,
lighting, and exposure duration). As the retention interval becomes longer, the opportunity for intervening
events to alter the memory also becomes greater, and other variables may interact with the retention
interval to impair performance (see also discussion of memory in Chapter 4). During the retention
interval, the ability to accurately identify faces of other races drops off especially quickly, relative to
same-race accuracy.98 Also, for those eyewitnesses who initially express less confidence in their
95

One month is the most commonly encountered delay by British police. G. Pike, N. Brace, and S. Kynan, The
visual identification of suspects: Procedures and practice, (London: Policing and Reducing Crime Unit, 2002), cited
by K. A. Deffenbacher, B. H. Bornstein, E. K. McGorty, and S. D. Penrod, “Forgetting the Once-Seen Face:
Estimating the Strength of an Eyewitness's Memory Representation,” Journal of Experimental Psychology: Applied
14(2): 139–150, 142 (2008). Law enforcement authorities may have little control over the time required to identify a
suspect and obtain the cooperation of the eyewitness to participate in an identification procedure. Thus, retention
interval has commonly been considered an estimator variable in eyewitness identification studies.
96
See J. Dysart and R. C. L. Lindsay, “The Effects of Delay on Eyewitness Identification Accuracy: Should we
be concerned?” in The Handbook of Eyewitness Psychology: Volume II: Memory for People, ed. R. C. L. Lindsay,
D. F. Ross, J. D. Read, and M. P. Toglia. (Mahwah: Lawrence Erlbaum and Associates, 2006), 361–73.
97
K. A. Deffenbacher, B. H. Bornstein, E. K. McGorty, and S. D. Penrod, “Forgetting the Once-Seen Face:
Estimating the Strength of an Eyewitness’s Memory Representation,” Journal of Experimental Psychology: Applied
14(2):139–50, 142 (2008). More than twenty of the published studies included in the meta-analysis found no
significant effect of retention interval.
98
J. L. Marcon, C. A. Meissner , M. Frueh , K. J. Susa, and O. H. MacLin, “Perceptual identification and the
cross-race effect,” Visual Cognition 18:5: 767–779 (2010) (finding that the cross-race effect was more pronounced
when the retention interval was lengthened). C. A. Meissner and J. C. Brigham, “Thirty years of investigating the
own-race bias in memory for faces—A meta-analytic review,” Psychology, Public Policy, and Law 7(1): 3–35, 24
(2001) (meta-analysis finding that as retention time increased “participants increasingly adopted a more liberal
response criterion when responding to other-race faces. This liberal response criterion indicated that participants

68
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
identification, there is a greater decrease in accuracy of identification when the retention interval is
longer.99
CONCLUSION
Research on eyewitness identification has appropriately identified the variables that may affect an
individual’s ability to make an accurate identification. Early research findings played an important role in
alerting law enforcement, prosecutors, defense counsel, and the judiciary to factors that might influence
the accuracy of identifications. In some jurisdictions, eyewitness identification research was used to
improve policies and procedures and to educate and train officers. However, much remains unsettled in
many areas of eyewitness identification research.
While past research appropriately identified system and estimator variables that may affect an
individual’s ability to make an accurate identification, this research might be strengthened in several
ways. Greater collaboration between the police, the courts, and researchers might lead to increased
consensus on research agendas and the conceptualization of variables to be examined. More attention to
reproducibility and transparency is needed in the selection of data collection strategies and reporting of
data. Analyses need to be reported completely, including estimates of effects, confidence intervals, and
significance levels. Further, in order to be useful to stakeholders, the statistical findings of this research
need to be translated back into terms that can be readily understood by practice and policy decisionmakers.
Further, our understanding of errors in eyewitness identification will benefit from more effective
research designs, more informative statistical measures and analyses, more probing analyses of research
findings, and more sophisticated systematic reviews and meta-analyses. In view of the complexity of the
effects of both system and estimator variables, and their interactions, on eyewitness identification
accuracy, better experimental designs that incorporate selected combinations of these variables (e.g.,
presence or absence of a weapon, lighting conditions, etc.) will elucidate those variables with meaningful
influence on eyewitness performance, which can inform law enforcement practice of eyewitness
identification procedures. To date, the eyewitness literature has evaluated procedures mostly in terms of a
single diagnosticity ratio or an ROC curve; even if uncertainty is incorporated into the analysis, many
other powerful tools for evaluating a “binary classifier” are worthy of consideration.100
When primary studies such as those described above are available in sufficient quantities, it is
important that their results are synthesized using systematic reviews which conform to current best
standards,101,102 These quantitative reviews would necessarily employ transparent, reproducible
procedures for locating all relevant published and unpublished research; employ independent, duplicate
procedures for selection of studies, extraction of data, and assessment of risk of bias; use meta-analytic
procedures which account for the heterogeneity of outcomes both within and across studies; and interpret
confidence intervals around pooled effects in a way that is readily understandable by stakeholders. These
required less evidence from memory [e.g., familiarity or memorability of the face] to respond that they had
previously seen an other-race face.”).
99
J. Sauer, N. Brewer, T. Zweck, and N. Weber, “The Effect of Retention Interval on the Confidence–Accuracy
Relationship for Eyewitness Identification,” Law and Human Behavior 34: 337–347 (2010) (finding greater
overconfidence at lengthy retention intervals).
100
T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning: Data Mining, Inference, and
Prediction, 2d ed. (New York: Springer, 2009).
101
A. Liberati, D. G. Altman, J. Tetzlaff, C. Mulrow, P. C. Gøtzsche, et al., “The PRISMA Statement for
Reporting Systematic Reviews and Meta-Analyses of Studies That Evaluate Health Care Interventions: Explanation
and Elaboration,” PLoS Medicine 6(7): e1000100. doi:10.1371/journal.pmed.1000100 (2009).
102
Institute of Medicine, Finding what works in health care: Standards for systematic reviews (Washington,
D.C.: The National Academies Press, 2011).

69
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
systematic reviews (which would be regularly updated as new studies are conducted) can be used to
further refine the research agenda in eyewitness identification research and to establish priorities for
funding of additional primary research.

70
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

6
Findings and Recommendations
Eyewitnesses make mistakes. Our understanding of how to improve the accuracy of eyewitness
identifications is imperfect and evolving. In the previous chapters, we described law enforcement
procedures to elicit accurate eyewitness identifications; the courts’ handling of eyewitness identification
evidence; the science of visual perception and memory as it applies to eyewitness identifications; and the
contributions of scientific research to our understanding of the variables that affect the accuracy of
identifications. On the basis of its review, the committee offers its findings and recommendations for
• identifying and facilitating best practices in eyewitness procedures for the law enforcement
community;
• strengthening the value of eyewitness identification evidence in court; and
• improving the scientific foundation underpinning eyewitness identification.
OVERARCHING FINDINGS
The committee is confident that the law enforcement community, while operating under
considerable pressure and resource constraints, is working to improve the accuracy of eyewitness
identifications. These efforts, however, have not been uniform and often fall short as a result of
insufficient training, the absence of standard operating procedures, and the continuing presence of actions
and statements at the crime scene and elsewhere that may intentionally or unintentionally influence
eyewitness’ identifications.
Basic scientific research on human visual perception and memory has provided an increasingly
sophisticated understanding of how these systems work and how they place principled limits on the
accuracy of eyewitness identification (see Chapter 4).1 Basic research alone is insufficient for
understanding conditions in the field, and thus has been augmented by studies applied to the specific
practical problem of eyewitness identification (see Chapter 5). Applied research has identified key
variables that affect the accuracy and reliability of eyewitness identifications and has been instrumental in
informing law enforcement, the bar, and the judiciary of the frailties of eyewitness identification
testimony.
A range of best practices has been validated by scientific methods and research and represents a
starting place for efforts to improve eyewitness identification procedures. A number of law enforcement
agencies have, in fact, adopted research-based best practices. This report makes actionable
recommendations on, for example, the importance of adopting “blinded” eyewitness identification
procedures. It further recommends that standardized and easily understood instructions be provided to
eyewitnesses and calls for the careful documentation of eyewitness’ confidence statements. Such
improvements may be broadly implemented by law enforcement now. It is important to recognize,
however, that, in certain cases, the state of scientific research on eyewitness identification is unsettled.

1

Basic research on vision and memory seeks a comprehensive understanding of how these systems are
organized and how they operate generally. The understanding derived from basic research includes principles that
enable one to predict how a system (such as vision or memory) might behave under specific conditions (such as
those associated with witnessing a crime), and to identify the conditions under which it will operate most effectively
and those under which it will fail. Applied research, by contrast, empirically evaluates specific hypotheses about
how a system will behave under a particular set of real-world conditions.

71
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
For example, the relative superiority of competing identification procedures (i.e., simultaneous versus
sequential lineups) is unresolved.
The field would benefit from collaborative research among scientists and law enforcement personnel
in the identification and validation of new best practices that can improve eyewitness identification
procedures. Such a foundation can be solidified through the use of more effective research designs (for
example, those that consider more than one variable at a time, and in different study populations to ensure
reproducibility and generalizability), more informative statistical measures and analyses (i.e., methods
from statistical machine learning and signal detection theory to evaluate the performance of binary
classification tasks), more probing analyses of research findings (such as analyses of consequences of
data uncertainties), and more sophisticated systematic reviews and meta-analyses (that take account of
current guidelines, including transparency and reproducibility of methods).
In view of the complexity of the effects of both system and estimator variables and their interactions
on eyewitness identification accuracy, better experimental designs that incorporate selected combinations
of these variables (e.g., presence or absence of a weapon, lighting conditions, etc.) will elucidate those
variables with meaningful influence on eyewitness performance, which can, in turn, inform law
enforcement practice of eyewitness identification procedures. To date, the eyewitness literature has
evaluated procedures mostly in terms of a single diagnosticity ratio or an ROC (Receiver Operating
Characteristic) curve; even if uncertainty is incorporated into the analysis, many other powerful tools for
evaluating a “binary classifier” are available and worthy of consideration.2 Finally, syntheses of
eyewitness research has been limited to meta-analyses that have not been conducted in the context of
systematic reviews. Systematic reviews of stronger research studies need to conform to current standards
and be translated into terms that are useful for decision makers.
The committee offers the following recommendations to strengthen the effectiveness of policies and
procedures used to obtain accurate eyewitness identifications.
RECOMMENDATIONS TO ESTABLISH BEST PRACTICES
FOR THE LAW ENFORCEMENT COMMUNITY
The committee’s review of law enforcement practices and procedures, coupled with its
consideration of the scientific literature, has identified a number of areas where eyewitness identification
procedures could be strengthened. The practices and procedures considered here involve acquisition of
data that reflect a witness’ identification and the contextual factors that bear on that identification. A
recurrent theme underlying the committee’s recommendations is development of, and adherence to,
guidelines that are consistent with scientific standards for data collection and reporting.
Recommendation #1: Train All Law Enforcement Officers in Eyewitness Identification
The resolution and accuracy of visual perceptual experience, as well as the fidelity of our memories
to events perceived, may be compromised by many factors at all stages of processing (see Chapter 4).
Unknown to the individual, perceptual experiences are limited by uncertainties and biased by
expectations. Memories are forgotten, reconstructed, updated, and distorted. An eyewitness’s memory
can be contaminated by a wide variety of influences, including interaction with the police.
The committee recommends that all law enforcement agencies provide their officers and agents
with training on vision and memory and the variables that affect them, on practices for minimizing
contamination, and on effective eyewitness identification protocols. In addition to instruction at the police
academy, officers should receive periodic refresher training, and officers assigned to investigative units
should receive in-depth instruction. Dispatchers should be trained not to “leak” information from one
caller to the next and to ask for information in a non-leading way. Police officers should be trained to ask
2

T. Hastie, R. Tibshirani, and J.H. Friedman, The Elements of Statistical Learning: Data Mining, Inference, and
Prediction (New York: Springer, 2009).

72
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
open ended questions, avoid suggestiveness, and efficiently manage scenes with multiple witnesses (e.g.,
minimize interactions among witnesses).
Recommendation #2: Implement Double-Blind Lineup and Photo Array Procedures
Decades of scientific evidence demonstrate that expectations can bias perception and judgment and
that expectations can be inadvertently communicated.3 Even when lineup administrators scrupulously
avoid comments that could identify which person is the suspect, unintended body gestures, facial
expressions, or other nonverbal cues have the potential to inform the witness of his or her location in the
lineup or photo array.
Double-blinding is central to the scientific method because it minimizes the risk that experimenters
might inadvertently bias the outcome of their research, finding only what they expected to find. For
example, in medical clinical trials, double-blind designs are crucial to account for experimenter biases,
interpersonal influences, and placebo effects.
To minimize inadvertent bias, double-blinding procedures are sometimes used in which the test
administrator does not know the composition of the photo array or lineup. If administrators are not
involved with construction of the lineup and are unaware of the placement of the potential suspect in the
sequence, they cannot influence the witness.
Some in the law enforcement community have responded to calls for double-blind lineup
administration with concern, citing the potential for increased financial costs and human resource
demands. The committee believes there are ways to reduce these costs and recommends that police
departments consider procedures and new technologies that increase efficiency of data acquisition under
double-blind procedures or those procedures that closely approximate double-blind procedures. If an
administrator who does not know the identity of the suspect cannot be assigned to the task, a non-blind
administrator (one knowing the status of the individuals in the lineup) might use a computer-automated
presentation of lineup photos. If computer-based presentation technology is unavailable, the administrator
could place photos in numbered folders that are then shuffled, as is current practice in some jurisdictions.
The committee recommends blind (double-blind or blinded) administration of both photo arrays
and live lineups and the adoption of clear, written policies and training on photo array and live lineup
administration. Police should use blind procedures to avoid the unintentional or intentional exchange of
information that might bias an eyewitness. The “blinded” procedure minimizes the possibility of either
intentional or inadvertent suggestiveness and thus enhances the fairness of the criminal justice system.
Suggestiveness during an identification procedure can result in suppression of both out-of-court and incourt identifications and thereby seriously impair the prosecutions’s ability to prove its case beyond a
reasonable doubt. The use of double-blind procedures will eliminate a line of cross-examination of
officers in court.
Recommendation #3: Develop and Use Standardized Witness Instructions
The committee recommends the development of a standard set of easily understood instructions to
use when engaging a witness in an identification procedure.
Witnesses should be instructed that the perpetrator may or may not be in the photo array or lineup
and that the criminal investigation will continue regardless of whether the witness selects a suspect.
Administrators should use witness instructions consistently in all photo arrays or lineups, and can use prerecorded instructions or read instructions aloud, in the manner of the mandatory reading of Miranda
Rights. Accommodations should be made when questioning non-English speakers or those with restricted
linguistic ability. Additionally, the committee recommends the development and use of a standard set of
instructions for use with a witness in a showup.
3

See Box 2-1.

73
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Recommendation #4: Document Witness Confidence Judgments
Evidence indicates that self-reported confidence at the time of trial is not a reliable predictor of
eyewitness accuracy.4 The relationship between the witness’ stated confidence and accuracy of
identifications may be greater at the moment of initial identification than at the time of trial. However, the
strength of the confidence-accuracy relationship varies, as it depends on complex interactions among such
factors as environmental conditions, persons involved, individual emotional states, and more. 5
Expressions of confidence in the courtroom often deviate substantially from a witness’ initial confidence
judgment, and confidence levels reported long after the initial identification can be inflated by factors
other than the memory of the suspect. Thus, the committee recommends that law enforcement document
the witness’ level of confidence verbatim at the time when she or he first identifies a suspect, as
confidence levels expressed at later times are subject to recall bias, enhancements stemming from
opinions voiced by law enforcement, counsel and the press, and to a host of other factors that render
confidence statements less reliable. During the period between the commission of a crime and the formal
identification procedure, officers should avoid communications that might affect a witness’ confidence
level. In addition, to avoid increasing a witness’ confidence, the administrator of an identification
procedure should not provide feedback to a witness. Following a formal identification, the administrator
should obtain level of confidence by witness’ self-report (this report should be given in the witness’ own
words) and document this confidence statement verbatim. Accommodations should be made for nonEnglish speakers or those with restricted linguistic ability.
Recommendation #5: Videotape the Witness Identification Process
The committee recommends that the video recording of eyewitness identification procedures
become standard practice.
Although videotaping does have drawbacks (e.g., costs, witness advocates opposing videotaping of
witnesses’ faces, and witnesses not wanting to be videotaped), it is necessary to obtain and preserve a
permanent record of the conditions associated with the initial identification. When necessary, efforts
should be made to obtain non-intrusive recordings of the initial identification process and to
accommodate non-English speakers or those with restricted linguistic ability. Measures should also be
taken to protect the identity of eyewitnesses who may be at risk of harm because they make an
identification.
RECOMMENDATIONS TO STRENGTHEN THE VALUE
OF EYEWITNESS IDENTIFICATION EVIDENCE IN COURT
The best guidance for legal regulation of eyewitness identification evidence comes not from
constitutional rulings, but from the careful use and understanding of scientific evidence to guide factfinders and decision-makers. The Manson v. Brathwaite test under the Due Process Clause of the U.S.
Constitution for assessing eyewitness identification evidence was established in 1977, before much
4

See, e.g., C. M. Allwood, J. Knutsson, and P. A. Granhag, “Eyewitnesses under influence: How feedback
affects the realism in confidence judgements,” Psychology, Crime, and Law 12(1): 25–38 (2006); B. H. Bornstein
and D. J. Zickafoose, “’I know I know it, I know I saw it’: The stability of the confidence-accuracy relationship
across domain,” Journal of Experimental Psychology-Applied 5(1): 76–88 (1999); P. A. Granhag, L. A. Stromwall,
and C. M. Allwood, “Effects of reiteration, hindsight bias, and memory on realism in eyewitness confidence,”
Applied Cognitive Psychology 14(5): 397–420 (2000); H. L. Roediger, III, J. T. Wixted, and K. A. DeSoto, “The
Curious Complexity between Confidence and Accuracy in Reports from Memory” in Memory and Law, ed. L.
Nadel and W. P. Sinnott-Armstrong (Oxford: Oxford University Press, 2012).
5
See, e.g., J. M. Talarico and D. C. Rubin, “Confidence, Not Consistency, Characterizes Flashbulb Memories,”
Psychological Science 14(5): 455–461 (September 2003).

74
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
applied research on eyewitness identification had been conducted. That test evaluates the “reliability” of
eyewitness identifications using factors derived from prior rulings and not from empirically validated
sources. As critics have pointed out, the Manson v. Brathwaite test includes factors that are not diagnostic
of reliability. Moreover, the test treats factors such as the confidence of a witness as independent markers
of reliability when, in fact, it is now well established that confidence judgments may vary over time and
can be powerfully swayed by many factors. While some states have made minor changes to the due
process framework, (e.g., by altering the list of acceptable “reliability” factors; see Chapter 3), wholesale
reconsideration of this framework is only a recent development (e.g., the recent decisions by state
supreme courts in New Jersey and Oregon; see Chapter 3).
Recommendation #6: Conduct Pretrial Judicial Inquiry
Eyewitness testimony is a type of evidence where (as with forms of forensic trace evidence)
contamination may occur pre-trial. Judges rarely make pre-trial inquiries about evidence in criminal cases
without one of the parties first raising an objection. In cases involving eyewitness evidence, however,
parties may not be sufficiently knowledgeable about the relevant scientific research to raise concerns.
Judges have an affirmative obligation to insure the reliability of evidence presented at trial. To meet
this obligation, the committee recommends that, as appropriate, a judge make basic inquiries when
eyewitness identification evidence is offered. While the contours of such an inquiry would need to be
established on a case-by-case basis, at a minimum, the judge could inquire about prior lineups, what
information had been given to the eyewitness before the lineup, what instructions had been given to the
eyewitness in connection with administering the lineup, and whether the lineup had been administered
“blindly.” The judge could also entertain requests from the parties for additional discovery, and could ask
the parties to brief any issues raised by these inquiries. A judge also could review reports of the
eyewitness’ confidence and any recordings of the identification procedures. When assessing the reliability
of an identification, a judge could also inquire as to what eyewitness identification procedures the agency
had in place and the degree to which they were followed. Both pre-trial judicial inquiries and any
subsequent judicial review would create an incentive for agencies to adopt written eyewitness
identification procedures and to document the identifications themselves.
If these initial inquiries raise issues with the identification process, a judge could conduct a pre-trial
hearing to review the reliability and admissibility of eyewitness identification evidence and to assess how
it should be treated at trial if found admissible. If indicia of unreliable eyewitness identifications are
present, the judge should apply applicable law in deciding whether to exclude the identifications or
whether some lesser sanction is appropriate. As discussed in the sections that follow, a judge may limit
portions of the testimony of the eyewitness. A judge can also ensure that the jury is provided with a
scientific framework within which to evaluate the evidence.
Recommendation #7: Make Juries Aware of Prior Identifications
The accepted practice of in-court eyewitness identifications can influence juries in ways that crossexamination, expert testimony, or jury instructions are unable to counter effectively. Moreover, as
research suggests (see Chapters 4 and 5), the passage of time since the initial identification may mean that
a courtroom identification is a less accurate reflection of an eyewitness’ memory. In-court confidence
statements may also be less reliable than confidence judgments made at the time of an initial out-of-court
identification; as memory fails and/or confidence grows disproportionately. The confidence of an
eyewitness may increase by the time of the trial as a result of learning more information about the case,
participating in trial preparation, and experiencing the pressures of being placed on the stand.
An identification of the kind dealt with in this report typically should not occur for the first time in
the courtroom. If no identification procedure was conducted during the investigation, a judge should
consider ordering that an identification procedure be conducted before trial. In any case, whenever the
75
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
eyewitness identifies a suspect in the courtroom, it is important for jurors to hear detailed information
about any earlier identification, including the procedures used and the confidence expressed by the
witness at that time. The descriptions of prior identifications and confidence at the time of those earlier
out-of-court identifications provide more useful information to the fact-finders and decision-makers.
Accordingly, the committee recommends that judges take all necessary steps to make juries aware of
prior identifications, the manner and time frame in which they were conducted, and the confidence level
expressed by the eyewitness at the time.
Recommendation #8: Use Scientific Framework Expert Testimony
The committee finds that a scientific framework describing what factors may influence a witness’
visual experience of an event and the resolution and fidelity of that experience, as well as factors that
underlie and influence subsequent encoding, storage, and recall of memories of an event, can inform the
fact-finder in a criminal case. As discussed throughout this report, many scientifically established aspects
of eyewitness memory are counterintuitive and may defy expectations. Jurors will likely need assistance
in understanding the factors that may affect the accuracy of an identification. In many cases this
information can be most effectively conveyed by expert testimony.
Contrary to the suggestion of some courts, the committee recommends that judges have the
discretion to allow expert testimony on relevant precepts of eyewitness memory and identifications.
Expert witnesses can explain scientific research in detail, capturing the nuances of the research, and
focusing their testimony on the most relevant research. Expert witnesses can convey current information
based on the state of the research at the time of a trial. Expert witnesses can also be cross-examined, and
limitations of the research can be expressed to the jury.
Certainly, qualified experts will not be easy to locate in a given jurisdiction; and indigent defendants
may not be able to afford experts absent court funds. Moreover, once the defense secures an expert, the
prosecution may retain a rebuttal expert, adding complexity to the litigation. Further investigation may
explore the effectiveness of expert witness presentation of relevant scientific findings compared with jury
instructions. Until we have a clearer understanding of the strengths and weaknesses of this technique, the
committee views expert testimony as an appropriate and effective means of providing the jury with
information to assess the strength of the eyewitness identification.
Expert witnesses should not be permitted to testify without limits. An expert explaining the relevant
scientific framework can describe the state of the research and focus on the factors that are particularly
relevant in a given case. However, an expert must not be allowed to testify beyond the limits of his or her
expertise. Although current scientific knowledge would allow an expert to inform the jury of factors
bearing on their evaluation of an eyewitness’ identification, the committee has seen no evidence that the
scientific research has reached the point that would properly permit an expert to opine, directly or through
an equivalent hypothetical question, on the accuracy of an identification by an eyewitness in a specific
case.
In many jurisdictions, expert witnesses who can testify regarding eyewitness identification evidence
may be unavailable. In state courts, funding for expert witnesses may be far more limited than funding in
federal courts. The committee recommends that local jurisdictions make efforts to ensure that defendants
receive funding to obtain access to qualified experts.
Recommendation #9: Use Jury Instructions as an Alternative Means to Convey Information
The committee recommends the use of clear and concise jury instructions as an alternative means
of conveying information regarding the factors that the jury should consider.

76
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Jury instructions should explain, in clear language, the relevant principles. Like the New Jersey
instructions,6 the instructions should allow judges to focus on factors relevant to the specific case, since
not all cases implicate the same factors. Jury instructions do not need to be as detailed as the New Jersey
model instructions and do not need to omit all reference to underlying research. With the exception of the
New Jersey instructions, jury instructions have tended to address only certain subjects, or to repeat the
problematic Manson v. Brathwaite language, which was not intended as instructions for jurors.
Appropriate legal organizations, together with law enforcement, prosecutors, defense counsel, and
judges, should convene a body to establish model jury instructions regarding eyewitness identifications.
RECOMMENDATIONS TO IMPROVE THE SCIENTIFIC FOUNDATION
UNDERPINNING EYEWITNESS IDENTIFICATION RESEARCH
Basic scientific research on visual perception and memory provides important insight into the
factors that can limit the fidelity of eyewitness identification (see Chapter 4). Research targeting the
specific problem of eyewitness identification (see Chapter 5) complements basic scientific research.
However, this strong scientific foundation remains insufficient for understanding the strengths and
limitations of eyewitness identification procedures in the field. Many of the applied studies on key factors
that directly affect eyewitness performance in the laboratory are not readily applicable to actual practice
and policy. Applied research falls short because of a lack of reliable or standardized data from the field, a
failure to include a range of practitioners in the establishment of research agendas, the use of disparate
research methodologies, failure to use transparent and reproducible research procedures, and inadequate
reporting of research data. The task of guiding eyewitness identification research toward the goal of
evidence-based policy and practice will require collaboration in the setting of research agendas and
agreement on methods for acquiring, handling, and sharing of data.
Recommendation #10: Establish a National Research Initiative on Eyewitness Identification
To further our understanding of eyewitness identification, the committee recommends the
establishment of a National Research Initiative on Eyewitness Identification (hereinafter, the Initiative).
The Initiative should involve the academic research community, law enforcement community, the federal
government, and philanthropic organizations. The Initiative should (1) establish a research agenda to
guide research for the next decade; (2) formulate practice- and policy-relevant research questions; (3)
identify opportunities for additional data collection; (4) systematically review research to examine
emerging findings on the impact of system and estimator variables; (5) translate research findings into
policies and procedures that are both practical and appropriate for law enforcement; and (6) set priorities
and timelines for issues to be addressed, the conduct of research, the development of best practices, and
formal assessments.
The committee notes that there appear to be few existing partnerships between the scientific
community and law enforcement organizations and therefore recommends that The National Science
Foundation (NSF) and the National Institute of Standards and Technology (NIST) take a leadership role
working with other federal agencies, such as the National Institute of Justice (NIJ), the Bureau of Justice
Statistics (BJS), and the Federal Bureau of Investigation (FBI), to support such collaborations.
6

New Jersey Criminal Model Jury Instructions, Identification (July 19, 2012), available at:
http://www.judiciary.state.nj.us/pressrel/2012/jury_instruction.pdf; New Jersey Court Rule 3:11, Record of an Outof-Court Identification Procedure (July 19, 2012), available at:
http://www.judiciary.state.nj.us/pressrel/2012/new_rule.pdf, New Jersey Court Rule 3:13-3, Discovery and
Inspection (July 19, 2012), available at: http://www.judiciary.state.nj.us/pressrel/2012/rev_rule.pdf.

77
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
The impact on society of innocents being incarcerated while perpetrators remain free, in conjunction
with limited federal resources, highlights the need for both public and private support for this Initiative.
To enhance the scientific foundation of eyewitness identification research and practice, the Initiative
should commit to the following:
a. Include a practice- and data-informed research agenda that incorporates input from law
enforcement and the courts and establishes methodological and reporting standards for research to
assess the fundamental performance of various aspects of eyewitness identification procedures as
well as synthesize research findings across studies.
b. Develop protocols and policies for the collection, preservation, and exchange of field data
that can be used jointly by the scientific and law enforcement communities. Data collection
procedures used in the field should be developed to ensure the relevance of the collected data, to
facilitate analysis of the data, and to minimize potential bias and loss of data through incomplete
recording strategies.
Law enforcement agencies should take the lead in collecting, maintaining, and sharing relevant data
from the field. Much of the data that would be useful for the evaluation of eyewitness identification
procedures have been collected in the form of administrative records and may be readily adapted for
use in research. Comprehensive data should be collected on lineup composition and witness
selections (i.e., fillers, non-identifications, and position of suspect in lineup).
c. Develop and adopt guidelines for the conduct and reporting of applied scientific research on
eyewitness identification that conform to the highest scientific standards. All eyewitness research,
including field-based studies, laboratory-based studies, and research synthesis, should use rigorous
research methods and provide detailed reporting of both methods and results, including (1) preregistration of all study protocols; (2) investigation of research questions and hypotheses informed
by the needs of practice and policy; (3) adoption of strict operationalization of key measures and
objective data collection; (4) development of experimental designs informed by analytical concerns;
(5) use of proper statistical procedures that account for the often non-traditional nature of data in this
field (e.g., estimates of effects with appropriate statements of uncertainty, multiple responses from
different scenarios from the same individuals, effects of order and time of presentation when
important, treatment of extreme observations or outliers); (6) reporting of participant recruitment
and selection and assignment to conditions; (7) complete reporting of findings including effect sizes
and associated confidence intervals for both significant and non-significant effects; and (8)
derivation of conclusions that are grounded firmly in the findings of the study, are framed in the
context of the strengths and limitations of study methodology, and clearly state their implications for
practice and policy decisions.
Strict adherence to guidelines for eyewitness identification research will result in more credible
research findings that can guide policy and practice. Research that conforms to guidelines will
withstand rigorous scrutiny by peers, will be verifiable through replication, and will permit inclusion
in systematic reviews, leading to greater confidence in the validity and generalizability of findings.
d. Adopt rigorous standards for systematic reviews and meta-analytic studies. Meta-analyses of
primary studies should be conducted only in the context of systematic reviews that locate and
critically appraise all research findings, including those from unpublished studies. Analyses should
consistently appraise and account for possible biases in the included research. Studies that do not
adequately conduct or report research methods, such as randomization, should be identified in the
findings. Sensitivity analyses considering impacts of lower quality or inadequately reported studies
on pooled effect estimates should be conducted and reported. When attempting to draw conclusions
78
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
from studies with missing data, reviewers should first attempt to contact the authors of the research
for additional information. When missing data cannot be retrieved from researchers, imputation
methods should, if used, be specific, transparent, and reproducible. Statistical methods for metaanalysis should conform to current best practice, using models appropriate to the level of
heterogeneity of results across studies, computing both point estimates and confidence intervals
around effect sizes, and translating the results of meta-analyses into terms that are both
understandable and useful to practice and policy decision-makers.
e. Provide basic instruction for police, prosecutors, defense counsel, and judges on aspects of the
scientific method relevant to eyewitness identifications procedures (e.g., the rationale for blinded
administration), including principles of research design and the uncertainties associated with data
analysis. Training should cover the importance of data collection and interpretation, including the
role of standardized eyewitness identification procedures and documentation of witness statements
of confidence. Competencies acquired through such training (quantitative reasoning, understanding
principles of research design, and recognition of data uncertainties) are likely to apply to issues
beyond eyewitness identification. For example, the knowledge and skills from training can be
applied to other issues that personnel face, either in forensic science technologies or in process
administration, evaluation, and quality improvement. Similarly, scientists will benefit from a greater
knowledge of legal issues, standards, and procedures related to the problem of eyewitness
identification. Training of both communities (law and science) will enhance communication and
lead to productive collaborations.
The collaborative research initiative between researchers and law enforcement communities will be
challenging as it will necessitate (1) standardized police procedures;7 (2) systematic valid evidence
collection and data entry and analysis; and (3) education and training for both researchers and law
enforcement professionals on the differences between these two communities in their use of terms and
considerations of standards of evidence and uncertainties in data. These three elements of a collaborative
initiative are critical to advancing the science related to eyewitness identifications, as each bears directly
on the integrity of the foundation upon which the efficacy and validity of current and future practices will
be judged. Without such a foundation, practical advances in our scientific understanding are unlikely to
occur.
The committee further recommends that the Initiative support research to better understand the
following: (1) the variables that affect the accuracy, precision, and reliability of eyewitness
identifications, and how those variables interact and vary in practice; (2) the (possibly joint) impact of
estimator and system variables on both identification accuracy and response bias; (3) best practices for
probing witness memory with the least potential for bias or contamination; (4) best strategies to assess
witnesses’ confidence levels when making an identification; and (5) appropriate types of instructions for
police, witnesses, and juries to best inform and facilitate the collection and interpretation of eyewitness
identifications; (6) photo array composition and procedures; (7) identification procedures in the field
(showups); (8) innovative technologies that might increase the reliability of eyewitness testimony (e.g.,
algorithm-based computer face recognition software, computer administered photo arrays, and mobile
technologies with photo identification programs); and (9) the most effective means of informing jurors
how to consider the factors that affect the strengths and weaknesses of eyewitness identification evidence.

7

The term standardized procedures refers to the notion that professionals reliably follow the same set of steps or
procedures. Such standardization ensures that data across cases can be considered comparable and, to a greater
extent, more reliable. Although reliability is not equivalent to validity, it is essential before researchers can assess
questions of validity. Without standardized procedures, valid comparisons between departments and regions of the
country cannot be achieved.

79
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Recommendation #11: Conduct Additional Research on System and Estimator Variables
Among the many variables that can affect eyewitness identification, the procedures for constructing
a lineup have received the greatest attention in recent years. As discussed in Chapter 5, the question as to
whether a simultaneous or sequential lineup is preferred is a specific case of the more general question of
what conditions might improve the performance of an eyewitness. The answer to that question depends
upon the criteria used to evaluate performance, and much of the debate has thus focused on the analysis
tools for evaluation. These tools have improved significantly over the years, beginning with the use of a
diagnosticity ratio, which uses the likelihood that the person identified is actually guilty as an evaluation
criterion. More recently, the diagnosticity ratio approach has been augmented by analysis of Receiver
Operating Characteristics (ROC analysis), which uses a measure of discriminability (i.e. a measure of
how well the witness can discriminate between different possible matches to his/her memory of the face
of the culprit) as an evaluation criterion. In principle, ROC analysis is a positive step, if only because it
incorporates more information (i.e. the earlier diagnosticity ratio is one component of the ROC analysis).
But a more complex question concerns how policy-makers and practitioners should weigh the two
evaluation criteria that have been considered thus far – likelihood of guilt and discriminability – when
making a decision about which lineup procedures to adopt. The answer is particularly nuanced because
the two criteria do not always lead to the same conclusion; one lineup procedure may yield poorer
discriminability while at the same time increasing the likelihood that the identified person is actually
guilty.
The committee concludes that there should be no debate about the value of greater discriminability –
to promote a lineup procedure that yields less discriminability would be akin to advocating that the lineup
be performed in dim instead of bright light. For this reason, the committee recommends broad use of
statistical tools that can render a discriminability measure to evaluate eyewitness performance. But a
lineup procedure that improves discriminability can yield greater or lesser likelihood of correct
identification, depending on how the procedure is applied (see Chapter 5). For lineup procedures that
yield greater discriminability, greater likelihood of correct identification would appear preferable and can
be achieved by methods that elicit a more conservative response bias, such as a sequential (relative to
simultaneous) lineup procedure.8 The committee thus recommends a rigorous exploration of methods
that can lead to more conservative responding (such as witness instructions) but do not compromise
discriminability.
In view of these considerations of performance criteria and recommendations about analysis tools,
can we draw definitive conclusions about which lineup procedure (sequential or simultaneous) is
preferable? At this point, the answer is no. Using discriminability as a criterion, there is, as yet, not
enough evidence for the advantage of one procedure over another. The committee thus recommends that
caution and care be used when considering changes to any existing lineup procedure, until such time as
there is clear evidence for the advantages of doing so. From a larger perspective, the identification of
factors (such as specific lineup procedures or states of other system variables) that can objectively
improve eyewitness identification performance must be among the top priorities for this field. This leads
us to three additional recommendations.
a. The committee recommends a broad exploration of the merits of different statistical tools for use
in the evaluation of eyewitness performance. ROC analysis represents an improvement over a single
diagnosticity ratio, yet there are well-documented quantitative shortcomings to the ROC approach.
But are there alternatives? As noted in Chapter 5, the task facing an eyewitness is a binary
classification task and there exist many powerful statistical tools for evaluation of binary
classification performance that are widely used, for example, in the field of machine learning. While
8

The committee stresses, however, that adoption of a more conservative response bias necessitates a
compromise by which fewer lineup “picks” are made overall and thus fewer guilty suspects are identified (see
Chapter 5).

80
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
none of these tools has been vetted for application to the problem of eyewitness identification, they
offer a potentially rich resource for future investigation in this field.
b. The alternative (sequential) lineup procedure was introduced as part of an effort to improve
eyewitness performance. While, as noted above, it remains unclear whether the procedure has
improved eyewitness performance, that goal is still primary. In an effort to achieve that goal, many
studies over the past three decades have explored the possibility that other factors may also affect
performance, but until recently these investigations have not evaluated performance using a
discriminability measure. The committee therefore recommends a broad exploration of the effects
of different system variables (e.g. additional variants on lineup procedures, witness lineup
instructions) and estimator variables (e.g. presence or absence of weapon, elapsed time between
incident and identification task, levels of stress), and – importantly –interactions between these
variables, using either the ROC approach or other tools for evaluation of binary classifiers that can
be shown to have advantages over existing analytical methods.
c. Building upon the committee’s call for a practice- and data-informed research agenda that
incorporates input from law enforcement and the courts and establishes methodological and
reporting standards for research, the committee recommends that the scientific community engaged
in studies of eyewitness identification performance work closely with law enforcement to identify
other system and estimator variables that might influence performance and practical issues that
might preclude certain strategies for influencing performance. In addition, the committee
recommends that policy decisions regarding changes in procedure should be made on the basis of
evidence of superiority and should be made in consultation with police departments to determine
which procedure yields the best combination of performance and practicality.
CONCLUSION
Eyewitness identification can be a powerful tool. As this report indicates, however, the malleable
nature of human visual perception, memory, and confidence; the imperfect ability to recognize
individuals; and policies governing law enforcement procedures can result in mistaken identifications
with significant consequences. New law enforcement training protocols, standardized procedures for
administering lineups, improvements in the handling of eyewitness identification in court, and better data
collection and research on eyewitness identification, can improve the accuracy of eyewitness
identifications.

81
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Appendixes

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

Appendix A
Biographical Information of Committee and Staff
Co-Chairs
Thomas D. Albright, Ph.D., (NAS) is Professor and Conrad T. Prebys Chair in Vision Research at the
Salk Institute for Biological Studies, where he joined the faculty in 1986. Dr. Albright is also Director of
the Salk Institute Center for the Neurobiology of Vision, Adjunct Professor of Psychology and
Neurosciences at the University of California, San Diego, and Visiting Centenary Professor at the Indian
Institute of Science, Bangalore.
Dr. Albright is an authority on the neural basis of visual perception, memory, and visually guided
behavior. Probing the relationship between the activity of brain cells and perceptual state, his laboratory
seeks to understand how visual perception is affected by attention, behavioral goals, and memories of
previous experiences. His discoveries address the ways in which context influences visual perceptual
experience and the mechanisms of visual associative memory and visual imagery. An important goal of
this work is the development of therapies for blindness and perceptual impairments resulting from
disease, trauma, or developmental disorders of the brain. A second aim of Dr. Albright’s work is to use
our growing knowledge of brain, perception, and memory to inform design in architecture and the arts,
and to leverage societal decisions and public policy.
Albright received a Ph.D. in psychology and neuroscience from Princeton University in 1983. He is a
recipient of numerous honors for his work, including the National Academy of Sciences Award for
Initiatives in Research. Dr. Albright is a member of the National Academy of Sciences, a fellow of the
American Academy of Arts and Sciences, a fellow of the American Association for the Advancement of
Science, and an associate of the Neuroscience Research Program. He is currently president of the
Academy of Neuroscience for Architecture; a member of the National Academy of Sciences Committee
on Science, Technology, and Law; and serves on the Scientific Advisory Committee for the Indian
National Brain Research Center.
Jed S. Rakoff, J.D., has been a United States District Judge for the Southern District of New York since
1996. Prior to his appointment, he was a federal prosecutor (1973–80) and a criminal defense lawyer at
two large New York law firms (1980–95). Judge Rakoff is coauthor of five books and the author of more
than 110 published articles, 500 speeches, and 1,200 judicial opinions. He has been an Adjunct Professor
at Columbia Law School since 1988, teaching upper class seminars in science and the law, class actions,
white collar crime, and the interplay of civil and criminal law.
Judge Rakoff is a Commissioner on the National Commission on Forensic Science and is a former member
of the Governance Board of the MacArthur Foundation Initiative on Law and Neuroscience. He was a
member of the National Research Council Committee on the Development of the Third Edition of the
Reference Manual on Scientific Evidence and the Committee on the Review of the Scientific Approaches
Used During the FBI’s Investigation of the 2001 Bacillus anthracis Mailings. He is a member of the
American Academy of Arts and Sciences and the American Law Institute. He is a Judicial Fellow at the
American College of Trial Lawyers, a former director of the New York Council of Defense Lawyers, and
former chair of the Criminal Law Committee, New York City Bar Association.
Judge Rakoff received a B.A. from Swarthmore College in 1964, an M.Phil. from Oxford University in
1966, and a J.D. from Harvard Law School in 1969.
85
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Members
William G. Brooks III is the Chief of the Norwood, Massachusetts Police Department. He began his
tenure on May 1, 2012. He served as the Deputy Chief with the Wellesley Police Department from 2000
to 2012. As Deputy Chief, Brooks was involved in hiring, discipline, administration, budgeting, training,
and multi-agency coordination. Prior to 2000, he served as a patrolman with the Westwood Police
Department from 1977 to 1982 and as an officer with the Norwood Police Department from 1982 to 2000.
In Norwood, he served as a patrolman and sergeant and as a detective sergeant for fourteen years,
supervising all criminal investigations conducted by detectives. Chief Brooks has been a police academy
instructor for thirty years and a presenter on eyewitness identification for six years. He presents nationally
on behalf of the Innocence Project, is a member of the Massachusetts Supreme Judicial Court’s Study
Committee on Eyewitness Identification, and was the 2012 recipient of the Innocence Network’s
Champion of Justice Award. Chief Brooks holds a master’s degree in criminal justice and is a graduate of
the FBI National Academy.
Joe S. Cecil, Ph.D., J.D, is a Project Director in the Division of Research at the Federal Judicial Center.
Currently, he is directing the Center’s Program on Scientific and Technical Evidence. As director, Dr.
Cecil is responsible for judicial education and training in the area of scientific and technical evidence and
served as principal editor of the first two editions of the Center’s Reference Manual on Scientific
Evidence, which is the primary source book on evidence for federal judges. He also has published several
articles on the use of court-appointed experts. Dr. Cecil is currently directing a research project that
examines the difficulties that arise with expert testimony in federal courts, with an emphasis on clinical
medical testimony and forensic science evidence. Other areas of Dr. Cecil’s research interest include
federal civil and appellate procedure, jury competence in complex civil litigation, and assessment of rule
of law in emerging democracies. Dr. Cecil serves on the editorial boards of social science and legal
journals. He previously served on the National Academies’ Panel on Confidentiality and Data Access, and
the Committee on Identifying the Needs of the Forensic Sciences Community. He currently is a member
of the National Academy of Sciences’ Committee on Science, Technology, and Law and was a member
of its Access to Research Data: Balancing Risks and Opportunities subcommittee. Dr. Cecil received his
law doctorate (in psychology) and law degree from Northwestern University.
Winrich Freiwald, Ph.D., is Assistant Professor, Laboratory of Neural Systems, The Rockefeller
University. Dr. Freiwald is interested in the neural processes that form object representations as well as
those that allow attention to make those representations available for social behavior and cognition. Dr.
Freiwald co-discovered a specialized neural machinery for face processing located in the temporal and
frontal lobes of the brain. He and his colleagues further showed that this machinery is composed of a
small network of a fixed number of face selective regions, termed face patches, each dedicated to a
different aspect of face processing and all closely connected with each other. Dr. Freiwald’s laboratory
aims to understand the inner workings of this system, from the level of individual cells to the interactions
of brain areas, in order to answer questions such as: How does face selectivity emerge in a single cell?
How is information transformed from one face patch to another? What is the contribution of each face
patch to different face recognition abilities like the recognition of a friend or a smile? How do the
different face patches interact in different tasks? And how is information extracted from a patch when a
perceptual decision is made?
Dr. Freiwald, a native of Oldenburg, Germany, performed his graduate work at the Max Planck Institute
for Brain Research in Frankfurt and received his Ph.D. from Tübingen University in 1998. He then joined
the Institute for Brain Research at the University of Bremen as a lecturer. Starting in 2001, he worked as a
postdoctoral fellow at the Massachusetts Institute of Technology, Massachusetts General Hospital,
Harvard Medical School, and the Hanse Institute for Advanced Study in Delmenhorst, Germany. He was
head of the primate brain imaging group at the Centers for Advanced Imaging and Cognitive Sciences in
86
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Bremen from 2004 to 2008 and a visiting associate at the California Institute of Technology in 2009. He
joined The Rockefeller University as assistant professor in 2009. Dr. Freiwald was named a Pew Scholar
in 2010, a McKnight Scholar in 2011, and a NYSCF – Robertson Neuroscience Investigator in 2013.
Brandon L. Garrett is the Roy L. and Rosamond Woodruff Morgan Professor of Law at the University
of Virginia Law School. Garrett joined the law faculty in 2005. His research and teaching interests
include criminal procedure, wrongful convictions, habeas corpus, corporate crime, scientific evidence,
civil rights, civil procedure, and constitutional law.
Mr. Garrett’s recent research includes studies of DNA exonerations, organizational prosecutions, and
eyewitness identification procedures in Virginia. In 2011, Harvard University Press published Mr.
Garrett’s book, Convicting the Innocent: Where Criminal Prosecutions Go Wrong, examining the cases of
the first 250 people to be exonerated by DNA testing. In 2013, Foundation Press published his coauthored casebook, Federal Habeas Corpus: Executive Detention and Post-Conviction Litigation. Mr.
Garrett is currently completing a new book, in contract with Harvard University Press, examining
corporate prosecutions.
Mr. Garrett attended Columbia Law School, where he was an articles editor of the Columbia Law Review
and a Kent Scholar. After graduating, he clerked for the Honorable Pierre N. Leval of the U.S. Court of
Appeals for the Second Circuit. He then worked as an associate at Neufeld, Scheck & Brustin LLP in
New York City.
Karen Kafadar, Ph.D., is Commonwealth Professor and Chair of Statistics at the University of Virginia.
Dr. Kafadar received her B.S. in Mathematics and M.S. in Statistics at Stanford University and her Ph.D.
in Statistics from Princeton University. Before joining the Statistics Department in 2014, she was
Mathematical Statistician at National Institute of Standards and Technology, member of the technical
staff at Hewlett Packard's RF/Microwave R&D Department, Fellow in the Division of Cancer Prevention
at National Cancer Institute, Professor and Chancellor's Scholar at University of Colorado-Denver, and
Rudy Professor of Statistics at Indiana University – Bloomington. Her research focuses on robust
methods, exploratory data analysis, characterization of uncertainty in the physical, chemical, biological,
and engineering sciences, and methodology for the analysis of screening trials, with awards from CDC,
American Statistical Association (ASA), and American Society for Quality.
Kafadar was editor of Technometrics and the review section of the Journal of the American Statistical
Association and is currently Biology & Genetics Editor for The Annals for Applied Statistics. She has
served on several National Research Council committees and is a past or present member on the
governing boards for ASA, Institute of Mathematical Statistics, International Statistical Institute, and
National Institute of Statistical Sciences. She is a Fellow of the ASA, the American Association for the
Advancement of Science, and the International Statistics Institute; she has authored more than 100 journal
articles and book chapters; and has advised numerous M.S. and Ph.D. students.
A.J. Kramer, J.D., is Federal Public Defender for the District of Columbia. He earned a Bachelor’s of
Arts from Stanford University (1975), followed by a Juris Doctorate from the Boalt Hall School of Law at
the University of California at Berkeley (1979). Mr. Kramer clerked for the Honorable Procter Hug, Jr., at
the United States Court of Appeals for the Ninth Circuit in Reno, Nevada. He spent seven years as an
Assistant Federal Public Defender in San Francisco, California, followed by three years as the Chief
Assistant Federal Public Defender in Sacramento, California. He taught legal research and writing at
Hastings College of the Law, University of California, San Francisco from 1982 to 1988. Mr. Kramer was
appointed Federal Public Defender for the District of Columbia in 1990.

87
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
A permanent faculty member at the National Criminal Defense College in Macon, Georgia, and at the
Western Trial Advocacy Institute in Laramie, Wyoming, Mr. Kramer is a Fellow of the American College
of Trial Lawyers. He is currently a member of the American Bar Association Criminal Justice Section
Council and a member of the United States Judicial Conference Advisory Committee on the Rules of
Evidence.
Scott McNamara, J.D., graduated from Syracuse University with a major in mathematics. Mr.
McNamara attended Vermont Law School, graduating cum laude in 1991. On July 20, 1992, he became
an Oneida County Assistant District Attorney. As such, he handled thousands of cases with a
concentration in narcotic and homicide prosecutions. McNamara was the Bureau Chief of the Narcotics
Unit for twelve years, and he was also the First Assistant District Attorney for six years. During his years
in the District Attorney’s Office, he was a member and the lead prosecutor assigned to the Oneida County
Drug Task Force. He also chaired the Oneida County District Attorney’s Office Death Penalty
Committee. From 2001 to 2006, Mr. McNamara represented the District Attorney’s Office on the Joint
Terrorism Task Force. In January of 2007, Mr. McNamara took office as the Oneida County District
Attorney and has since been elected, and re-elected, by the citizens of Oneida County. His tenure as
District Attorney has been one of proactive engagement and problem-solving. He has created an
Economic Crime Unit, a Conviction Integrity Unit, and he has appointed a community liaison to improve
communication and accessibility between the District Attorney’s Office and the diverse population it
serves. In addition, Mr. McNamara initiated a strategy of video recording all police interrogations in
Oneida County. He has always maintained that his goal as the county’s chief law enforcement officer is to
continue the legacy of bringing justice to those victimized by crime while recognizing the need to
safeguard and enhance fairness within the legal system.
For ten years, Mr. McNamara taught search and seizure at the Mohawk Valley Police Academy. He was
also an adjunct instructor at Mohawk Valley Community College, where he taught both criminal law and
constitutional criminal procedural law. McNamara currently is an adjunct instructor at Utica College,
where he teaches legal concepts of criminal fraud.
Charles Alexander Morgan III, M.D., is Associate Clinical Professor of Psychiatry, Yale University
School of Medicine. Over the course of twenty years at Yale University and the Neurobiological Studies
Unit of National Center for Posttraumatic Stress Disorder, Dr. Morgan’s neurobiological and forensic
research has established him as an international expert in posttraumatic stress disorder (PTSD), in
eyewitness memory, and in human performance under conditions of high stress. He is a forensic
psychiatrist and has testified as an expert on memory and PTSD at the International Tribunal on War
Crimes, the Hague, Netherlands. Dr. Morgan is subject matter expert in the selection and assessment of
U.S. Military Special Operations and Special Mission Units. His work has provided insight into the
psycho-neurobiology of resilience in elite soldiers and has contributed to the training mission of Army
special programs. For his work in the special operations community Dr. Morgan was awarded the U.S.
Army Award for Patriotic Service in 2008. In 2010 Dr. Morgan was awarded the Sir Henry Welcome
Medal and Prize for his research on enhancing cognitive performance under stress in special operations
personnel. In 2011, Dr. Morgan deployed to Afghanistan as an operational advisor with the Asymmetric
Warfare Group.
Elizabeth A. Phelps, Ph.D., is Silver Professor of Psychology and Neural Science at New York
University. Her research examines the cognitive neuroscience of emotion, learning, and memory. Her
primary focus has been to understand how human learning and memory are changed by emotion and to
investigate the neural systems mediating their interactions. She has approached this topic from a number
of different perspectives, with an aim of achieving a more global understanding of the complex relations
between emotion and memory. As much as possible, Dr. Phelps has tried to let the questions drive the
research, not the techniques or traditional definitions of research areas. Dr. Phelps has used a number of
88
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
techniques (behavioral studies, physiological measurements, brain-lesion studies, fMRI) and has
collaborated with a number of people in other domains (social and clinical psychologists, psychiatrists,
neuroscientists, economists, physicists). Dr. Phelps received a Ph.D. in neuroscience from Princeton
University.
Daniel J. Simons, Ph.D., is a professor in the department of psychology at the University of Illinois,
where he heads the Visual Cognition Laboratory. His research explores the limits of awareness and
memory, the reasons why we often are unaware of those limits, and the implications of such limits for our
personal and professional lives. He is best known for his research that demonstrates how people are far
less aware of their visual surroundings than they think.
Dr. Simons received his B.A. from Carleton College and his Ph.D. in experimental psychology from
Cornell University. He then spent five years on the faculty at Harvard University before being recruited to
Illinois in 2002. He has published more than fifty articles for professional journals, and his work has been
supported by the National Institutes of Health, the National Science Foundation, and the Office of Naval
Research. He is a Fellow and Charter Member of the Association for Psychological Science and an Alfred
P. Sloan Fellow, and he has received many awards for his research and teaching, including the 2003 Early
Career Award from the American Psychological Association. His research adopts methods ranging from
real-world and video-based approaches to computer-based psychophysical techniques, and it includes
basic behavioral measures, survey and individual difference methods, simulator studies, and training
studies. This diversity of approaches helps establish closer links between basic research on the
mechanisms of attention, perception, memory, and awareness and how those mechanisms operate in the
real world.
In addition to his scholarly research, Dr. Simons is the co-author (with Christopher Chabris) of the New
York Times bestselling book, The Invisible Gorilla. He has penned articles for the New York Times, The
Wall Street Journal, the Los Angeles Times, and the Chicago Tribune (among others), and he appears
regularly on radio and television.
Anthony D. Wagner, Ph.D., is a Professor of Psychology and Neuroscience and Co-Director, Center for
Cognitive and Neurobiological Imaging, Stanford University. He is also Director of the Stanford Memory
Laboratory. At Stanford since 2003, Dr. Wagner’s research explores how the brain supports learning,
memory, and executive function. In addition to his basic science, his research examines memory
dysfunction in clinical populations and the role of neuroscience evidence in legal and educational settings.
He is on the faculty in the Psychology Department, and participates in the Neurosciences Program, the
Symbolic Systems Program, the Human Biology Program, and the Stanford Center for Longevity.
Externally, he is a member of the MacArthur Foundation’s Research Network on Law and Neuroscience.
He is a Fellow of the American Association for the Advancement of Science, and a recipient of the
American Psychological Association’s Distinguished Scientific Award for Early Career Contribution,
among other honors. Dr. Wagner received a Ph.D. in psychology from Stanford University in 1997.
Joanne Yaffe, Ph.D., is Professor, College of Social Work, University of Utah and Adjunct Professor of
Psychiatry, College of Medicine, University of Utah. Her scholarly interests are in evidence based
practice and using scientific knowledge for policy and practice decisions. She is particularly interested in
the synthesis of research through systematic reviews and meta-analysis, and, with colleagues in the
United Kingdom, was funded by the Cochrane Collaboration to develop guidelines for reporting
systematic reviews without included studies. She is affiliated with the Social Welfare Coordinating Group
and the Knowledge Translation Group of the Campbell Collaboration and has worked with the Methods
Group of the Cochrane Collaboration. Dr. Yaffe is a member of the International Advisory Group for
CONSORT-SPI, which has developed guidelines for the reporting of randomized trials for complex social
and psychological interventions. Dr. Yaffe received a B.S. in Psychology from University of
89
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Massachusetts, an M.S.W. from the University of Michigan, and a Ph.D. in Social Work and Psychology
from the University of Michigan. She has advanced training in systematic reviews and meta-analysis.
Staff
Anne-Marie Mazza, Ph.D., is the Director of the Committee on Science, Technology, and Law. Dr.
Mazza joined the National Academies in 1995. She has served as Senior Program Officer with both the
Committee on Science, Engineering and Public Policy and the Government-University-Industry Research
Roundtable. In 1999 she was named the first director of the Committee on Science, Technology, and Law,
a newly created activity designed to foster communication and analysis among scientists, engineers, and
members of the legal community. Dr. Mazza has been the study director on numerous Academy reports
including, Reference Manual on Scientific Evidence, 3rd Edition (2011); Review of the Scientific
Approaches Used During the FBI’s Investigation of the 2001 Anthrax Letters (2011); Managing
University Intellectual Property in the Public Interest (2010); Strengthening Forensic Science in the
United States: A Path Forward, (2009); Science and Security in A Post 9/11 World,(2007); Reaping the
Benefits of Genomic and Proteomic Research: Intellectual Property Rights, Innovation, and Public
Health, 2005; and Intentional Human Dosing Studies for EPA Regulatory Purposes: Scientific and
Ethical Issues, 2004. Between October 1999 and October 2000, Dr. Mazza divided her time between The
National Academies and the White House Office of Science and Technology Policy, where she served as
a Senior Policy Analyst responsible for issues associated with a Presidential Review Directive on the
government-university research partnership. Before joining the Academy, Dr. Mazza was a Senior
Consultant with Resource Planning Corporation. She is a fellow of the American Association for the
Advancement of Science. Dr. Mazza was awarded a B.A., M.A., and Ph.D., from The George
Washington University.
Arlene F. Lee, J.D., is the Board Director for the Committee on Law and Justice (CLAJ). Prior to joining
CLAJ, Ms. Lee was the Director of Policy at the Center for the Study of Social Policy, where she focused
on helping federal and state elected officials develop research-informed policies and funding to improve
results for children and families. In this capacity, she oversaw PolicyforResults.org, a leading national
resource for results-based policy. Previously she was the Executive Director of the Maryland Governor’s
Office for Children, where she chaired the Children’s Cabinet and was responsible for the cabinet’s fund
of 60+ million dollars annually. She has served as the Deputy Director of the Georgetown University
Center for Juvenile Justice Reform, Director of the Federal Resource Center for Children of Prisoners and
Youth Strategies Manager for the Governor’s Office of Crime Control and Prevention. Ms. Lee is also the
author of numerous articles and co-authored The Impact of the Adoption and Safe Families Act on
Children of Incarcerated Parents. She has a B.A. in Sociology from Washington College and a J.D. from
Washington College of Law, American University. As a result of her work, Ms. Lee was named one of
Maryland’s Top 100 Women and has received three Governor’s Citations.
Steven Kendall, Ph.D., is Program Officer for the Committee on Science, Technology, and Law. Dr.
Kendall has contributed to numerous Academy reports including the Reference Manual on Scientific
Evidence, 3rd Edition (2011), Review of the Scientific Approaches Used During the FBI’s Investigation of
the 2001 Anthrax Mailings (2011), Managing University Intellectual Property in the Public Interest
(2010); and Strengthening Forensic Science in the United States: A Path Forward (2009). Dr. Kendall
completed his Ph.D. in the Department of the History of Art and Architecture at the University of
California, Santa Barbara, where he wrote a dissertation on 19th century British painting. Dr. Kendall
received his M.A. in Victorian Art and Architecture at the University of London. Prior to joining the
National Research Council in 2007, he worked at the Smithsonian American Art Museum and The
Huntington in San Marino, California.

90
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Karolina Konarzewska is Program Assistant for the Committee on Science, Technology, and Law. Ms.
Konzarzewska received a B.A. in Political Science from the College of Staten Island, City University of
New York and an M.A. in International Relations, New York University. Prior to joining The National
Academies, she worked at various research institutions in Washington, D.C., where she covered political
and economic issues pertaining to Europe, Russia, and Eurasia.

91
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

Appendix B
Committee Meeting Agendas
Meeting 1
Washington, DC
Monday, 2 December 2013
OPEN SESSION
8:00

Continental Breakfast

8:30

Opening Remarks and Introductions
Co-chairs:
Thomas D. Albright, Salk Institute for Biological Studies
Jed S. Rakoff, U.S. District Court for the Southern District of New York

8:45 – 9:30

Charge to the Committee
Speaker: Anne Milgram, Laura and John Arnold Foundation

9:30–11:00

The Science of Memory – A Dynamic Process
Speakers:
Daniel L. Schacter, Harvard University (via videoconference)
John T. Wixted, University of California, San Diego

11:00–11:15

Break

11:15–12:00

Overview of Eyewitness Identification
Speaker: Gary L. Wells, Iowa State University

12:00–1:00

Lunch

1:00–2:30

Meta-Analytical Reviews of System and Estimator Variables
Speakers:
Nancy Steblay, Augsburg College
Christian A. Meissner, Iowa State University
Kenneth Deffenbacher, University of Nebraska at Omaha

2:30–3:00

Strengths and Weaknesses of Eyewitness Research Methodologies
Speaker: Steven Penrod, John Jay College of Criminal Justice

3:00–3:30

General Acceptance of Eyewitness Testimony Research
Speaker: Saul Kassin, John Jay College of Criminal Justice

3:30–3:45

Break

3:45–4:15

Simultaneous and Sequential Lineups
Speaker: Roy S. Malpass, University of Texas at El Paso
93
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
4:15–5:15

Perspectives on Eyewitness Identification
Speakers:
John Firman, International Association of Chiefs of Police
David LaBahn, Association of Prosecuting Attorneys
Kristine Hamann, National District Attorney’s Association
Barry Scheck, The Innocence Project

Tuesday, 3 December 2013
CLOSED SESSION: 8:00–9:15
OPEN SESSION
9:30–10:15

Police Practices
Speakers:
Joseph Salemme, Chicago Police Department
Rob Davis, Police Executive Research Forum

10:15–11:45

Judicial Findings and Recommendations – Including Jury Instructions
Speakers:
The Honorable Robert J. Kane, Supreme Judicial Study Group on
Eyewitness Identification (MA)
The Honorable Geoffrey Gaulkin, Special Master, State v. Henderson (NJ)
The Honorable Paul De Muniz, Oregon Supreme Court
The Honorable Barbara Hervey, Texas Court of Criminal Appeals

11:45–12:30

Research on Jury Instructions
Speakers:
Shari Seidman Diamond, Northwestern University and American Bar Foundation
David V. Yokum, University of Arizona

CLOSED SESSION: 12:30–2:00
Meeting 2
Washington, DC
Thursday, 6 February 2014
OPEN SESSION
8:30–8:45

Opening Remarks and Introductions
Co-chairs:
Tom Albright, Salk Institute for Biological Studies
Jed Rakoff, U.S. District Court for the Southern District of New York

8:45–9:30

The Illinois Pilot Program on Sequential Double-Blind Identification Procedures
Speaker: Sheri Mecklenburg, U.S. Department of Justice

9:30–10:15

Face Recognition and Human Identification
Speaker: P. Jonathon Phillips, National Institute of Standards and Technology

94
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
10:15–10:30

Break

10:30–11:15

Evaluating Eyewitness Research in Court: Moving from General
to Specific Inference
Speaker: John Monahan, University of Virginia

11:15–12:00

Eyewitness Identification from the Perspective of State Attorney Generals
Speaker: Peter Kilmartin, State of Rhode Island

12:00–12:45

Lunch

12:45–1:30

Costs and Benefits of Eyewitness Identification Reforms
Speaker: Steven E. Clark, University of California, Riverside

1:30–2:30

Misinformation and the Creation of False Memories
Speaker: Elizabeth Loftus, University of California, Irvine – via videoconference

2:30–3:15

Obtaining Better Descriptive Information: The Use of the Cognitive Interview
Speaker: Ronald Fisher, Florida International University

CLOSED SESSION: 3:30–5:30
Friday, 7 February 2014
CLOSED SESSION: 8:00–2:00
Meeting 3
Washington, DC
Thursday, 24 April 2014
OPEN SESSION
10:30

Welcome
Co-chairs:
Tom Albright, Salk Institute for Biological Studies
Jed Rakoff, U.S. District Court for the Southern District of New York

10:35–11:30

Photo Arrays in Eyewitness Identification Procedures
Speaker: Karen L. Amendola, Police Foundation

CLOSED SESSION: 11:45–5:00
Friday, 25 April 2014
CLOSED SESSION: 8:30–3:00

95
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

Appendix C
Consideration of Uncertainty in Data on the
Confidence-Accuracy Relationship and the Receiver
Operating Characteristic (ROC) Curve
What has happened is history. What might have happened is science and technology. So what you are
really interested in is what might have happened if you could do it all over again.
John W. Tukey, 18 November 1992, in a
discussion of assessing the uncertainty in cancer
mortality rates at the National Cancer Institute

Both the receiver operating characteristic (ROC) and the confidence-accuracy relationship involve
data (usually, as the proportions of participants in a given study that meet some criterion) and hence are
subject to various sources of uncertainty, including measurement error, random variations from external
conditions, and biases (such as the tendency to respond “conservatively” or “liberally”; see examples of
these biases in Chapter 5). Appendix C focuses on quantification of uncertainty in only some of the errors
caused by measurement and other random sources. Because the confidence-based ROC curve is justified
by an implicit assumption that confidence and accuracy are related, the first section of this appendix
discusses the incorporation of uncertainty when assessing the strength of the confidence-accuracy
relationship, and the second section does the same for the ROC curve. In what follows, HR denotes the hit
rate (or “sensitivity” of a procedure on which the confidence-accuracy relationship or ROC is being
constructed), and FAR (or, 1 –specificity; see Chapter 5) denotes the false alarm rate.1
CONFIDENCE-ACCURACY RELATIONSHIP
When authors talk about the confidence-accuracy relationship, they usually are referring to a
correlation coefficient or to a slope of the line fitted to the points (C, A), where a measure of the
eyewitness’ expressed confidence level C is on the x-axis, and a measure of the witnesses’s accuracy A is
on the y-axis. However one measures the significance of the confidence-accuracy relationship (e.g., in
either a correlation coefficient or a slope of the line fitted to the [C, A] points), it is important to note that
both expressed confidence level (C) and reported accuracy (A) are based on data and thus are subject to
uncertainty, both from random and systematic sources of variation and from biases (see, e.g., Chapter 5
for examples of biases and other variables, such as the implementation of a lineup procedure). In this
appendix, we consider the effects of uncertainty in only “A” and “C” in assessing the strength of the
confidence-accuracy relationship. Ideally, one would repeat the incident multiple times and assess the
error in the repetitions. Usually such repetition is not possible, and one must rely on approximate
measures of uncertainty, as best one can, in the (C, A) points. Approaches for characterizing the
uncertainty in the confidence-accuracy relationship, using data in the published literature, follow.

1

The data cited here are used simply for convenience, as the publications from which they arose provided
sufficient details about them that the illustrations can be generated.

97
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
Consider the following data:2
1) n1 = 44 participants who expressed “Low” confidence (confidence ratings 1,2,3); their overall
accuracy was stated as 61%. Taking the median of these three confidence ratings, C1 = 2 and
A1 = 0.61. The estimated standard error of this proportion is (0.61 · 0.39/44)1/2 = 0.0735.
2) n2 = 203 participants who expressed “Medium” confidence (confidence ratings 4,5,6); their
overall accuracy was stated as 71%. Taking the median of these three confidence ratings, C2 = 5
and A2 = 0.71. The estimated standard error of this proportion is (0.71 · 0.29/203)1/2 = 0.0318.
3) n3 = 326 participants who expressed “High” confidence (confidence rating 7); their overall
accuracy was stated as 85%. Thus, C3 = 7 and A3 = 0.85. The estimated standard error of this
proportion is (0.85 · 0.15/326)1/2 = 0.0198.
A plot of these three data points might suggest a highly convincing relationship between accuracy and
confidence. However, the relationship is not “statistically significant” when assessed via a weighted linear
regression (where weights are inversely proportional to either the standard errors or the variances), nor via
an unweighted Pearson correlation coefficient or a Spearman’s rank correlation coefficient (which
depends less on the assignment of “Low,” “Medium,” and “High” as 2, 5, 7, respectively, than do the
other two methods). Separate tests comparing the proportions 0.85 (“High”) versus either 0.71
(“Medium”) or 0.61 (“Low”) are “statistically significant,” but not the test for comparing the proportions
0.71 (“Medium”) and 0.61 (“Low”). Statistical significance is hard to achieve with only three data points.
Moreover, none of these tests takes into account the potential for error in the self-reported “C” values
(2,5,7), which, as discussed in the previous paragraph, is likely to exist.
Consider a second set of data, reported in Juslin et al.3 In this article, the authors considered two
lineup conditions, denoted as “suspect-similarity” and “culprit-description.”4 The authors correctly note
that the identification rates at each expressed confidence level for these two conditions are very similar;
hence, as the condition had no effect on identification accuracy, one might as well pool “successes/trials”
across the two conditions to reduce the uncertainty in each of the accuracy rates and thus gain greater
power.
Even after combining the two conditions, however, the numbers of trials in the 10 ECL categories
(0.1 = “10% confident,” 0.2 = “20% confident” ... 1.0 = “100% confident”) are not very high (the ten
numbers range from 7 for ECL = 20% to 45 for ECL = 90%). To increase the chances of seeing a
meaningful relationship between confidence and accuracy, the authors pool 0.1 with 0.2, 0.3
with 0.4, 0.5 with 0.6, 0.7 with 0.8, and, 0.9 with 1.0. Although Table 2 in Juslin et al. provides the counts
(numbers of trials), it does not tabulate the accuracies (numbers of correct responses). One can estimate
these accuracies by weighted averages of the displayed percentages shown in the plots in their Figure 25 for
2

These data are cited in H. L. Roediger III, J. T. Wixted, and K. A. DeSoto. “The Curious Complexity between
Confidence and Accuracy in Reports from Memory” in Memory and Law, ed. L. Nadel and W. P. SinnottArmstrong (Oxford: Oxford University Press, 2012), p. 109, who in turn cite Odinot et al. [G. Odinot, G.Wolters,
and P.J. van Koppen, “Eyewitness memory of a supermarket robbery: A case study of accuracy and confidence after
3 months, Law and Human Behavior 33: 506–514 (2009)] as the source of these data, from nine “central witnesses”
(the other 5 witnesses were not interviewed by the police). The sample sizes (44, 203, 326) apparently arise from
having “averaged across different categories (person descriptions, object descriptions, and action details) for the
nine central witnesses interviewed in that study”; see J. T. Wixted et al., “Confidence Judgments Are Useful in
Eyewitness Identifications: A New Perspective,” submitted to Applied Psychology 2014, p. 17.
3
Juslin et al., “Calibration and Diagnosticity of Confidence in Eyewitness Identification: Comments on What
Can Be Inferred From the Low Confidence-Accuracy Correlation,” Journal of Experimental Psychology: Learning,
Memory, and Cognition 22(5): 1304-1316 (September 1996).
4
See J. T. Wixted et al., “Confidence Judgments Are Useful in Eyewitness Identifications: A New Perspective,”
submitted to Applied Psychology 2014, p. 13.
5
See pages 1310-1311 of Juslin et al. for the data in their Table 1 and Figure 2, respectively.

98
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
the “suspect-similarity condition” (“A” = 0.27, 0.38, 0.51, 0.55, 0.87; n = 15, 21, 25, 29, 51) and for the
“culprit- description condition” (“A” = 0.18, 0.66, 0.63, 0.90, 0.91; n = 10, 18, 28, 41, 37). In the confidence
level categories (15%, 35%, 55%, 75%, 95%), the accuracies (with their standard errors and the total sample
sizes on which they are based following them in parentheses) are, respectively, 23.4% (8.5%, n=25), 50.9%
(8.0%, n=39), 52.6% (6.9%, n=53), 75.5% (5.1%, n=70), and 88.7% (3.4%, n=88). For these data, both the
unweighted correlation coefficient, 0.9766 (t-statistic = 7.865, p-value 0.004) and the slope of the weighted
linear regression (points weighted inversely proportional to their standard errors), 0.773 (standard error
0.085, p-value 0.003), are statistically significant, in that such convincing data of a relationship between
correlation and accuracy would be unlikely to arise if, in fact, no association existed.
Another method for assessing the significance of the unweighted correlation is through the simulation
of a large number of trials on the basis of the data that were observed. For each trial, one can first
simulate five confidence values, uniformly distributed between the endpoints that were observed: c1 is
uniformly distributed between (0.05, 0.25) (mean is the observed 0.15); c2 is uniformly distributed
between (0.25, 0.45) (mean is the observed 0.35); ... c5 is uniformly distributed between (0.85, 1.00).
Next, one simulates five proportions using the observed conditions: a1 is a binomial variate (n = 25, p =
0.234) divided by n = 25; a2 is a binomial variate (n = 39, p = 0.509) divided by n = 39; ... a5 is a binomial
variate (n = 88, p = 0.887) divided by n = 88. For each trial with five simulated c values and their five
corresponding a values, one calculates a Pearson correlation coefficient. Figure C-1 shows a plot of the
five data points, with limits of one standard error on the estimated accuracies (left panel) and the
histogram of the 1,000 simulated Pearson correlation coefficients (right panel). The median is 0.9534
(close to the observed 0.9766), the upper and lower quartiles are 0.916 and 0.977, and the central 90% of
the 1,000 values lie between 0.8650 and 0.993. Thus, an approximate 90% confidence interval for the true
correlation coefficient, (0.865, 0.993) definitely does not include zero, a further indication of the
significance of the Pearson correlation coefficient.
The example above illustrates the importance of incorporating known uncertainty in the estimated
accuracy for the confidence level category. The relationship between confidence and accuracy should take
into account (1) the repeated responses of a limited number of “eyewitnesses” in the study, and (2) the
uncertainty in an eyewitness’ “expressed confidence level.” The 2009 National Research Council report,
Strengthening Forensic Science in the United States: A Path Forward, cited studies in which fingerprint
examiners reached different conclusions when presented with exactly the same evidence at a later time.6
Quite possibly, in many of these laboratory studies on which these confidence-accuracy relationships are
based, participants may express different levels of confidence if presented with exactly the same set of
circumstances and procedures six months later.
The existing literature ranges on its assessment of the significance of the confidence-accuracy
relationship, with some articles suggesting a very strong relationship and many others suggesting that it is
weak or nonexistent. The lack of significance in the confidence-accuracy relationship may result from
other factors not taken into account. For example, Smalarz and Wells7 suggest that restricting the plot to
only those data corresponding to “choosers” may strengthen the relationship. Other factors that might
affect the relationship include the presence or absence of weapon, the level of stress during the incident,
and the length of exposure to the perpetrator. Roediger and colleagues state that 8

6

National Research Council, Strengthening Forensic Science in the United States: A Path Forward
(Washington, D.C.: The National Academies Press, 2009), p. 139.
7
L. Smalarz and G. L. Wells, “Eyewitness certainty as a system variable,” in Reform of Eyewitness
Identification Procedures, ed. B. L. Cutler (Washington D.C.: American Psychological Association, 2013), 161–77.
8
H. L. Roediger III, J. T. Wixted, and K. A. DeSoto, “The curious complexity between confidence and accuracy
in reports from memory,” in Memory and Law, ed. L. Nadel and W. Sinnott-Armstrong (Oxford: Oxford University
Press, 2012), 84–118.

99
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
the simple assumption usually made that confidence and accuracy are always tightly linked is
wrong…the relation between confidence and accuracy depends on the method of analysis, on the
target material being remembered, on who is doing the remembering, and (in situations where
memory is tested by recognition) on the nature of the lures and distractors. In addition, there is more
than one way to measure the relationship between confidence and accuracy, and not every way is
equally relevant to what courts of law would like to know about the issue.
Studies that incorporate numerous variables, as well as soliciting a confidence statement at various times
(e.g., immediately, or 10 minutes after the incident, or 1 hour after the incident), need to be conducted.
RECEIVER OPERATING CHARACTERISTIC ANALYSIS
A receiver operator characteristic (ROC) is a reliable, time-honored assessment of test performance.
ROC has been used for decades in the medical test diagnostic literature. As noted in Chapter 5,
conventionally two procedures were compared using a single diagnosticity ratio: DR = HR/FAR = hit
rate/false alarm rate, or sensitivity / (1 – specificity). Wixted and colleagues9 observed that the
diagnosticity ratio, DR, can vary depending on an eyewitness’ expressed confidence level (ECL) and
hence proposed the use of an ECL-based ROC curve to compare two lineup procedures (simultaneous
versus sequential).
The ECL-based ROC curve for a given procedure (e.g., simultaneous) is constructed as follows:
1) Collect participants in a study and subject them to the experimental conditions.
2) For each participant, record whether she or he accurately selected the correct suspect or
accurately passed over the filler, and the expressed confidence level in the decision.
3) Collect all the responses for participants who answered “100% confident” (say, n1 of them) and
record the combined FAR (false alarm rate, or 1 – specificity) and HR (hit rate, or sensitivity)
across n1 participants, (FAR1, HR1).
4) Repeat step 3 for all participants who answered “90% confident” (or higher; say, n0.9 of them),
resulting in the data pair (FAR0.9, HR0.9).
5) Repeat step 3 for all participants who answered “80% confident” (or higher; say, n0.8), resulting in
the data pair (FAR0.8, HR0.8).
6) Continue to repeat step 3 for the groups of participants who answered “70% confident” ... “10%
confident” (or higher; say, n0.7...n0.1 of them).
7) Plot the ten data pairs, (FAR1, HR1), ..., (FAR0.1, HR0.1).
This plot results in the ROC curve, whose points (HR, FAR) correspond to different ECLs.
The plotted points usually are connected by straight lines, and the slope of the ROC curve at each of
those plotted points represents the DR corresponding to that confidence category. The ROC curve
illustrates the separate DRs rather than calculating a single DR collapsed across all confidence categories.
As with the confidence-accuracy relationship, it is important to recognize the uncertainty in the estimated
(FAR, HR) data points. How does the uncertainty in FAR and HR, and hence in the diagnosticity ratio (DR
= HR/FAR), translate into uncertainty into the ROC curve?

9

L. Mickes, H. D. Flowe, and J. T. Wixted, “Receiver operating characteristic analysis of eyewitness memory:
Comparing the diagnostic accuracy of simultaneous and sequential lineups,” Journal of Experimental Psychology:
Applied 18, 361–376 (2012). See especially pp. 362–65 for a description of ROC analysis in the medical literature
and applied to the eyewitness identifications.

100
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
The effect of uncertainty in estimates of HR, FAR, DR (= HR/FAR) on the ROC curve can be seen by
simulating new HR and FAR rates, assuming that the given HR and FAR rates are true “means” from the
simulated distributions. As a first example, consider the set of data from Brewer and Wells10 which is
cited by Mickes et al. in their Table 1 (p. 367). The data are: HR = (.090,.237,.320,.355,.370); FAR
=(.002,.015,.030,.038,.041), leading to five diagnosticity ratios (rounded) DR = ( 45,16,11,9,9 ). The
article states that the experiment involved 1200 participants.
As above, one can simulate each of the five hit rates and the five false alarm rates, with 4,000
independent trials and 1,200 participants, in such a way that the means of the five distributions of hit rates
(HRs) and the means of the five distributions of false alarm rates (FARs) equal the values observed in the
experiment (e.g., (0.090, 0.237, 0.320, 0.355, 0.370) for HR and (0.002, 0.015, 0.030, 0.038, 0.041) for
FAR), leading to five distributions of 4000 diagnosticity ratios (HR/FAR). For example, consider
simulating 1,200 individuals whose HR is 0.090 = 9.0%. One expects that, on average, about (9%) ×
1,200 = 108 of the simulated 1,200 participants will have “hits.” When repeating this trial of 1,200
individuals, the number might be 110, or 95, or some other number around, but usually not exactly, 108.
Repeating the trial 4,000 times, one can average the 4,000 numbers (e.g., 108, 110, 95…), yielding a
mean simulated HR. The advantage is that one can also use the 4,000 numbers to calculate a standard
deviation.11 One repeats exactly the same exercise for the five FAR rates, yielding a mean FAR and a
standard deviation, SDFAR. As noted in Chapter 5, in real life, HR and FAR will be estimated on the same
set of 1,200 participants, so the two numbers, HR and FAR, in the five (HR, FAR) pairs, will be
correlated. In the simulation, HR and FAR are independent, so the estimated uncertainties are likely to be
optimistic; the real uncertainties could well be larger. One can then plot three sets of points (each set
contains five points): (1) (mean HR, mean FAR) (this plot should look qualitatively similar to the one in
Figure 6(A) in Mickes et al. (2012) (p.371); (2) (mean HR − SDHR, mean FAR − SDFAR) (these points
should lie somewhat below the points plotted in (1)); and (3) (mean HR + SDHR, mean FAR + SDFAR)
(these points should lie somewhat above the points plotted in (1)).
Figure C-2 shows bands of one standard error in both HR and FAR, illustrating one source of
uncertainty in estimating these two quantities. The same approach to calculating uncertainties was used
for the two sets of (HR, FAR) values given by the “simultaneous” and “sequential” data in Mickes et al
(2012) Table 3 (p. 372). The text indicates that Experiment 1A used n = 598 participants, so the
simulation assumed n = 600. In Figure C-3, “M” refers to “siMultaneous” (solid, black), and “Q” refers to
“seQuential” (dash, red). Note that the “M” and “Q” points look roughly in the same pattern as in their
Figure 6A (p. 371). Note the substantial overlap in the bands of “one standard deviation” surrounding
each of the data points, indicating no “statistically significant” differences between the “M”
(simultaneous) and “Q” (sequential) points.12 If one were to take into account the effects of using the
same eyewitness in the same study with different responses to different tasks, the variability would be
even larger.
When the same exercise for the data in Experiment 2 (n=631) is repeated, similarly ambiguous results
(see Figure C-4) are obtained. As Mickes and colleagues suggest, the differences between simultaneous
and sequential are even less impressive, and especially so once bands of one standard errors around the
points are shown.

10

N. Brewer and G. L. Wells, “The confidence-accuracy relationship in eyewitness identification: Effects of
lineup instructions, foil similarity, and target-absent base rates,” Journal of Experimental Psychology: Applied
12(1): 11–30 (2012) (as cited by Mickes et al., Table 1, p. 367).
11
Or Standard Deviation Hit Rate (SDHR) which also can be obtained from standard formulas for the standard
deviation of the binomial distribution. See G. Snedecor and W. Cochran, Statistical Methods, Sixth Ed. (Ames,
Iowa: Iowa State University Press, 1967).
12
The bands of two standard deviations would overlap even more.

101
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
These further analyses on these published data sets suggest the following conclusions.
1) The strength of the confidence-accuracy relationship involves uncertainty in the measures of both
A (accuracy) and C (confidence), as well as other factors that can influence the relationship.
2) An ROC curve incorporates more information than a single DR (diagnosticity ratio = HR/FAR)
using a third variable (different test thresholds in the medical literature; in the present context,
different expressed confidence levels (ECLs); i.e., HR and FAR at different expressed confidence
levels). As is true with any data, the data from which an ROC is constructed (FARs, HRs,
expressed confidence levels) have uncertainty, and that uncertainty is passed on to the ROC. A
naive comparison of two ROCs without recognizing that uncertainty can be misleading. As with
any tool, one must be careful in how one draws inferences when comparing ROC curves.
3) Other methods for comparing two procedures (in which the outcome is a binary classification
such as “identification” / “no identification” of an individual) exist in other literature.13
These analyses considered only the most obvious form of random measurement error. The ROC may
be influenced by other sources of bias; these sources are not considered or displayed in the plots shown
here (see Chapter 5). Also, the ROC curve takes into consideration only the probability that an eyewitness
who makes a positive identification of a suspect has correctly identified the true culprit (positive
predictive value); it does not take into consideration the rule-out probability that an eyewitness who fails
to make an identification of a suspect has correctly recognized that the suspect is not the true culprit
(negative predictive value) (see Chapter 5).

FIGURE C-1 Data Inferred from Juslin et al 14
13

See, e.g., T. Hastie, R. Tibshirani, and J.H. Friedman, The Elements of Statistical Learning: Data Mining,
Inference, and Prediction (New York: Springer, 2009) for a discussion on classification and evaluation methods of
statistical machine learning research.
14
Adapted from Juslin et al., “Calibration and Diagnosticity of Confidence in Eyewitness Identification:
Comments on What Can Be Inferred From the Low Confidence-Accuracy Correlation,” Journal of Experimental
Psychology: Learning, Memory, and Cognition 22(5): 1304-1316 (September 1996). The left panel plots
confidence-accuracy data from p. 1311. Data are pooled into five categories; accuracies are inferred from p. 1313.
Data are shown with limits of one standard error and weighted least squares regression line. The right panel is a
histogram of 1000 simulated Pearson correlation coefficients, using data from 5 categories shown in right panel. The

102
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

FIGURE C-2 Data from Brewer and Wells15

FIGURE C-3 Data from Experiment 1A in Mickes, Flowe, and Wixted 16

central 90% of the simulated values lie between 0.853 and 0.993, indicating that the true unweighted Pearson
correlation coefficient is significantly different from zero. Courtesy of Karen Kafadar.
15
Adapted from N. Brewer and G. L. Wells, “The confidence-accuracy relationship in eyewitness identification:
Effects of lineup instructions, foil similarity, and target-absent base rates,” Journal of Experimental Psychology:
Applied 12(1): 11–30 (2006). This data was cited by L. Mickes, H. D. Flowe, and J. T. Wixted, “Receiver
Operating Characteristic Analysis of Eyewitness Memory: Comparing the Diagnostic Accuracy of Simultaneous
Versus Sequential Lineups,” Journal of Experimental Psychology-Applied 18(4): 361–376 (2012). Courtesy of
Karen Kafadar.

103
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS

FIGURE C-4 Data from Experiment 2 in Mickes, Flowe, and Wixted 17

ALTERNATIVE ANALYSIS TO CONFIDENCE-BASED ROC
FOR COMPARING PROCEDURES
As noted in Chapter 5, the diagnosticity ratio [hit rate / false alarm rate = HR/FAR = sensitivity/(1 –
specificity)] can depend not only on an eyewitness’ tendency toward “conservative” or “liberal”
identification (as measured by expressed confidence level), but also on numerous other factors, including:
(1) lineup procedure (e.g., two levels: simultaneous versus sequential); (2) presence or absence of a weapon
(two levels; more levels could be considered, such as gun, knife, towel, none); (3) stress (e.g., three levels:
high, medium, low); (4) elapsed time between incident and exam (e.g., three levels: 30 min, 2 hours, 1 day);
(5) race difference (e.g., two levels: same or different race; or four levels: eye-witness/culprit = white/white;
white/non-white; non-white/white; non-white/non-white; non-white/white); (6) participant (e.g., N levels,
corresponding to N participants).
If a study is sufficiently large, one could develop a performance metric for each participant in the study
corresponding to each of these conditions. For example, one could construct an ROC curve and calculate as
the performance metric the the logarithm of the area under the curve, or log(AUC), for each person and each
condition in the study. One could also use as a performance metric the logarithm of the odds (log odds) of a
correct decision; e.g., log(HR/(1-HR)) or log((1-FAR)/FAR).
Consider the following approach:
Let yijklmnr denote the log(AUC) or a log odds (or another performance metric) for the rth trial using
participant n (n = 1, ...,N) for procedure i, weapon level j, stress level k, time condition ℓ, and crossrace effect m.18 One could write:
16

Adapted from L. Mickes, H. D. Flowe, and J. T. Wixted, “Receiver Operating Characteristic Analysis of
Eyewitness Memory: Comparing the Diagnostic Accuracy of Simultaneous Versus Sequential Lineups,” Journal of
Experimental Psychology-Applied 18(4): 361–376 (2012). Courtesy of Karen Kafadar.
17
Adapted from L. Mickes, H. D. Flowe, and J. T. Wixted, “Receiver Operating Characteristic Analysis of
Eyewitness Memory: Comparing the Diagnostic Accuracy of Simultaneous Versus Sequential Lineups,” Journal of
Experimental Psychology-Applied 18(4): 361–376 (2012). Courtesy of Karen Kafadar.

104
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
yijklmnr = μ + αi + βj + γk + δl + φm + (αβ)ij + ...(interactions)... + ϵijklmnr
where μrepresents the overall average log(AUC) or log odds across all conditions, the next six terms
reflect the main effects of A (lineup procedure: i = 1 for sequential and i = 2 for simultaneous); B
(weapon: j = 1 for presence and j = 2 for absence of weapon); C (stress level: k = 1 for low, k = 2 for
medium, k = 3 for high); D (elapsed time between incident and report: ℓ = 1 for 30 minutes, ℓ = 2 for 2
hours, ℓ = 3 for 1 day); E (cross-race effect: m = 1 for same race and m = 2 for different races); F
(participant effect: n = 1, 2, ...,N participants); “(interactions)” reflects the joint effect of two or more
factors together; and the last term, ϵijklmnr represents any random error in the rth trial that is not specified
from the previous terms (e.g., measurement, ”ECL,” multiple trials). This approach would allow one to
separate the effects of the different factors, to assess which factors have the greatest influence on the
outcome (here, logarithm of the area under the ROC curve: bigger is better), and to evaluate the
importance of these factors relative to variation among “eyewitnesses.” It may be that eyewitnesses are
the greatest source of variability, dominating the effects of all other factors. Or it may be that, in spite of
person-to-person variability, one or more factors still stand out as having strong influence on the outcome.
Note that (1) other covariates could be included, such as age and gender of participant; and (2) the ROC
curve need not be defined in terms of expressed confidence level thresholds, if a more sensitive measure
of response bias (tendency toward “liberal” versus “conservative” identifications) can be developed.
For example, C.A. Carlson and M.A. Carlson19 use partial area under the curve, or pAUC, as a
summary measure of the information in an ROC curve (bigger is better), for each of twelve different
conditions defined by three factors: (1) Procedure, three levels: simultaneous (SIM: suspect in position 4),
sequential (SEQ2: suspect in position 2), sequential (SEQ5: suspect in position 5); (2) Weapon focus, two
levels: present versus absent; (3) Distinctive feature, two levels: present versus absent. The data are
provided in their Table 3, along with 95% confidence intervals.20 Because the length of a confidence
interval is proportional to the standard error, pAUC values with shorter confidence intervals correspond to
smaller standard errors and hence should have higher weights. The logarithms of the reported pAUC
values and weights (reciprocals of the lengths of the reported confidence intervals) are given below in
Table C-1.
For the Carlson study, the data on all N = 2,675 participants (720 undergraduates and 1,955
SurveyMonkey respondents) were combined, and expressed confidence levels were solicited on a 7-point
scale. Variations in the twelve log(pAUC) values can be decomposed into three main effects (one each for
procedure, weapon, and feature), and their two-way interactions. (The raw data may permit a more
detailed analysis.) The data can be analyzed using a less complex model than that stated above (because
the model has fewer terms):
yijk = μ + αi + βj + γk + (αβ)ij + (αγ)ik + (βγ)jk + ϵijk
where yijk denotes (5 + log(pAUC)) for procedure i (i = 1, 2, 3), weapon condition j (j = 1, 2), and feature
k (k = 1, 2); μ represents the overall average log(pAUC) across all conditions; αi represents the effect of
procedure i; βj represents the effect of weapon condition j; γk represents the effect of feature condition k;
and the next three terms reflect the three two-factor interactions between the main factors. The analysis of
18

When the performance metric is a log odds, this model is known as logistic regression; see, e.g., F. Harrell,
Regression Modeling Strategies (New York: Springer-Verlag, 2001). A model where the performance metric is
log(AUC) was studied by F. Wang and C. Gatsonis. See F. Wang and C. Gatsonis, “Hierarchical models for ROC
curve summary measures: Design and analysis of multi-reader, multi-modality studies of medical tests,” Statistics in
Medicine 27: 243-256 (2008).
19
C. A. Carlson and M. A. Carlson, “An evaluation of lineup presentation, weapon presence, and a distinctive
feature using ROC analysis,” Journal of Applied Research in Memory and Cognition 3(2): 45–53 (2014).
20
Carlson and Carlson , p. 49.

105
Copyright © National Academy of Sciences. All rights reserved.

Identifying the Culprit: Assessing Eyewitness Identification

PREPUBLICATON COPY – UNEDITED PROOFS
variance, where log(pAUC) values are weighted according to the values in the last column of Table C-1,
is given in Table C-2 below. None of the factors is significant.21 It must be stressed that the complete set
of raw data may yield a more powerful analysis with different results, as might a different summary
measure of the ROC curve, such as AUC, or area under the ROC curve.22
TABLE C-1 Conditions and Logarithms of Reported pAUC Valuesa
Condition
1

Procedure
SIM

Weapon
Yes

Feature
Yes

5 + log(pAUC)
1.31112

Weight
47.6

2

SIM

Yes

No

1.72983

33.3

3

SIM

No

Yes

0.92546

55.6

4

SIM

No

No

1.87643

45.5

5

Seq2

Yes

Yes

1.49344

47.6

6

Seq2

Yes

No

1.22774

47.6

7

Seq2

No

Yes

1.08798

52.6

8

Seq2

No

No

1.58875

41.7

9

Seq5

Yes

Yes

1.70316

38.5

10

Seq5

Yes

No

0.98262

58.8

11

Seq5

No

Yes

0.65719

66.7

12
Seq5
No
No
1.49344
55.6
a
Adapted from data on pAUC from Table 3 in C. A. Carlson and M. A. Carlson. 2014. “An evaluation of lineup
presentation, weapon presence, and a distinctive feature using ROC analysis,” Journal of Applied Research in
Memory and Cognition 3(2): 45–53. The addition of “5” to “log(pAUC) is simply to avoid negative numbers; the
inferences from the analysis remain unchanged. Courtesy of Karen Kafadar.

TABLE C-2 Analysis of Variance Table for log(pAUC)a
Source of Variation
Procedure

Degrees of Freedom
2

Sum of Squares
8.04

Mean Square
4.02

F-statistic
1.129

p-value
0.470

Weapon

1

2.94

2.94

0.826

0.460

Feature

1

14.72

14.72

4.138

0.179

Procedure×Weapon

2

0.59

0.30

0.083

0.923

Procedure×Feature

2

10.41

5.21

1.463

0.406

Weapon ×Feature

1

34.80

34.80

9.780

0.089

Residuals

2
7.12
3.56
Adapted from data on pAUC from Table 3 in C. A. Carlson and M. A. Carlson. 2014. “An evaluation of lineup
presentation, weapon presence, and a distinctive feature using ROC analysis,” Journal of Applied Research in
Memory and Cognition 3(2): 45–53. Courtesy of Karen Kafadar.
a

21

We can decompose the two degrees of freedom in the sum of squares for Procedure (three levels), 8.04, into
two single degree of freedom contrasts, seq2 versus seq5 (4.14), and sim versus the average of seq2 and seq5 (3.90),
and consider all pairwise interaction terms among the four “main effects.” All single degree-of-freedom effects
remain non-significant, in either this weighted analysis or in an unweighted analysis.
22
For a discussion of the advantages and disadvantages of using AUC versus pAUC as a summary measure, see
S. D. Walter, “The partial area under the summary ROC curve,” Statistics in Medicine 24(13): 2025–2040 (July
2005).

106
Copyright © National Academy of Sciences. All rights reserved.