Skip navigation

Ca State Auditor the Report on Alternative Sanctions Profiling Program 2011

Download original document:
Brief thumbnail
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
Department of Corrections
and Rehabilitation
The Benefits of Its Correctional Offender
Management Profiling for Alternative Sanctions
Program Are Uncertain
September 2011 Report 2010-124

Independent NONPARTISAN
TRANSPARENT Accountability

The first five copies of each California State Auditor report are free. Additional copies are $3 each, payable by
check or money order. You can obtain reports by contacting the Bureau of State Audits at the following address:
California State Auditor
Bureau of State Audits
555 Capitol Mall, Suite 300
Sacramento, California 95814
916.445.0255 or TTY 916.445.0033
OR
This report is also available on the World Wide Web http://www.bsa.ca.gov
The California State Auditor is pleased to announce the availability of an on-line subscription service. For
information on how to subscribe, please contact the Information Technology Unit at 916.445.0255, ext. 456,
or visit our Web site at www.bsa.ca.gov.
Alternate format reports available upon request.
Permission is granted to reproduce reports.
For questions regarding the contents of this report,
please contact Margarita Fernández, Chief of Public Affairs, at 916.445.0255.

Elaine M. Howle
State Auditor

CALIFORNIA STATE AUDITOR

Doug Cordiner
Chief Deputy

Bureau of State Audits

555 Capitol Mall, Suite 300

S a c r a m e n t o, C A 9 5 8 1 4

September 6, 2011	

916.445.0255

916.327.0019 fax

w w w. b s a . c a . g o v

2010-124

The Governor of California
President pro Tempore of the Senate
Speaker of the Assembly
State Capitol
Sacramento, California 95814
Dear Governor and Legislative Leaders:
As requested by the Joint Legislative Audit Committee, the California State Auditor presents this
audit report concerning the Department of Corrections and Rehabilitations’ (Corrections) use of
the Correctional Offender Management Profiling for Alternative Sanctions (COMPAS), which is a
software tool that helps to identify the characteristics that cause offenders to commit crimes so they
can participate in rehabilitative programs and thereby lessen their likelihood of reoffending.
Our report concludes that the benefits from Corrections’ use of COMPAS are, at best, uncertain.
Specifically, Corrections’ use of COMPAS in its reception centers—facilities where inmates entering
the correctional system are evaluated and assigned to a prison—does not meaningfully affect its
decision‑making concerning prison assignments, and by extension, the rehabilitative programs
inmates might access at those facilities. Our discussions with staff from eight of Corrections’
12 reception centers revealed that other non-COMPAS factors, such as an inmate’s security level and
limited bed space at receiving prisons, play more prominent roles in determining where inmates can
be housed. Furthermore, the COMPAS core assessment identifies up to five different needs; however,
Corrections has rehabilitative programs that address only two. Corrections has not established
regulations defining how COMPAS assessments are to be used despite legal requirements to do so.
As a result, we recommend that Corrections suspend its use of COMPAS until it issues regulations
addressing these issues and develops a methodology for measuring whether COMPAS is achieving
its intended outcome, helping Corrections reduce the State’s recidivism rate.
Our review also revealed other problems with Corrections’ deployment of COMPAS that negatively
affect its usefulness. Some correctional staff we spoke with at reception centers and parole offices
indicated a lack of acceptance of COMPAS, suggesting the need for further training or clarification
regarding COMPAS’s value. Further, Corrections use of COMPAS for placing inmates into its in‑prison
rehabilitative programs is limited to its substance abuse program. However, we found that many in this
program either lack COMPAS assessments or have a low COMPAS-identified need for substance
abuse treatment. Moreover, relatively few inmates with moderate to high substance abuse treatment
needs, as determined through the COMPAS core assessment, are assigned to a treatment program.
Finally, we found that Corrections lacks accounting records demonstrating how much it cost to fully
deploy and implement COMPAS at its reception centers, prisons, and parole offices. The lack of such
information prevents Corrections from demonstrating accountability for its spending on COMPAS.
Respectfully submitted,

ELAINE M. HOWLE, CPA
State Auditor

California State Auditor Report 2010-124

September 2011

Contents
Summary	

1

Introduction	

7

Audit Results
The Benefits of the Correctional Offender Management Profiling for
Alternative Sanctions Program Have Yet to Be Realized	

17

Corrections Cannot Demonstrate How Much It Cost to Fully Deploy
and Administer COMPAS 	

38

Recommendations	

42

Response to the Audit
Department of Corrections and Rehabilitation	

45

California State Auditor’s Comments on the Response From the 	
Department of Corrections and Rehabilitation	

49

vii

viii

California State Auditor Report 2010-124

September 2011

Blank page inserted for reproduction purposes only.

California State Auditor Report 2010-124

September 2011

Summary
Results in Brief

Audit Highlights . . .

The Department of Corrections and Rehabilitation (Corrections)
intends to use the Correctional Offender Management Profiling for
Alternative Sanctions (COMPAS) software to help identify factors
that cause inmates to commit crimes, so they can participate in
such rehabilitative programs as substance abuse treatment or
vocational education to reduce their likelihood of reoffending,
thereby reducing overcrowding in the State’s prisons. California’s
high recidivism rates and difficulties with prison overcrowding are
well documented. In its October 2010 outcome evaluation report,
Corrections reported that 67.5 percent of all felons released during
fiscal year 2005–06 returned to prison within three years. Further,
in May 2011 the U.S. Supreme Court issued a ruling upholding
the authority of a lower court to require that California reduce
its inmate population to 137.5 percent of the design capacity of its
correctional institutions. As of June 30, 2011, Corrections had more
than 144,000 inmates in its various institutions, which were designed
to accommodate only 80,000.

Our review of the Department of Corrections
and Rehabilitation’s (Corrections) use
of Correctional Offender Management
Profiling for Alternative Sanctions (COMPAS)
highlighted the following:

However, the prospects that COMPAS will play a meaningful role
in helping Corrections ultimately reduce prison overcrowding and
lower its recidivism rates are, at best, uncertain. Corrections uses
gender-specific versions of two different COMPAS assessments.
The COMPAS core assessment identifies the needs of inmates
entering the prison system, while the COMPAS reentry assessment
evaluates inmates who are about to reenter society on parole.
Our review found Corrections’ use of COMPAS during its parole
planning process is not consistently enforced, while its use in
reception centers—where inmates are initially evaluated and
assigned to a prison—does not appear to affect decisions on
prison assignments and, by extension, the rehabilitative programs
inmates might access at those facilities.
Corrections’ process at its 12 reception centers for assigning
inmates to prisons is complex and considers factors such as
an inmate’s history of violence, medical needs, gang affiliations,
and the available bed space at suitable facilities that can
accommodate the inmate’s security requirements. Our observations
at one reception center and discussions with Corrections’ staff at
seven others revealed that prison assignments are often not
based on COMPAS. Instead, the inmate’s security level and the
weekly placement restrictions imposed by Corrections’ Population
Management Unit—the unit responsible for coordinating inmate
movement within the prison system—are the primary determinants
of prison assignment.

»» It is uncertain if COMPAS will help
Corrections reduce prison overcrowding
and lower recidivism rates.
• 	 Corrections’ use of COMPAS during its
parole planning process is inconsistent.
• 	 Prison assignments made by its
reception centers are often not based
on COMPAS.
»» According to some staff at the reception
centers, there is a lack of buy-in on
COMPAS across the institutions and the
tool does not seem beneficial.
»» Corrections has not issued regulations
on COMPAS as required nor has it provided
training to some staff on how to use the
assessment tool.
»» The value of using COMPAS to assess an
inmate’s needs is limited—COMPAS core
assessment identifies up to five different
needs; however, Corrections has
rehabilitative programs that only
address two.
»» Since Corrections instructed staff to use
COMPAS only when placing inmates in its
substance abuse treatment program, we
question whether COMPAS is a valuable
tool for identifying inmates’ needs.
»» Corrections has not tracked the actual
costs of deploying COMPAS and cannot
explain how it monitored the project for
any cost overruns.

1

2

California State Auditor Report 2010-124

September 2011

Our discussions with staff at the reception centers also suggested
that some do not see value in using COMPAS. For example,
one classification staff representative—who is responsible for the
final decision on where an inmate will be sent—told us that she
rotates to different prisons each week and that there is a lack of
buy-in on COMPAS across the institutions, noting that the tool
does not seem beneficial. The classification staff representative
also indicated Corrections has not trained staff in her role on how
to use COMPAS. Corrections’ project manager for COMPAS
confirmed that these individuals have not been trained, explaining
that reception center staff are still in the process of developing
procedures for using COMPAS. According to the project
manager, this lack of training would preclude classification staff
representatives from consistently considering COMPAS results.
An assistant warden at another reception center questioned whether
COMPAS identified inmate needs that his staff had not already
gleaned from inmate files. Furthermore, we noticed that Corrections
has established underground regulations concerning COMPAS’s use
because they were not adopted in accordance with the California
Administrative Procedures Act, nor has it discussed COMPAS in
its operating manual.
The value of using COMPAS to assess an inmate’s needs is
also limited because few programs exist within prisons to treat
those needs. The COMPAS core assessment identifies up to
five different needs; however, Corrections has rehabilitative
programs that address only two—its academic/vocational education
and substance abuse treatment programs. It has no programs to
address criminal thinking, anger and violence, or family criminality.
Corrections’ secretary acknowledged that limited rehabilitative
programs could cause some to question the value of using COMPAS
but maintained that doing so was still worthwhile, since it avoids
guessing which inmates need what services and potentially wasting
limited resources on the wrong population. However, the secretary’s
comments contrast with the fact that more inmates with COMPAS
core assessments have substance abuse problems when compared to
the number of those with academic/vocational needs, yet the inmate
capacity for academic/vocational education programs is five times
that of substance abuse treatment programs. Considering that
Corrections has instructed its staff to consider COMPAS assessment
results only when placing inmates in its substance abuse treatment
program, it is questionable whether COMPAS is a valuable tool for
determining what rehabilitative programs Corrections needs.
Corrections has also cited budget cuts from fiscal year 2009–10
as a main reason for limited rehabilitative programs. In the
2009 Budget Act, the Legislature required Corrections to cut an
unspecified amount from its rehabilitative programs and submit a
report on its plan to absorb these cuts while maintaining effective

California State Auditor Report 2010-124

September 2011

rehabilitative programs. Its progress report to the Legislature in
April 2010 indicated that it shortened the duration of its substance
abuse treatment program, which had been from 12 to 36 months,
to three months and reduced the number of prisons at which the
program is offered. Corrections also reported it developed new
education models that prioritize inmate placement based on their
risk of reoffending and the time remaining until their release. These
budget cuts caused spending on rehabilitative programs to be
reduced by $69.1 million during fiscal year 2010–11.
Our review also raised questions as to whether Corrections was
actually putting inmates in the correct rehabilitative programs
based on COMPAS. All 11 of its institutions with a substance
abuse treatment program treat only a limited number of inmates
with moderate to high needs as determined by COMPAS. In
February 2011 Corrections had nearly 2,600 inmates with moderate
to high needs for substance abuse treatment as determined
by COMPAS, but only 800 with moderate to high needs were
assigned to substance abuse treatment during that same month.
An additional 740 inmates who had no COMPAS assessment and
310 who were assessed as having a low need for substance abuse
treatment were enrolled in the program that month. The fact
that Corrections’ staff can place inmates in the substance abuse
treatment program without a COMPAS assessment raises more
questions as to whether COMPAS is a valuable screening tool to
identify inmates’ needs. In addition, Corrections has not developed
a plan to measure COMPAS’s impact on reducing recidivism, and
thus its value is uncertain.
Using the COMPAS reentry assessment as part of the parole
planning process has potential benefits for inmates but may not yield
lower recidivism rates. Within 240 days of an inmate’s release to
parole, Corrections’ staff administer a reentry assessment that results
in a case plan to assist the inmate with transitioning to life outside
of prison. The case plan provides the inmate with tasks and goals
to pursue after being released, to address identified problems. In
addition, the case plan provides the inmate with contact information
for programs and resources in the community where he or she will
be paroled. Reentry assessments can also potentially provide value to
parole agents, eliminating the need to create case plans.
However, the value of using COMPAS reentry assessments on
inmates being paroled is questionable, since the inmates are not
required to adhere to the goals and tasks in the case plans and
parole agents’ use of COMPAS has not been consistently enforced.
One parole agent indicated he ignores the COMPAS assessments
because he has a better knowledge of the area and can recommend
programs and services that are free, or at least more affordable.
According to the associate director of the Division of Adult Parole

3

4

California State Auditor Report 2010-124

September 2011

Operations (Parole Operations), parole agents need more in-depth
training in how to apply COMPAS assessments to a parole plan that
will ultimately reduce recidivism. Parole Operations is currently
engaged in a pilot project that clarifies how parole agents should
use COMPAS. Some enhancements of the pilot project include
requiring—as a condition of parole—that parolees follow the
goals and tasks identified in the COMPAS case plan as modified
by the parole agent. Corrections has also developed a form that
requires parole agents to specify goals and how much time parolees
should spend over a certain period of time on meeting those
goals. However, the ultimate success of COMPAS relies on staff ’s
willingness to use the assessment tool.
Finally, Corrections has not tracked the actual costs of deploying
COMPAS. Our discussions with Corrections’ budget and accounting
staff revealed that they are unaware of any accounting codes in
their systems to track actual costs related to COMPAS deployment.
Corrections’ deployment of COMPAS has occurred in waves, with
its parole units and reception centers beginning to use COMPAS in
2006 and 2007, respectively. Corrections’ 33 prison institutions began
using COMPAS in March 2011. Our discussions with the COMPAS
project’s controller and technical project manager indicated that they
have not reported the actual costs associated with prison deployment
to the California Technology Agency (Technology Agency) because
Corrections’ accounting system is not set up to separately track
and report such costs, including staffing costs for Corrections’
information technology personnel associated with deployment
of COMPAS core assessments to the 33 prison institutions.
Furthermore, Corrections could not support a total of $14.6 million
in actual COMPAS costs that it reported to the Legislature in the
governor’s proposed budgets for fiscal years 2008–09 and 2009–10.
As a result, Corrections cannot demonstrate accountability for its
spending on COMPAS or explain how it monitored the project
for any cost overruns.
Recommendations
To ensure that the State does not spend additional resources on
COMPAS while its usefulness is uncertain, Corrections should
suspend its use of the COMPAS core and reentry assessments until
it has done the following:
•	 Issued regulations and updated its operations manual to define
how Corrections’ use of COMPAS will affect decision making
regarding inmates, such as clarifying how COMPAS results will
be considered when sending inmates to different prison facilities,
enrolling them in rehabilitative programs to address their criminal
risk factors, and developing expectations for those on parole.

California State Auditor Report 2010-124

September 2011

•	 Demonstrated to the Legislature that it has a plan to measure
and report COMPAS’s effect on reducing recidivism. Such a
plan could consider whether inmates enrolled in a rehabilitative
program based on a COMPAS assessment had lower recidivism
rates than those provided rehabilitative programming as a result
of non-COMPAS factors.
Once Corrections resumes its use of the COMPAS core and reentry
assessments, it should take the following steps to better ensure that
COMPAS is a valuable inmate assessment and planning tool:
•	 Provide ongoing training to classification staff representatives,
parole agents, and others who may administer or interpret
COMPAS assessment results.
•	 Develop practices or procedures to periodically compare the
demand for certain rehabilitative programs, as suggested by
a COMPAS core assessment, to the existing capacity to treat
such needs.
To ensure transparency and accountability for costs associated with
information technology projects such as COMPAS, Corrections
should take the following actions:
•	 Disclose that it lacks accounting records to support certain
COMPAS expenditure amounts it reported to the Technology
Agency and seek guidance on how to proceed with future
reporting requirements for its deployment of the COMPAS
core assessment to its adult institutions.
•	 Develop policies to ensure that accounting or budget management
personnel are involved in the planning phase of future information
technology projects so that appropriate accounting codes are
established for tracking and reporting actual project costs.
Agency Comments
Corrections disagrees with our recommendation to temporarily
suspend its use of COMPAS. Although Corrections acknowledges
that it should do more to communicate the value of COMPAS to
its employees and provide ongoing training to key staff, it believes
suspending COMPAS will result in the loss of the progress it
has made, including five years of training, and that it would take
years for it to regain its momentum. Corrections did not provide
responses to our specific recommendations, but indicated that it
would provide additional information in its 60-day, six-month, and
one-year responses.

5

6

California State Auditor Report 2010-124

September 2011

Blank page inserted for reproduction purposes only.

California State Auditor Report 2010-124

September 2011

Introduction
Background
The mission of the Department of Corrections and Rehabilitation
(Corrections) is to enhance public protection through safe and
secure incarceration of offenders, effective parole supervision, and
rehabilitative strategies to successfully reintegrate offenders into
California’s communities. Corrections’ budget for fiscal year 2011–12
is $10.1 billion, and it estimates that it will oversee 163,000 inmates
and 107,000 parolees.
Origins of Correctional Offender Management Profiling for
Alternative Sanctions Assessments
In the Budget Act of 2006, the Legislature directed Corrections
to contract with correctional program experts to complete an
assessment of California’s adult prison and parole programs
designed to reduce recidivism. As a result, Corrections created the
Expert Panel on Adult Offender Reentry and Recidivism Reduction
Programs (Expert Panel) composed of individuals from across the
nation with expertise in the field of corrections. The Expert Panel
issued its report in June 2007, recommending that Corrections
select and use a tool that would identify an offender’s criminal risk
factors—attributes directly linked to criminal behavior—so that
rehabilitative programs could be identified to treat such factors.
The Expert Panel also developed a detailed, sequential description
of how California should apply eight principles and practices
to deliver effective rehabilitative programming that it called the
California Logic Model. Concurrently, as a result of legislation
(Assembly Bill 900 (AB 900)) chaptered in May 2007, state law
was amended to require Corrections to conduct assessments of all
inmates and to place them in programs that will aid in their reentry
to society and that will most likely reduce chances of reoffending.
Citing AB 900 and the Expert Panel’s report, Corrections initiated a
plan to implement the Correctional Offender Management Profiling
for Alternative Sanctions (COMPAS) assessments, described in
the next section, for identifying the criminal risk factors of inmates
entering the prison system and those approaching their parole dates.
Deployment and Use of COMPAS
COMPAS is a family of software products that can be used for
offender assessment, classification, and case management that
was created and is owned by the Northpointe Institute for Public
Management, Incorporated (Northpointe). These software products
contain various modules that can be customized by Northpointe

7

8

California State Auditor Report 2010-124

September 2011

to meet the needs of its clients. Northpointe’s COMPAS product
includes the COMPAS core and COMPAS reentry modules, which
can be used to predict incoming and outgoing inmates’ risk of
reoffending and to assess their criminal risk factors. Both of these
modules measure criminal risk factors from different perspectives.
The COMPAS core module focuses on events and circumstances
in inmates’ pasts that led to their criminal acts, such as crime in
their social environment or a family history of jail or incarceration.
The COMPAS reentry module focuses on criminal risk factors of
inmates being released and reentering society, such as difficulty
finding a job or controlling their temper. Northpointe owns the
rights to both of these modules. Therefore, Corrections can make
system modifications only through Northpointe.
Originally, Corrections intended to use the portions of COMPAS
that assess both the risk that an inmate will reoffend and the inmate’s
criminal risk factors. However, the risk of reoffending portion had
not been tested to ensure that it was appropriate for California’s
offender population. In lieu of validating Northpointe’s proprietary
risk assessment, Corrections opted to develop its own risk
assessment, called the California Static Risk Assessment (California
risk assessment). According to Corrections, it began using the
California risk assessment in February 2009 and ultimately entered
into an agreement with Northpointe in January 2010 to include the
California risk assessment within COMPAS.
Corrections administers COMPAS’s core needs assessments in its
reception centers—where incoming inmates are sent for processing,
testing, and assignment to prisons—and in its prison institutions
if inmates have not already received a COMPAS core assessment.
To administer the COMPAS core assessment, Corrections’ staff
meet with inmates one-on-one and read them the multiple-choice
questions and record their responses. See Table 1 for the number
of questions in each type of assessment. For the most part, the
information provided by inmates cannot be verified because it
comes from their personal experiences, and Corrections’ staff
generally have limited knowledge of these experiences. According
to a COMPAS system administrator at Corrections’ headquarters,
the information in COMPAS may be accessed by certain
Corrections’ headquarters staff, as well as by a variety of prison
staff, including information system analysts, prison management,
correctional counselors and supervisors, support staff, and mental
health staff. In addition, various other Corrections’ units can access
COMPAS, including the Office of Correctional Safety, the Office of
Research, and the Data and Performance Management Unit.
After a COMPAS core needs assessment is performed, the
COMPAS software produces a bar chart identifying whether an
inmate has a high, moderate, or low need in five areas included

California State Auditor Report 2010-124

September 2011

in the assessment: substance abuse, criminal thinking, family
criminality, academic and vocational education, and anger and
violence. The identified needs help Corrections know which
programs might help reduce the inmate’s risk of reoffending after
release. Currently, Corrections uses various versions of COMPAS,
depending upon an inmate’s gender and whether the offender is
entering or leaving prison.
Table 1
Number of Questions in Each Correctional Offender Management Profiling
for Alternative Sanctions Needs Assessment
MALE
INSTITUTIONS

FEMALE
INSTITUTIONS

Correctional Offender Management Profiling for
Alternative Sanctions (COMPAS) core assessment

93

188

COMPAS reentry assessment

45

126

Sources:  Northpointe Institute for Public Management, Incorporated, and COMPAS
assessment questionnaires.

As shown in Figure 1 on the following page, Corrections started
using COMPAS in 2006 for inmates preparing to be released,
and the software was subsequently deployed to reception centers;
deployment for use in the prison institutions began only recently,
in early 2011. Prior to its contract with Northpointe in 2010,
Corrections procured services from Northpointe through various
purchase orders.
When implementing COMPAS in 2006 within the Division of
Adult Parole Operations, Corrections’ vision was to help prepare
inmates for community reintegration through prerelease planning
and the use of an automated risk and needs assessment tool.
Corrections implemented COMPAS as a part of its efforts to
establish a preparole planning and placement program.
In response to AB 900, Corrections expanded its use of COMPAS
beyond parole into its reception centers and then into the prison
institutions. As a result of these expansions, Corrections’ staff will
evaluate and review inmates’ criminal risk factors at various points
in time during their incarceration. Figure 2 on page 11 depicts
Corrections’ vision for how it ultimately plans to use COMPAS at
its reception centers, prisons, and parole offices.

9

10

California State Auditor Report 2010-124

September 2011

Figure 1
Correctional Offender Management Profiling for Alternative Sanctions Implementation Time Line
March 2006
Division of Adult Parole Operations begins assessing inmates within 240 days of parole using the Correctional
Offender Management Profiling for Alternative Sanctions (COMPAS) risk and needs assessment*
January 2007
COMPAS results available at parole field units
May 2007
Governor approves AB 900, requiring the Department of Corrections and Rehabilitation (Corrections)
to assess all inmates to place them in programs that will reduce the chance of reoffending
June 2007
COMPAS risk and needs assessments introduced in four reception centers
Expert Panel recommends that Corrections assess criminal risk factors
to determine needed rehabilitative programs‡
January 2008
Corrections develops and submits its feasibility study
report on deploying COMPAS core assessment to its
prison institutions

April 2010
Corrections enters into an agreement with Northpointe† to
deploy the COMPAS core assessment to its prison institutions
October 2010
The Office of the State Chief Information
Officer§ approves a reduction in the scope
and extends the timeline of Corrections’
deployment of COMPAS core assessment
to prison institutions

June 2008
COMPAS risk and needs assessment introduced
in remaining eight prison reception centers
February 2009
Corrections completes
COMPAS core assessment
implementation training for
correctional counselors in
reception centers

2006

2007

2008

2009

March 2011
Division of Adult Programs
directs counselors in prison
institutions to conduct
COMPAS core assessments

2010

2011

Sources:  Corrections guidance memoranda and its operations manual.
*	 When Corrections first started conducting COMPAS assessments for inmates nearing parole, it used the COMPAS risk and needs assessments,
which was replaced subsequently by the COMPAS reentry assessment.
†	 Northpointe Institute for Public Management, Incorporated.
‡	 Corrections created the Expert Panel on Adult Offender Reentry and Recidivism Reduction Programs in response to legislative direction
to contract with correctional program experts to complete an assessment of California’s adult prison and parole programs designed to
reduce recidivism.
§	 In September 2010 the Office of the State Chief Information Officer was renamed the California Technology Agency.

Corrections’ Process for Procuring COMPAS and the Role
of Northpointe
Corrections indicated that, before selecting COMPAS as its risk
and needs assessment tool for its preparole planning initiative, it
reviewed four companies in depth that provided risk and needs
assessment instruments, of which COMPAS was one. According
to Corrections, it established criteria for product standards
that included cost, technical assistance, training, automation,
reassessment capabilities, prediction, risk potential, classification
requirements, robustness of program, and comprehensiveness.
Ultimately, it selected Northpointe’s COMPAS risk and needs
assessments because Corrections felt that COMPAS most
closely mirrored its business rules, was user friendly, and was
compatible with the ever-changing needs of Corrections and
the offender population.

California State Auditor Report 2010-124

September 2011

Figure 2
Department of Corrections and Rehabilitation’s Vision of Correctional Offender Management Profiling for
Alternative Sanctions’ Role as Inmates Move Within the Prison System

RECEPTION CENTER

Inmate sentenced to prison

PAROLE

PAROLE FIELD OFFICE
Parole agent develops case plan
based on COMPAS reentry results
and goals for the parolee.

CLASSIFICATION

ENDORSEMENT

Inmate evaluated for medical, psychological,
and educational needs and security level.
CORRECTIONAL OFFENDER MANAGEMENT
FOR ALTERNATIVE SANCTIONS (COMPAS)
CORE ASSESSMENT CONDUCTED

Classification staff representative
reviews inmate file and endorses
inmate to a prison institution.
COMPAS CORE
ASSESSMENT CONSIDERED*

PRISON INSTITUTION

PAROLE PLANNING AND PLACEMENT

ANNUAL INMATE REVIEWS

Parole planning staff conduct a COMPAS
reentry assessment within 240 days of
release to identify needs once paroled.

Corrections’ staff provide inmates with
initial/annual classification reviews to
determine rehabilitative program needs.
COMPAS CORE ASSESSMENT
CONDUCTED†

COMPAS REENTRY ASSESSMENT CONDUCTED

COMPAS REENTRY
ASSESSMENT CONSIDERED
Inmate receives rehabilitative programs.
Sources:  Department of Corrections and Rehabilitation’s (Corrections) guidance memoranda and its operations manual.
*	 Corrections intends that classification staff representatives at reception centers ensure that inmates are sent to prisons that are consistent with their
risk and needs scores. For example, inmates with a high risk to reoffend and an indication of a drug problem should be sent to a prison with a
substance abuse treatment program.
†	 Corrections’ staff conduct COMPAS core assessments during initial/annual inmate reviews only if assessments were not conducted when inmates
were in a reception center. Corrections plans to have Northpointe Institute for Public Management, Incorporated develop an annual follow‑up
assessment version of COMPAS for use in its prison institutions.

11

12

California State Auditor Report 2010-124

September 2011

Corrections acquired COMPAS in 2005 through a leveraged
procurement agreement that the Department of General Services
(General Services) had established to allow state departments
to buy information technology products, software, and certain
services at preestablished prices. When General Services did not
renew the leveraged procurement agreement beginning in fiscal
year 2005–06, Corrections entered into subsequent purchase
agreements directly with Northpointe. However, when Corrections
decided to expand COMPAS to the general-population prisons, it
executed a contract with Northpointe for $3.97 million with a term
running from April 30, 2010, through June 30, 2012.
Northpointe’s role in the development and implementation of
COMPAS initially focused primarily on providing the software
licenses, modifying COMPAS as requested by Corrections, and
later assisting with the development and delivery of COMPAS
training to Corrections’ staff. In its monthly status report to
the California Technology Agency for March 2011, Corrections
reported that all of its 33 prison institutions had received training
and instructions to begin using COMPAS.
COMPAS and the Strategic Offender Management System
Corrections plans to implement the Strategic Offender Management
System (SOMS), a fully automated system that will consolidate
existing databases and records and replace manual paper processes.
According to Corrections’ planning documents, SOMS and
COMPAS will share certain inmate data. Specifically, SOMS will
provide COMPAS with offender identification data, such as the
offender’s name and Social Security number. COMPAS will provide
SOMS with criminal risk factor data for offenders via a real-time
Web services interface. Corrections plans to have Northpointe
modify COMPAS to allow for an interface with SOMS in subsequent
released versions. According to the technical architect on the SOMS
project, SOMS will not interface with COMPAS until sometime
late in 2012.
Scope and Methodology
The Joint Legislative Audit Committee (audit committee)
directed the Bureau of State Audits (bureau) to review Corrections’
use of the COMPAS assessment tool for identifying the rehabilitative
needs of its inmates. Specifically, the audit committee directed us
to review the goals for using COMPAS and evaluate whether any of
Corrections’ policies, procedures, or planned changes to COMPAS
will facilitate achieving these goals. The audit committee also asked
us to determine the cost to implement COMPAS, including the

California State Auditor Report 2010-124

September 2011

original budget and the actual amount spent to date as well as the
costs to administer and maintain COMPAS. To the extent that
COMPAS is not complete or fully deployed, the audit committee
asked us to identify the budget and projected timeline to fully
complete and deploy the program; determine if the contractor’s
scope of work is sufficient to complete and fully deploy COMPAS;
determine if the plans for the program contain a clear path for
completing COMPAS, which includes key milestones and their
respective costs and completion time frames; and assess whether
such estimated costs and timelines are attainable.
The audit committee also directed us to examine how Corrections
is currently using COMPAS and whether its use is meeting
established goals and objectives. Specifically, the audit committee
asked us to determine whether all new and existing inmates have
been assessed and, if not, the timeline for assessing all inmates;
the sources of data recorded in COMPAS and whether such
information is verified beforehand; whether COMPAS evaluations
result in the identification of education programs at a higher level
than those the inmate completed during a prior incarceration; the
extent to which inmates are enrolled in programs identified through
assessments performed with COMPAS; and whether Corrections
has measured COMPAS’s affect on recidivism rates and, if not,
whether Corrections is collecting the necessary information for this
measurement. Finally, the audit committee directed us to review
and assess Corrections’ plan for incorporating COMPAS into the
SOMS project.
To address the audit committee’s objectives, we performed the
procedures shown in Table 2 on the following page. Additionally,
we relied on data provided by Corrections. The U.S. Government
Accountability Office (GAO), whose standards we follow, requires
us to assess the sufficiency and appropriateness of computerprocessed information. To comply with this standard, we assessed
the reliability of Corrections’ Offender Based Information System
(OBIS) for the purpose of identifying the number of unique inmates
who were released to parole during fiscal years 2007–08 through
2010–111 (as displayed in Table 7 on page 33). Specifically, we
performed data-set verification procedures, performed electronic
testing of key data elements, and assessed the accuracy of OBIS.
We did not perform completeness testing of OBIS because the
source documents for this system are stored at the 33 adult inmate
institutions located throughout the State, making such

1	

For each of the three systems discussed here, fiscal year 2010–11 includes data for the period
July 1, 2010, through February 20, 2011, only.

13

14

California State Auditor Report 2010-124

September 2011

Table 2
Methods of Addressing Audit Objectives
AUDIT OBJECTIVE

METHOD

Review and evaluate the goals and objectives of the
Correctional Offender Management Profiling for Alternative
Sanctions (COMPAS) program. In addition, review and
assess the Department of Corrections and Rehabilitation’s
(Corrections) policies, procedures, and planned changes for
determining whether the goals and objectives will be met.

Reviewed relevant laws and COMPAS-related documents, such as Corrections’ feasibility study
report; Corrections’ internal administrative memoranda defining the goals and use of COMPAS
within its reception centers, prison institutions, and parole units; and Corrections’ communication
with the Legislature regarding how it intends to use COMPAS. Finally, we interviewed Corrections’
staff from eight reception centers and two parole offices to better understand COMPAS’s role in
Corrections’ business practices.

Determine the cost to implement COMPAS, including
budgeted amounts and the actual costs to date. Further,
determine the cost to administer and maintain the
COMPAS program.

Reviewed Corrections’ initial and revised budget for its deployment of COMPAS to its prison
institutions as reported to the Office of the State Chief Information Officer.* Corrections lacked
accounting records that specifically tracked actual COMPAS costs. The lack of accounting information
prevented us from fully addressing this audit objective.

If COMPAS is not complete and fully deployed, to the
extent possible:

Corrections has deployed COMPAS to its reception centers, prison institutions, and parole units.
Nevertheless, we performed the following:

a) Identify the budget and projected timeline to fully
complete and deploy the program.

• Reviewed available budget and project schedule information.

b) Determine if the contractor’s scope of work is sufficient to
complete and fully deploy the program.

• Obtained the scope of work contained in Corrections’ contract with its vendor, Northpointe
Institute for Public Management, Incorporated, for the deployment of COMPAS to its
prison institutions.

c) Determine if the plans for the program contain a clear
path for completing COMPAS that includes key milestones
and their respective costs and completion timelines.
d) Assess whether such estimated costs and timelines
are attainable.
To the extent possible, examine how Corrections is using
COMPAS, the scope of its use, and the extent to which it
is operating as designed by determining, at a minimum,
the following:
a) Whether all new and existing inmates have been
assessed. If not, identify the timeline for assessing
all inmates.
b) The sources of data recorded in COMPAS, whether such
data is verified before being entered into COMPAS, and
who has access to data in the COMPAS system.
c) Whether COMPAS evaluations consistently result in the
identification of educational programs at a higher level
than those an inmate successfully completed previously.
In addition, determine the extent to which inmates are
enrolled in programs that address the needs identified
through COMPAS.
d) Whether Corrections has measured the COMPAS
program’s impact on any preliminary recidivism rates.
If not, determine whether Corrections is collecting the
appropriate data to measure its impact on recidivism
rates in the future.

Review and assess Corrections’ plan for incorporating
COMPAS into the Strategic Offender Management System
(SOMS) project.

We obtained Corrections’ various policy memoranda that exclude certain inmates from receiving
a COMPAS assessment. To determine the proportion of the total inmate and parolee populations
for fiscal years 2007–08 through 2010–11† that received a COMPAS assessment, we analyzed
data collected from Corrections’ Offender Based Information System, Distributed Data Processing
System, and the COMPAS database.‡ We also interviewed Corrections’ staff to determine when
they intend to assess all inmates.
We observed Corrections’ staff administer COMPAS assessments as part of the reception center and
preparole planning process at one prison institution, Deuel Vocational Institution. We also considered
whether staff verify information prior to entering the information in COMPAS. To determine who has
access to COMPAS data, we interviewed Corrections’ project manager for COMPAS deployment.
We evaluated Corrections’ policies for considering COMPAS results when placing inmates into
rehabilitative programs. For February 20, 2011, we compared the number of inmates whose
COMPAS core assessment determined they had a moderate to high need for substance abuse
treatment and were within four to 12 months of their earliest possible release date with the
number of inmates who were ultimately placed in rehabilitative programs.
We were unable to determine whether COMPAS assessments consistently result in the identification
of educational programs at a higher level than those an inmate successfully completed during a prior
incarceration because data concerning inmates completing rehabilitative programs is not recorded
in a database and, therefore, was not retrievable for this type of analysis.
Reviewed Corrections’ periodic reports on recidivism measures and determined whether COMPAS
was discussed. Interviewed Corrections’ staff to determine whether there is a plan to assess
COMPAS’s affect on recidivism.
Interviewed Corrections’ staff involved with the implementation of the SOMS and COMPAS to
determine Corrections’ vision for how the two systems will share information, and when this
functionality will exist.

Sources:  Bureau of State Audits’ analysis of audit request 2010-124 and its analysis of information and documentation identified in the Method column of
the table above.
*	 In September 2010 the Office of the State Chief Information Officer was renamed the California Technology Agency.
†	 Fiscal year 2010–11 includes data for the period July 1, 2010, through February 20, 2011, only.
‡	 We assessed the reliability of the data contained in each of these systems and discuss the results beginning on page 13.

California State Auditor Report 2010-124

September 2011

testing impractical. We did not identify any errors when performing
data-set verification procedures, nor did we identify any illogical
information in the key fields used in our analysis. For accuracy
testing, we selected a random sample of 29 transactions and tested
the accuracy of nine key fields. Of the nine key fields tested, we
found errors in two key fields. Based on our testing and analysis,
we found that OBIS is of undetermined reliability for the purpose
of identifying the number of inmates who were released to parole
during fiscal years 2007–08 through 2010–11.
In addition, we assessed the reliability of Corrections’ Distributed
Data Processing System (DDPS) for the purpose of identifying the
number of inmates who were housed in an institution or camp
during fiscal years 2007–08 through 2010–11 (as displayed in
Table 6 on page 32). Specifically, we performed data-set verification
procedures, performed electronic testing of key data elements, and
assessed the accuracy of DDPS. We did not perform completeness
testing of DDPS because the source documents for this system are
stored at the 33 adult inmate institutions located throughout the
State, making such testing impractical. We did not identify any
errors when performing data-set verification procedures, nor did
we identify any illogical information in the key fields used in our
analysis. We were unable to test the accuracy of DDPS for the period
July 1, 2007, through February 19, 2010, because Corrections had
previously destroyed the hard-copy source documents in accordance
with its record retention schedule. Therefore, we selected a random
sample of five transactions for the period February 20, 2010, through
February 20, 2011, and tested the accuracy of five key fields and
found no errors. Based on our testing and analysis, we found that
DDPS is of undetermined reliability for the purpose of identifying
the number of inmates who were housed in an institution or camp
during fiscal years 2007–08 through 2010–11.
Finally, we assessed the reliability of Corrections’ COMPAS
database for the purpose of identifying the number of inmates
who received at least one COMPAS assessment (as displayed in
tables 6 and 7 on pages 32 and 33, respectively) and the number
of inmates housed in an institution or camp on February 20, 2011,
who were identified as having a moderate to high criminal risk
factor in any of the five areas assessed by the COMPAS core
assessment (as displayed in Table 3 and Figure 4 on pages 23 and 27,
respectively). Specifically, we performed data-set verification
procedures and electronic testing of key data elements and did not
identify any issues. We did not perform accuracy and completeness
testing of COMPAS because it is a paperless system, and thus,
hard-copy source documentation was not available for us to
review. Alternatively, following GAO guidelines, we reviewed
selected system controls, which included general and business
process application controls. General controls are the policies and

15

16

California State Auditor Report 2010-124

September 2011

procedures that apply to all or a large segment of Corrections’
information systems and help ensure their proper operation.
Business process application controls are directly related to a
specific computerized application, COMPAS in this case, and help
to ensure that transactions are complete, accurate, and available.
The preliminary results of this review indicate that Corrections has
weaknesses in its general controls associated with a large segment
of its information systems. The specifics of this review will be
published in a separate management letter.

California State Auditor Report 2010-124

September 2011

Audit Results
The Benefits of the Correctional Offender Management Profiling for
Alternative Sanctions Program Have Yet to Be Realized
Although the Department of Corrections and Rehabilitation
(Corrections) has been conducting Correctional Offender
Management Profiling for Alternative Sanctions (COMPAS)
assessments for inmates about to be released from prison since
2006 and for inmates entering prison since 2007, the benefits of
these assessments are unclear. This is mainly due to the limited
opportunities for inmates to participate in Corrections’ existing
in‑prison rehabilitative programs, because factors other than
COMPAS assessments—such as security and available bed space—
take priority in determining where inmates are assigned for housing
and, by extension, the rehabilitative programs they might receive at
these facilities. Also, Corrections generally has limited capacity in
its rehabilitative programs. Even in the only rehabilitative program
area where COMPAS plays a role in determining admission—
substance abuse treatment—a limited number of inmates with
COMPAS‑identified needs receive treatment. In addition,
Corrections does not have a plan to evaluate whether COMPAS
will help it reach its ultimate goal of reducing prison overcrowding
and recidivism. Further, according to Corrections, parole agents do
not always use COMPAS assessments in developing parole plans
for inmates about to be released from prison and parolees are not
required to adhere to their COMPAS-generated case plans.
Corrections’ Process for Assigning Inmates to Prison Facilities and to
Rehabilitative Programs Limits the Usefulness of COMPAS
Corrections uses a complex and multifaceted process for
assigning inmates to prison facilities. The complexity of this
process has limited Corrections’ opportunity to effectively
use COMPAS core assessments. Corrections’ intent is to use
COMPAS core assessments to identify criminal risk factors—
attributes directly linked to criminal behavior—and then identify
rehabilitative programs to address these factors. Corrections
began using COMPAS core assessments in 2008 at its 12 reception
centers—facilities that temporarily house incoming inmates
until prison assignments are made—so it could make better
placement decisions.
State regulations require Corrections’ staff to consider many factors
when making prison placement decisions. For example, state
regulations require that each inmate be assigned to a prison facility
with a security level based on his or her placement score—derived
from a variety of factors such as the inmate’s length of sentence,

17

18

California State Auditor Report 2010-124

September 2011

The eight reception centers
indicated that COMPAS assessments
do not play a significant role when
deciding where inmates should
be housed.

gang activity, mental health, history of prior incarcerations, and
behavior during those incarcerations. Further, these regulations
instruct Corrections’ staff to consider other administrative factors
when assigning inmates, such as any known enemies housed
within a prison, the inmate’s escape risk, medical conditions, prior
or current sexual offenses, and history of violence. According to
officials from eight of Corrections’ 12 reception centers, these
criteria are the highest-priority determinants of where an inmate
can be housed. As a result, the eight reception centers indicated
that COMPAS assessments do not play a significant role when
deciding where inmates should be housed. This perspective was
confirmed through our own observations at one reception center,
where we observed several inmate interviews during which a
reception center’s staff provided inmates with their placement
scores and with the opportunity to request placement at particular
prison facilities, as long as they were consistent with the inmates’
placement scores.
Limited bed space at receiving prisons can also reduce the
opportunity for COMPAS to play a role when assigning inmates to
prisons and, by extension, the rehabilitative programs inmates
might join. For example, Corrections’ Population Management Unit,
the unit responsible for coordinating the movement of inmates
between prisons, issues weekly guidance on the prisons into which
each reception center can place inmates. For example, for the week
of March 7, 2011, a male inmate in the High Desert State Prison
reception center requiring placement in a prison with a security
level of 2—low to moderate security—could be sent to only one of
two institutions. The weekly guidance also specifies any overall
placement priorities, such as assignments to conservation camps
or out-of-state correctional facilities. The March 7, 2011, guidance
included a list of eight such priorities, of which placing inmates in a
substance abuse treatment program was the last. Three classification
staff representatives—individuals with responsibility for deciding
where an inmate is placed at the conclusion of their stay at a
reception center—with whom we spoke told us that they base their
prison assignment decisions primarily on the weekly guidance to the
extent the priorities are compatible with the inmate’s classification
level and medical or psychiatric needs.
Corrections’ ability to use COMPAS core assessments to place
inmates in rehabilitative programming is further limited since
the assessments are being used only for placing inmates in the
substance abuse treatment program, even though the COMPAS
core assessment identifies an inmate’s needs in four other areas.
When it expanded its use of COMPAS core assessments to
general-population inmates, Corrections issued a March 2011
memorandum to prison wardens and other high-level staff
indicating that correctional staff will use the criminal risk factor

California State Auditor Report 2010-124

September 2011

scores from COMPAS in conjunction with established priority
placement models to assign offenders to programs and locations
consistent with their needs. The March 2011 memorandum
referenced an attached inmate placement priority matrix
(priority matrix).
Corrections’ priority matrix establishes criteria for correctional
staff to follow when deciding which inmates should participate
in certain rehabilitative programs. Our review of the priority
matrix found that COMPAS results are to be considered only
when the inmate has a need for substance abuse treatment. The
placement matrix establishes six different priority groups for
substance abuse. In general, inmates who are within 12 months
of release from prison and demonstrate a moderate to high risk of
committing another crime once released receive priority placement
in the program. Results from an inmate’s COMPAS assessment
indicating a moderate to high need for substance abuse treatment
can increase an inmate’s priority but is not a primary factor. For
example, inmates with a moderate to high need for substance
abuse treatment—either as identified by the COMPAS assessment
or based on a review of the inmate’s file—can fall to the bottom
half of the priority matrix if they have a longer amount of time to
serve or if their risk of reoffending is low. We also noted that certain
inmates were ineligible for the substance abuse treatment program,
such as those who have had discipline problems within the last
12 months, those who are active gang members or sex offenders,
and those who are subject to deportation. An inmate’s priority
for placement into Corrections’ other rehabilitative programs—
namely, its academic and vocational programs—does not depend
on COMPAS but is based instead on the inmate’s educational
background as determined by other non-COMPAS assessments—
such as the test of adult basic education—and the inmate’s ability to
benefit from the program.
Corrections Could Do More to Promote Understanding of COMPAS
Core Assessments
Corrections has not implemented COMPAS core assessments
in a way that allows its staff to understand the value of using this
tool. According to one classification staff representative who
indicated that she has been with Corrections for 24 years and
rotates each week to a different prison in her region, a lack of
buy-in exists across prison institutions regarding the benefits
of COMPAS. The classification staff representative also indicated
that staff in her position have received no training in incorporating
COMPAS results into their decisions when placing inmates into

An inmate’s priority for placement
into Corrections’ academic programs
does not depend on COMPAS but is
based on non‑COMPAS assessments.

19

20

California State Auditor Report 2010-124

September 2011

prison facilities, and as a result she does not feel that Corrections’
headquarters has effectively communicated its expectations with
respect to the COMPAS assessments.
We spoke with the COMPAS user project manager regarding
the extent to which classification staff representatives have been
trained, and we were told that training has not been provided
because reception centers are still developing procedures for using
COMPAS. Another classification staff representative indicated
that considering COMPAS core assessments as meaningful criteria
for assigning inmates to prisons is not feasible due to a lack of
rehabilitative programs and the other factors they need to consider,
such as an inmate’s classification placement score and gang
affiliations, among others. An associate warden at the California
Correctional Facility for Women told us he doubts whether
COMPAS core assessments identify things that his staff do not
know about inmates based on the information already available. For
example, the associate warden indicated that an inmate’s substance
abuse problem can be identified based on the individual’s arrest
history, which was done prior to using COMPAS core assessments.
The associate warden also indicated that he does not think many of
his staff see the benefit of COMPAS core assessments.

We see limited evidence that
Corrections has reengineered its
practices or involved stakeholders
such as the staff who are expected
to use COMPAS core assessments
in prisons.

When the Expert Panel on Adult Offender Reentry and Recidivism
Reduction Programs (Expert Panel) made its recommendations
to Corrections that it take an evidence-based approach to
rehabilitating inmates, it mentioned that reducing recidivism is
not simply a matter of identifying evidence-based programs that
produce results, but that the greater challenge lies in changing how
Corrections operates. The Expert Panel indicated that political
and correctional agency leaders must engage in organizational
reengineering tasks to successfully move from traditional
warehousing or custodial practices to evidence-based rehabilitation
principles and practices. The Expert Panel also discussed the need
for Corrections to collaborate with stakeholders as part of this
process. However, we see limited evidence that Corrections has
reengineered its practices or involved stakeholders such as the staff
who are expected to use COMPAS core assessments in prisons.
For example, Corrections has issued regulations and adopted a
department operations manual that define how its staff are to
interact with inmates, covering topics such as how inmates are
processed, evaluated, and ultimately released. However, neither
Corrections’ regulations nor its operations manual discusses how
the use of COMPAS fits within Corrections’ current practices.
Although Corrections has issued several memoranda since 2008
that attempt to refine and clarify its policy on COMPAS core
assessments, these rules were established without public input.

California State Auditor Report 2010-124

September 2011

Our legal counsel reviewed three of these memoranda and
concluded that they established underground regulations
because they were not adopted in accordance with the California
Administrative Procedures Act. That law requires departments
and agencies to give the public an opportunity to participate in the
adoption of state regulations to ensure that such regulations are
clear, necessary, and legally valid. Moreover, that same law prohibits
enforcement of underground regulations. Corrections’ operations
manual also specifies that its stakeholders shall have input on
the development of its regulations, clarifying that a rule, order,
or standard may be considered a regulation if it affects decisions
concerning inmate custody, discipline, classification, programming,
release date, and other factors, such as those that might influence
the terms or conditions of parole. Had Corrections solicited public
input from its stakeholders and adopted regulations, it might have
been in a better position to ensure that its application of COMPAS
in the prison setting is legally enforceable, would be understood
by its employees, and would have a demonstrable effect on inmate
decision making. Until the Expert Panel’s recommendations
are heeded, we believe Corrections will continue to implement
COMPAS in a way that provides limited value.
Corrections’ Limited Rehabilitative Programs Lessen the Value of Its
Needs Assessment Tool
In its August 2008 memorandum to reception center and prison
staff regarding the use of COMPAS, Corrections explained that
alleviating California’s prison overcrowding problem and improving
rehabilitation outcomes was based, in part, on getting the right
person to the right program at the right time. However, the
value of using COMPAS core assessments to identify an inmate’s
criminal risk factors—and thereby the rehabilitative programming
required—has been diminished by the lack of programs to treat
such factors. Corrections has rehabilitative programs that address
only two of the five criminal risk factors currently being identified
with COMPAS core assessments in reception centers—academic/
vocational education and substance abuse treatment. In an
April 2011 letter to a member of the Legislature, Corrections’
secretary recognized that Corrections does not have the financial
resources to provide comprehensive education and substance abuse
treatment to offenders. Nevertheless, the letter stated that the
secretary believes assessing inmate needs is still worthwhile because
doing so avoids having Corrections guess which inmates need what
services and potentially wasting limited resources on the wrong
population. However, the secretary’s comments stand in contrast
to the fact that more inmates receiving COMPAS core assessments
have moderate to high substance abuse problems when compared

Until the Expert Panel’s
recommendations are heeded, we
believe Corrections will continue to
implement COMPAS in a way that
provides limited value.

21

22

California State Auditor Report 2010-124

September 2011

to those in need of academic/vocational programs, yet the capacity
for academic/vocational programs is more than five times greater
than the capacity for substance abuse treatment.
Corrections has limited rehabilitative programs that address
COMPAS-identified needs. For example, Table 3 shows the
estimated number of inmates who have moderate to high criminal
risk factors in the five areas assessed by the COMPAS core
assessment, compared to Corrections’ estimated capacity to address
those needs. For our analysis, we had to estimate the number of
inmates who would likely have certain criminal risk factors, since
not every inmate has received a COMPAS core assessment. On
February 20, 2011, Corrections had an inmate population
of 146,111 individuals in its various institutions and camps, of which
35,814—or 24.5 percent—had received at least
one COMPAS core assessment. The text box
shows how many inmates receiving the
Number of Inmates for Whom a Correctional
Offender Management Profiling for Alternative
assessment had moderate to high criminal risk
Sanctions Core Assessment Identified
factors and the type of need identified.
Moderate or High Criminal Risk Factors
as of February 20, 2011

The inmates included in the numbers shown
in the text box can have multiple moderate
•	 Substance abuse treatment needs: 25,512
to high needs and can thus be included in
•	 Academic/vocational needs: 19,708
more than one category. By assuming that
those inmates who received COMPAS core
•	 Anger and violence needs: 16,285
assessments are representative of the entire
•	 Criminal thinking needs: 15,452
prison population, we were able to arrive at a
•	 Family criminality needs: 10,659
rough estimate of the potential need that exists
for rehabilitative programs. For example, based
Source:  Bureau of State Audits’ analysis of data collected from
the Department of Corrections and Rehabilitation’s Correctional
on the 35,814 inmates that had received COMPAS
Offender Management Profiling for Alternative Sanctions
core assessments, we estimate in Table 3 that
(COMPAS) database, Distributed Data Processing System, and
66,438 inmates in the total population have
Offender Based Information System.
As of February 20, 2011, 35,814 inmates had received at least
rehabilitative needs related to anger and violence.
one COMPAS core assessment.
However, Corrections has no rehabilitative
programs in place for this and two other criminal
risk factors identified through COMPAS.
When we asked Corrections for its perspective, the director of
rehabilitative programs indicated that while Corrections does not
contract for services to provide programs for assessed criminal risk
factors related to criminal thinking, anger and violence, and family
criminality, it has a number of programs provided by volunteers
and/or self-help groups that cover these topics, including Anger
Management, Families in Conflict, and Building a Marriage for
Success. Although Corrections makes these voluntary programs
available, we saw no evidence suggesting that Corrections directed
its staff to consider COMPAS core assessment results for placement
into these other programs.

California State Auditor Report 2010-124

September 2011

Table 3
Comparison of the Estimated Number of Inmates With Certain Criminal Risk
Factors to the Availability of Rehabilitative Programs by Category
Fiscal Year 2010–11
ESTIMATED NEEDS—
CONSIDERING TIME
LEFT TO SERVE

ANNUAL
PROGRAM
CAPACITY†

80,403

16,574

43,601

Substance abuse

104,082

17,495

7,886

Criminal thinking

63,040

NA

NA

Anger and violence

66,438

NA

NA

Family criminality

43,486

NA

NA

CRIMINAL RISK FACTOR IDENTIFIED

Academic/vocational education

OVERALL
ESTIMATED
NEEDS*

Sources:  Bureau of State Audits’ analysis of data collected from the Department of Corrections and
Rehabilitation’s (Corrections) Correctional Offender Management Profiling for Alternative Sanctions
(COMPAS) database, Distributed Data Processing System, and Offender Based Information System,
and Corrections’ program capacity estimates.
NA = Not applicable. Corrections has no rehabilitative programs in place for these criminal risk
factors identified by COMPAS.
*	 We calculated the overall estimated needs by determining the ratio of inmates with COMPAS core
assessments indicating a moderate or high need in each category to total inmates with at least
one COMPAS core assessment (35,814), and multiplying this ratio by the number of inmates who
were housed in an institution or camp on February 20, 2011 (146,111). Inmates with more than
one moderate to high assessed need are counted in each category.
†	 Unaudited program capacity information was provided by the chief of Corrections’ Data and
Performance Management Unit within the Office of Rehabilitative Program Planning
and Accountability.

Table 3 raises questions as to why academic/vocational programs
have more than five times the program capacity compared to
substance abuse treatment, particularly since substance abuse
is the criminal risk factor most often cited by COMPAS. For
example, the text box shows that 25,512 out of 35,814 inmates—
or 71 percent—with COMPAS core assessments had moderate or
high substance abuse treatment needs, whereas only 19,708—
or 55 percent—had moderate or high academic/vocational
programming needs. Corrections indicated that it reduced the
number of its in-prison substance abuse treatment programs
in response to budget reductions and that it prioritizes inmate
placement into its academic/vocational and substance abuse
treatment programs based on a variety of factors, such as the
inmate’s risk of reoffending, remaining time left to serve until
release, and degree of need—whether identified by the COMPAS
core assessment or other assessment tools.
The middle column in Table 3 shows our estimate of the number
of inmates who have moderate to high needs for academic/
vocational and substance abuse treatment programs and who

23

24

California State Auditor Report 2010-124

September 2011

generally fall within Corrections’ time-left-to-serve guidelines.2 As
the table illustrates, Corrections has more inmates with moderate
or high substance abuse treatment needs—17,495 inmates who
are within four to 12 months of release—than the 7,886 capacity
in its in‑prison substance abuse treatment program. In contrast,
Corrections has more capability to address academic/vocational
needs, with an annual capacity of 43,601 inmates, while we estimate
that there are 16,574 inmates with moderate or high needs who
are within 13 to 48 months of release. Our analysis in Table 3
does not consider whether the 16,574 or 17,495 inmates previously
received treatment for their needs since Corrections does not have
electronic records indicating past participation in rehabilitative
programs. When we asked Corrections for its perspective, the
director of rehabilitative programs explained that Corrections has
not historically been funded based on program need and that using
inmate need data to project budgetary and fiscal requirements
is a relatively new concept and one that has not been integrated
into its funding formulas. The director of rehabilitative programs
also explained that Corrections must comply with statutory
requirements, such as providing literacy programs in each state
prison. Further, Corrections indicated that it is still in the process
of determining the best way to determine needed capacity in its
rehabilitative programs based on inmate need, length of stay, and
other factors.

Corrections indicated that it
anticipated cutting at least
$250 million from rehabilitative
programs, including academic/
vocational education,
substance abuse treatment,
and other programs.

Corrections has cited funding reductions as a factor contributing to
its limited adult rehabilitative programs. During fiscal year 2009–10,
the Legislature recognized that Corrections’ budget included a
reduction in funding for inmate and parolee programs designed
to reduce recidivism. As a provision of the amended 2009 Budget
Act, the Legislature required Corrections to explain and report
by August 15, 2009, how it intended to achieve the unspecified
cuts. Corrections indicated that it anticipated cutting at least
$250 million from rehabilitative programs, including academic/
vocational education, substance abuse treatment, and other
programs for inmates and parolees. In its follow-up report to the
Legislature in April 2010 regarding its progress toward making these
cuts, Corrections explained that it had changed how it delivered
rehabilitative programs. Some of the more significant changes
included reducing the length of its substance abuse treatment
program from a 12- to 36-month program to a three‑month program
and reducing the number of prisons where this treatment is provided.
2	

Corrections’ guidance generally indicates that inmates should be admitted to academic and
vocational education programs when they have 48 months left to serve and to substance abuse
programs when they have 12 months left to serve. Further, because these programs take 12 and
three months, respectively, to complete, we limited our analysis to inmates with 13 to 48 months
left to serve for academic and vocational education and four to 12 months for substance abuse,
which allowed Corrections one month to get an inmate placed in the correct program.

California State Auditor Report 2010-124

September 2011

Additionally, Corrections reported that it developed new education
models that require inmates to be given priority for classes based on
their risk of reoffending and how much time they have remaining
until release. Corrections stated in April 2010 that $200 million
in reductions would come from funding for adult rehabilitative
programs, with another $50 million in cuts to be shared by parole
and female offender programs. Our review of expenditure records
maintained by the State Controller’s Office revealed that Corrections
has reduced some of its spending on rehabilitative programs. As
shown in Figure 3, Corrections’ spending on its Adult Education,
Vocation, and Offender programs was at its highest point during fiscal
year 2009–10 but fell by $69.1 million to $327.7 million disbursed
during fiscal year 2010–11.
Figure 3
Department of Corrections and Rehabilitation Disbursements for Adult
Education, Vocation, and Offender Programs
$400

Disbursements (in Millions)

350
300
250
200
150
100
50
0
2006–07

2007–08

2008–09

2009–10

2010–11

Fiscal Years
Source:  State Controller’s Office.
Notes:  Disbursements represent payments made during the fiscal years shown, regardless of
the year in which the Legislature provided the Department of Corrections and Rehabilitation
(Corrections) with the authority to spend these funds.
Disbursement amounts shown here are limited to Corrections’ activity under the Adult Education,
Vocation, and Offender programs (program 45) portion of its budget. This portion includes funding
for education, vocational, and in‑prison substance abuse programs for adult offenders. Beginning
in fiscal year 2010–11, this portion of Corrections’ budget was split into multiple programs
(programs 45, 46, and 48).

25

26

California State Auditor Report 2010-124

September 2011

Inmates With Substance Abuse Problems Are Not Always Placed in
Rehabilitative Programs, and Corrections Lacks a Plan to Measure
COMPAS’s Impact on Recidivism

Corrections’ 11 institutions with
substance abuse programs treat only
a limited number of inmates with
moderate to high substance abuse
treatment needs, as determined
by COMPAS.

Corrections uses COMPAS core assessments only to determine
an inmate’s admission to its substance abuse treatment program.
Limited capacity in the treatment program, along with other
exclusionary factors—such as gang membership or recent
disciplinary problems—can prevent an inmate with substance
abuse problems from receiving treatment. Our review found
that Corrections’ 11 institutions with substance abuse programs
treat only a limited number of inmates with moderate to high
substance abuse treatment needs, as determined by COMPAS. In
February 2011, the 11 institutions shown in Figure 4 housed nearly
2,600 inmates that had moderate to high substance abuse treatment
needs as identified through a COMPAS core assessment and that
were within four to 12 months of their earliest possible release
date. However, according to Corrections, only about 800 inmates
with moderate to high needs—as determined by COMPAS core
assessments—were assigned to the substance abuse treatment
program during that same month. An additional 740 inmates
assigned to the program had no COMPAS assessment on file, and
310 had a COMPAS assessment that indicated a low substance
abuse need. Corrections’ policies allow its staff to assign inmates
with a documented history of substance abuse to the substance
abuse treatment program, regardless of whether they have received
a COMPAS core assessment. However, these broad guidelines and
the limited controls Corrections has to ensure that staff consider
COMPAS likely contribute, in some part, to the statistics shown
in Figure 4. Finally, while Figure 4 shows that COMPAS’s impact
on getting inmates into the substance abuse program is at best
uncertain, Corrections currently lacks a plan to measure COMPAS’s
contribution to reducing recidivism and prison overcrowding, the
ultimate goal of using COMPAS in the first place.
Relatively few inmates with moderate to high substance abuse
treatment needs, as determined through a COMPAS core
assessment, are assigned to Corrections’ treatment program. For each
of the 11 institutions that have such a treatment program, Figure 4
provides two pieces of information. First, it depicts the number of
inmates with moderate to high substance abuse treatment needs as
determined by a COMPAS core assessment and who are roughly
within Corrections’ time-left-to-serve guidelines. For example,
Figure 4 shows that as of February 20, 2011, Avenal State Prison had
321 inmates with moderate to high substance abuse treatment needs
who were within four to 12 months of their earliest possible release
date. The figure also shows—based on Corrections’ February 4, 2011,
analysis of current participants in its substance abuse treatment
program as of that date—that Avenal State Prison had 93 inmates in

California State Auditor Report 2010-124

September 2011

the program who had moderate to high needs according to a COMPAS
core assessment, while another 75 had no COMPAS assessment, and 27
were assessed as having a low need. Although it is possible that some of
the 321 inmates at Avenal State Prison with moderate to high substance
abuse treatment needs, as well as the inmates at the other 10 institutions
that had similar needs, had previously received substance abuse
treatment prior to February 4, 2011, we could not determine the number
of inmates in this category because Corrections does not have electronic
records indicating an inmate’s history in rehabilitative programs.
Figure 4
Comparison of Inmate Needs Versus Placement in a Substance Abuse Treatment Program in February 2011
Inmates with moderate or high substance abuse treatment needs according to the Correctional Offender
Management Profiling for Alternative Sanctions (COMPAS) core assessment*
Inmates in the substance abuse treatment program with low needs, according to COMPAS†
Inmates in the substance abuse treatment program with no COMPAS assessment on file†
Inmates in the substance abuse treatment program with moderate or high needs, according to COMPAS†

400
350

Number of Inmates

300
250
200
150
100
50
0
on

al

en
Av

is
Pr
te
a
t
S
ni

r
lifo

Ca

l

na

o
cti

re
or
aC

n

ral
nt
Ce

ilit

tio

tu

ti
Ins

ac
’s F
n
e

om
aW

i

Ca

rn
lifo

li
Ca

for

nia

y
or
nf

tio

tu

ti
Ins

n

ia

rn

lifo
Ca

m
Wo
or
f
n

io

t
itu

t
Ins

ny

en

Me

lifo
Ca

ia
rn

M

olo
’s C
n
e

ilit

c
Fa
ing

in
Tra
al

on
cti

re

r
Co

k

c
hu

y

on

ey

all
aV

all
aw

C

ia
rn

ac
tF
en

tm
rea
T
e

us

b

A
ce
tan

bs

Su

ilit

is
Pr
te
a
t
S

nia

for

li
Ca

y
S
n,

o

is
Pr
te
a
t
S

ey

ll
Va

no

en

ola

m
Wo
or
f
n

o

is
Pr
e
t
Sta

fo

li
Ca

Institutions With a Substance Abuse Treatment Program
Sources:  Bureau of State Audits’ (bureau) analysis of data collected from the Department of Corrections and Rehabilitation’s (Corrections)
COMPAS database, Offender Based Information System, and Distributed Data Processing System, and the chief of Corrections’ Data and Performance
Management Unit.
*	 The bureau’s analysis of the number of inmates in an institution that had a moderate or high COMPAS core substance abuse treatment needs
assessment and four to 12 months left to serve as of February 20, 2011.
†	 These numbers were provided to us by the chief of the Data and Performance Management Unit and indicate the number of inmates with
a low or moderate to high substance abuse treatment need according to COMPAS, or that had no COMPAS core assessment on file but were
in a substance abuse treatment program in an institution on February 4, 2011. We did not verify the accuracy of these numbers.

27

28

California State Auditor Report 2010-124

September 2011

COMPAS core assessments may not
be a key factor in whether an inmate
gets into the in-prison substance
abuse treatment program.

Nevertheless, Figure 4 shows a significant number of inmates
participating in the substance abuse treatment program at certain
institutions with either no COMPAS core assessment or one showing
low needs, raising questions as to whether Corrections is maximizing
its use of the COMPAS core assessment to place the right inmates into
the substance abuse treatment program. As a result, COMPAS core
assessments may not be a key factor in whether an inmate gets into the
in-prison substance abuse treatment program, as suggested by Figure 4.
Our review also noted limited monitoring on the part of Corrections’
headquarters staff to ensure that its prison personnel consistently
consider and act on COMPAS core assessments. Given the importance
of COMPAS core assessments to Corrections’ efforts to ultimately
reduce prison overcrowding, we expected that Corrections would have
performed periodic quality control reviews of inmate assignments
into the substance abuse treatment program, such as by periodically
sampling different prisons to verify that COMPAS core assessments
were being used as called for in its placement policies. Instead,
Corrections’ oversight of staff’s use of COMPAS core assessments
is more limited. According to Corrections’ chief of Rehabilitative
Programming and Accountability (rehabilitative accountability chief ),
Corrections’ Adult Programs Division holds quarterly teleconferences
with the correctional staff at the institutions to have them discuss and
explain placement decisions that do not appear to meet requirements.
The basis for the discussion is a chart similar to Figure 4 that shows,
among other things, summary data of the characteristics of inmates
assigned to substance abuse treatment programs in the institutions
that offer such programs, including characteristics such as inmates’
time left to serve and COMPAS’s identified need for substance abuse
treatment. According to the rehabilitative accountability chief, most
explanations are that inmates’ central files reveal a history of substance
abuse. According to Corrections’ director of rehabilitative programs,
an in-depth file review may provide an indication of a substance abuse
treatment need that may not have been captured by the COMPAS
core assessment. However, Corrections’ explanation that its staff are
able to make placement decisions without considering COMPAS core
assessment results, and that such results may not identify a substance
abuse treatment need, only serves to raise more questions as to whether
the COMPAS core assessment is a valuable tool.
Corrections also does not have a plan to evaluate whether the
COMPAS core or reentry assessments will help it reach its ultimate
goal—reducing prison overcrowding and recidivism. When making
the business case for implementing COMPAS at its institutions,
Corrections indicated that it would provide a critical link to an
overall case management system to track the progress of offenders in
their individual rehabilitative programming, which would ultimately
lead to a reduction in recidivism. During the audit we asked the
director of Corrections’ Division of Internal Oversight and Research

California State Auditor Report 2010-124

September 2011

(research director) to explain how Corrections intended to evaluate
whether COMPAS was assisting it in meeting this goal. The research
director indicated that because COMPAS is an assessment tool
designed to identify the needs of offenders, and does nothing to alter
offender behavior, COMPAS alone is unable to have a direct impact
on recidivism. The research director explained that Corrections will
continue to evaluate the impact of rehabilitative programs on recidivism
rates and that it plans to use COMPAS data in future recidivism reports
as a way to describe its population of offenders, similar to how age and
gender are categorized in Corrections’ current recidivism report.
Measuring COMPAS’s effect on Corrections’ ultimate goal of
reducing recidivism would seem critical to ensure that it is a
valuable tool in achieving its intended outcome. To do this,
Corrections would need to collect information on which inmates
were placed in rehabilitative programs as a result of COMPAS and
whether these inmates eventually returned to prison once released.
Also, comparing recidivism rates between groups that received
rehabilitative programming as a direct result of COMPAS versus
those that received programming as a result of other considerations
could provide Corrections with important insight into how much
value COMPAS provides in ensuring that the right inmate gets
into the right program. Although certain inmate information may
not be stored electronically, such as historical data on an inmate’s
participation in or completion of the substance abuse treatment
program, Corrections could consider other approaches, such as
file reviews of a representative sample of inmates, to address these
questions. However, doing so would require a clear documentation
trail showing that COMPAS played a role in Corrections’ decision
to provide an inmate with rehabilitative programming. Corrections’
March 2011 memorandum to its staff regarding COMPAS’s
implementation for its prison institutions discussed the expectation
that classification committees document their decisions to assign
inmates to rehabilitative programs and to cite COMPAS core
assessment scores as appropriate. However, since this is a new policy
and the rollout of COMPAS to the prison institutions is fairly recent,
Corrections may need to monitor whether this is occurring to ensure
that it can ultimately measure COMPAS’s impact.
Corrections Does Not Intend to Use COMPAS to Assess All Inmates
In an effort to manage increased workload for reception center staff
and parole staff, Corrections has excluded certain offenders from
receiving COMPAS core and reentry assessments. As of March 31, 2011,
Corrections had completed COMPAS core assessments in its
receptions centers for nearly 80,000 inmates and COMPAS reentry
assessments for more than 171,000 inmates nearing parole.

29

30

California State Auditor Report 2010-124

September 2011

In January 2007 the Division of Adult Parole Operations (Parole
Operations) issued a memorandum indicating that COMPAS
reentry assessments were available for inmates nearing parole,
with certain exceptions as shown in Table 4. These exclusions were
modified twice, in February 2010 and in December 2010, to help
parole operations attain its goal of completing assessments on
100 percent of offenders eligible to receive one.
Table 4
Offenders Who Have Been Excluded From Receiving a Correctional Offender
Management Profiling for Alternative Sanctions Reentry Assessment
JANUARY
2007*

FEBRUARY
2010

DECEMBER
2010









Mexican national offenders pending deportation to Mexico
who have an active hold with U.S. Immigration and Customs
Enforcement or a prior deportation order





Other offenders with a prior U.S. Immigration and Customs
Enforcement deportation order



EXCLUSIONS

Offenders/parolees serving their entire sentence in a county jail
Enhanced outpatient†
U.S. Immigration and Naturalization Service/deportation cases

Parole violators returned to custody
Civil addicts‡
Interstate cases§
Offenders approved for nonrevocable paroleII
Offenders housed at locations other than prisons













Sources:  Department of Corrections and Rehabilitation procedures guide and policy memoranda.
*	 In January 2007 the Division of Adult Parole Operations issued a memorandum indicating that
Correctional Offender Management Profiling for Alternative Sanctions reentry assessments were
available only for parolees with a new commitment or for parole violators returned to custody.
The memorandum specifically excludes the first three populations listed in this table, but is silent
as to whether the remaining inmate populations are excluded.
†	 The Enhanced Outpatient Program provides care to mentally disordered inmate patients who
would benefit from a therapeutic environment that is less restrictive than inpatient settings. This
includes inmates participating in the Transitional Case Management Program, which prepares
seriously mentally disordered inmate patients for parole.
‡	 The Civil Addict Program provides treatment and rehabilitation to persons convicted of felonies
and misdemeanors whose crimes were attributable to the repeated use of narcotics.
§	 Interstate cases relate to inmates that are transferred from out of state to be supervised in
California, or California inmates being supervised by other states.
II	 Under nonrevocable parole, parolees are not returned to prison for a parole violation.

Certain inmates in the reception centers have also been excluded
from receiving a COMPAS core assessment, as shown in Table 5.
In September 2008 the associate director of the Division of Adult
Institutions notified staff at reception centers that they could exclude
certain inmates while they were familiarizing themselves with the

California State Auditor Report 2010-124

September 2011

COMPAS core assessments, but only if needed to ensure that all
inmates are processed expeditiously through the reception center.
In March 2009 the chief deputy secretary of adult operations issued
a memorandum acknowledging that the learning process and
proficiency in the COMPAS program by reception center counseling
staff would take time and provided an updated list of inmates to be
temporarily excluded from receiving an assessment.

Table 5
Offenders Who Have Been Excluded From Receiving a Correctional Offender Management Profiling for Alternative
Sanctions Core Assessment in the Reception Centers
EXCLUSIONS

SEPTEMBER 2008

Inmates paroling from a reception center with less than 60 days to serve
Inmates with a current or pending security housing unit term*
Inmates with sensitive needs
Inmates designated as enhanced outpatient level of care†
Inmates designated as security housing unit or psychiatric services unit‡
Inmates designated as mental health crisis bed§
Inmates paroling from a reception center with less than 90 days to serve
Inmates designated as low risk (score of 1) on the California Static Risk Assessment
Inmates designated as a parole violator returned to custody with less than 90 days to serve
Inmates with an active U.S. Immigration and Customs Enforcement hold
Inmates designated as condemned
Inmates with an earliest possible future parole date more than 10 years away
Inmates serving a life term with a minimum eligible future parole date more than 5 years away

MARCH 2009

MARCH 2010




















Source:  Department of Corrections and Rehabilitation policy memoranda.
*	 Security housing units house inmates whose conduct endangers the safety of others or the security of the institution.
†	 The Enhanced Outpatient Program (EOP) provides care to mentally disordered inmate patients who would benefit from a therapeutic
environment that is less restrictive than inpatient settings.
‡	 The psychiatric services unit provides secure housing and care for inmates with diagnosed psychiatric disorders who require placement in housing
equivalent to a security housing unit at the EOP level of the mental health delivery system.
§	 The mental health crisis bed provides short-term inpatient care for seriously mentally disordered inmate patients awaiting transfer to a hospital
program or being stabilized on medication prior to transfer to a less restrictive level of care.

According to a March 2010 memorandum, Corrections temporarily
suspended COMPAS assessments between August 2009 and
March 2010 as one of several measures to assist in expediting inmate
processing times at reception centers. The memorandum further
stated that upon resuming COMPAS assessments, temporary
exclusionary criteria for the COMPAS core assessments would be
applied, as shown in the last column in Table 5, in the reception

31

32

California State Auditor Report 2010-124

September 2011

centers in an effort to achieve compliance with state law, while at the
same time maintaining the demanding level of production necessary
to prevent backlogs in the county jails and reception centers.
Tables 6 and 7 show the history of Corrections in conducting
COMPAS core and reentry assessments of inmates for the last
four fiscal years. Table 6 compares the number of inmates in
Corrections’ institutions during fiscal years 2007–08 through 2010–11
to the number who had received a COMPAS core assessment in
each respective fiscal year. Table 7 compares the number of inmates
released to parole during the same period to the number who received
a COMPAS core or reentry assessment prior to parole during each
respective fiscal year. The data in these tables clearly show a significant
gap between the number of inmates under Corrections’ supervision
and the number assessed using COMPAS. However, Corrections
continues to gain ground on assessing inmates with the COMPAS
core assessment, having assessed 0.2 percent of the population in fiscal
year 2007–08 and 26.5 percent partway through fiscal year 2010–11.
Table 6
Inmates Assessed by the Department of Corrections and Rehabilitation
Using the Correctional Offender Management Profiling for Alternative
Sanctions Core Assessment
Fiscal Years 2007–08 Through 2010–11
FISCAL YEAR
2007–08

2008–09

2009–10

2010–11*

Number of inmates who were housed in an
institution or camp during the fiscal year†

276,998

269,960

258,123

208,714

Number of inmates who received at least
one Correctional Offender Management
Profiling for Alternative Sanctions
(COMPAS) core assessment as of the last
day of the fiscal year‡

453

30,174

44,572

55,231

Percentage of inmates who have received
a COMPAS core assessment

0.2%

11.2%

17.3%

26.5%

Sources:  Bureau of State Audits’ analysis of data collected from the Department of Corrections
and Rehabilitation’s COMPAS database, Distributed Data Processing System, and Offender Based
Information System.
*	 Fiscal year 2010–11 includes data for the period July 1, 2010, through February 20, 2011, only.
†	 These numbers represent the number of inmates who spent at least one day in an institution or
camp during each fiscal year.
‡	 If an inmate received a COMPAS core assessment in any one of these fiscal years, that assessment
would be counted for each subsequent year the inmate was incarcerated in an institution or camp.

California State Auditor Report 2010-124

September 2011

Table 7
Parolees Assessed by the Department of Corrections and Rehabilitation
Using a Correctional Offender Management Profiling for Alternative
Sanctions Core or Reentry Assessment
Fiscal Years 2007–08 Through 2010–11
FISCAL YEAR
2007–08

2008–09

2009–10

2010–11*

Number of inmates released to parole
during the fiscal year†

117,881

114,611

107,462

70,643

Number of inmates who received a
Correctional Offender Management
Profiling for Alternative Sanctions
(COMPAS) core or reentry assessment
within the 12-month period before
their first release to parole during the
fiscal year

62,206

57,447

35,986

14,946

Percentage of inmates who have received
a COMPAS core or reentry assessment

52.8%

50.1%

33.5%

21.2%

Sources:  Bureau of State Audits’ analysis of data collected from the Department of Corrections and
Rehabilitation’s COMPAS database and Offender Based Information System.
*	 Fiscal year 2010–11 includes data for the period July 1, 2010, through February 20, 2011, only.
†	 These numbers represent the number of inmates released to parole during the fiscal year. Our
analysis included an inmate’s initial release to parole as well as those inmates who were released to
parole after serving time for a parole violation.

Corrections does not have a timeline for assessing all inmates
and is not clear on its goal for completing assessments of
unexcluded, or eligible, inmates. A policy memorandum issued
in March 2011 announcing the implementation of COMPAS for
general‑population inmates states that all offenders, other than
those with a low risk of reoffending, will have a COMPAS needs
assessment completed within approximately 18 months of that
implementation. However, Corrections’ secretary stated in an
April 2011 letter to the chair of the Senate Rules Committee that
Corrections estimates it will take approximately 18 months—to
September 2012—to have the majority of eligible inmates assessed.
The rehabilitative accountability chief stated that Corrections
plans to complete assessments for 100 percent of all unexcluded
inmates in the next 18 months. However, she stated that she is
still in the process of developing more exclusionary criteria for
general‑population inmates, which could affect Corrections’
timeline for assessing all inmates. Consequently, it is not clear at
this time how many inmates Corrections plans to assess or how
long it will take to assess them.

33

34

California State Auditor Report 2010-124

September 2011

Corrections’ COMPAS Reentry Assessment Tools May Benefit Paroling
Inmates, but Their Effect on Recidivism Is Unclear
The potential benefits from Corrections’ use of the COMPAS
reentry assessment on inmates approaching parole are more readily
apparent than the benefits of using COMPAS in reception centers.
Within 240 days of an inmate’s release to parole, Corrections’
staff administer a COMPAS reentry assessment, which results in
a case plan that can assist the prisoner with transitioning to life
outside of prison. The case plan considers the factors that caused
the inmate to commit a crime in the first place, such as problems
with substance abuse or anger management issues, and provides the
inmate with tasks and goals to pursue once released to address such
factors. The case plan also takes the additional step of providing the
inmate with contact information for programs and resources in
the communities where he or she will be paroled.
The COMPAS reentry assessment also has potential value
for the parole agents who monitor parolees. One output from
performing the assessment is a reentry summary document that
provides useful background information on the inmate, such as
information on substance abuse problems, education level, and
vocational skills. The reentry summary also provides the parole
agent with information obtained from a review of the inmate’s
file as compared to information obtained through an interview,
highlighting inconsistencies for the parole agent to consider when
developing the inmate’s parole plan.
Inmates are not consistently
compelled to follow their COMPAS
case plan as a condition of parole,
and parole agents—those
who monitor parolees—do
not appear to routinely use
COMPAS‑related information.

However, the benefits from administering COMPAS in the larger
context of Corrections’ goal of reducing recidivism is unclear, since
inmates are not consistently compelled to follow their COMPAS
case plan as a condition of parole, and parole agents—those who
monitor parolees—do not appear to routinely use COMPAS‑related
information when developing case plans and supervising parolees.
According to Parole Operations’ associate director, parole agents’
use of COMPAS has been inconsistent, and more training is
required to show them how to apply COMPAS assessments to a
parole plan that will address the inmate’s criminal risk factors and
thereby reduce recidivism. Corrections’ Parole Operations began
conducting a pilot project in August 2010 called the California
Parole Supervision and Reintegration Model (pilot project), which
includes requiring certain parolees, as a condition of their parole, to
address their criminal risk factors as established in their COMPAS
case plans or as directed by their parole agent.
Performing COMPAS reentry assessments can be beneficial to
inmates facing parole. According to one parole services associate
(PSA)—an individual who is responsible for administering the
reentry assessment to prison inmates—the case plan resulting from

California State Auditor Report 2010-124

September 2011

the reentry assessment can reduce the inmate’s stress by allowing
him or her the opportunity to begin planning for rehabilitation
sooner. For example, inmates can use their case plans to begin
contacting resources while still in prison to prepare to meet their
goals even before they meet with their parole agent. PSAs develop
an inmate’s case plan following a file review and an interview.
COMPAS reentry assessments for men contain 45 questions and
those for women contain 126 questions focusing on areas such
as whether the inmate has issues with anger and substance abuse
while also focusing on factors such as whether the inmate will
face housing or employability problems once released. The female
versions of the assessments include additional questions focusing
on personal relationships, parenting issues, and abuse. Once the
questionnaire is completed, the PSA can produce a bar chart from
COMPAS showing the degree to which the inmate has certain
needs that, if not addressed, could cause the individual to commit
another crime. In general, the COMPAS reentry assessment
classifies these criminal risk factors as either unlikely, probable, or
highly probable.
The value to the inmate receiving a reentry assessment is that the
case plan provides a listing of community resources that can be
accessed in response to higher-level needs identified by COMPAS.
According to its policies and procedures, Corrections’ staff are
required to develop a case plan that addresses all highly probable
and probable needs, establishing a minimum of one goal, task, and
service provider for each need. Our review of some of the
COMPAS case plans demonstrated that they appeared to provide
the inmate with specific and useful information. For example,
one case plan we reviewed furnished an inmate with information
on how to find short-term housing by providing an address and
phone number for a homeless shelter. This case plan also identified
resources to treat the inmate’s drug addiction by providing contact
information for various residential and outpatient drug programs.
Finally, the case plan gave the inmate helpful resources and contact
information for finding a job. For example, the case plan told the
inmate where he could go to get his identification card or driver’s
license, obtain his Social Security card, and develop job skills.
Parole agents can also benefit from using COMPAS reentry
assessments. According to Corrections’ June 2009 memorandum
regarding the mandatory use of COMPAS in Parole Operations, the
COMPAS reentry assessment is designed to provide a standardized
and consistent report on an offender’s needs for parole agents and
to propose community-based programs that can best assist the
offender in his or her reintegration efforts. Although Parole
Operations allows its parole agents to review and make changes to

COMPAS reentry assessments for
men contain 45 questions and those
for women contain 126 questions
focusing on areas such as whether
the inmate has issues with anger and
substance abuse.

35

36

California State Auditor Report 2010-124

September 2011

the COMPAS-generated case plan, the associate director of Parole
Operations stated that one of the intended benefits of the case plan
is to help reduce the parole agent’s workload.
However, the value of the COMPAS reentry assessment toward
reducing recidivism is uncertain because paroling inmates are not
required to adhere to the goals and tasks outlined in their COMPAS
case plan. Further, Parole Operations acknowledged that parole
agents’ use of COMPAS has been inconsistent and that more
training is required. Under state regulations, Corrections’ staff
must inform paroled inmates of the conditions they must follow at
least 45 days before they are released to parole. As discussed earlier,
Corrections’ regulations do not discuss how COMPAS is to be used
with its other practices. Our audit found that Parole Operations’
various administrative memoranda concerning the implementation
and use of COMPAS do not discuss how or whether goals or tasks
outlined in the COMPAS case plan can influence the conditions of
parole. Our legal counsel reviewed the June 2009 memorandum and
concluded that it was an unenforceable underground regulation.
We spoke with one parole agent in the Santa Rosa parole office
who indicated he tells parolees that addressing the goals in their
case plans is voluntary, but that failure to participate is noted
in their file and may affect parole decisions made by the Board
of Parole Hearings. A parole administrator who oversees multiple
parole officers echoed this perspective. However, without a process
to monitor that parolees actually address the criminal risk factors
identified on their case plans, the success of COMPAS in helping
Corrections reduce recidivism seems to rely on parolees voluntarily
addressing the factors that caused them to commit a crime.

One parole agent said he usually
does not review the COMPAS case
plan because his knowledge of the
area enables him to recommend
programs and services that are free
or at least more affordable.

There is also uncertainty as to the benefit of the COMPAS reentry
assessment because parole agents do not always see value in or
actually use the case plan that is produced. Some parole agents
we spoke with indicated that they may have more knowledge
of the available programs located in the communities where an
inmate will be paroled. According to one parole agent who works
in Stockton, he usually does not review the COMPAS case plan
because his knowledge of the area enables him to recommend
programs and services that are free or at least more affordable. The
Stockton parole agent explained that the COMPAS case plan often
includes programs and services outside of the parolee’s area or lists
programs that require payment.
A parole unit supervisor in the Stockton office confirmed this
perspective, stating that some parole agents prefer to develop
their own case plans as they believe they know the parolee and
the available resources better than COMPAS does. We also heard
similar statements from a parole agent in the Santa Rosa office, who

California State Auditor Report 2010-124

September 2011

indicated that he makes modifications to the case plan when the
services listed are no longer available, are not accessible, or are not
affordable. Although our audit did not focus on the location and
the affordability of programs cited on COMPAS case plans, our
review of some of these plans revealed that some of the suggested
services do cost money, while others are free. For example,
one parolee’s COMPAS case plan listed two residential sober
living facilities that cost $371 and $700 per month, respectively.
Other services listed on this same parolee’s case plan, such as
organizations that provide emergency and short-term housing,
did not specify whether or not they had a cost.
Parole Operations Has Made Efforts to Improve Staff Perspectives on
COMPAS Reentry
Parole Operations appears to recognize that it has more work to
do to ensure that COMPAS reentry is better understood and used
by its staff. In August 2010 the director of Parole Operations issued
an e-mail to Corrections’ parole staff indicating that COMPAS
reentry was not being used to the fullest extent of its capabilities
and imploring his staff to use the COMPAS reentry assessment
tool to enhance effectiveness. In May 2011 the associate director
of Parole Operations acknowledged that parole agents are not
consistently using the COMPAS reentry assessment and that
although some training has been provided, more in-depth training
is required on criminal risk factors and how to apply the COMPAS
reentry assessment to a parole plan that will address those needs to
ultimately reduce recidivism.
The associate director also pointed to a pilot project currently
underway within Parole Operations that should help improve
parole agents’ use of COMPAS. This pilot project, taking place
at select units within each of its four parole regions, includes
procedures that further explain the use of COMPAS reentry
when supervising parolees. Among the COMPAS-related
procedures in the pilot project are requirements that parolees
who are released after October 2010 have, as a condition of their
parole, a requirement that they work toward completing the tasks
established by their parole agent for addressing their criminal risk
factors. According to the pilot project’s guidelines, these needs
should be derived from the COMPAS reentry case plan or from
the parole agent’s direct observation of the parolee. Further, Parole
Operations has developed a form called the Goals and Progress
Report that requires parole agents to specify goals and how much
time parolees should spend over a certain period of time toward
meeting those goals. For example, one or two months after
receiving direction from a parole agent, the parolee will report
back on the same form indicating how much time was spent on

The associate director also pointed
to a pilot project currently underway
within Parole Operations that should
help improve parole agents’ use
of COMPAS.

37

38

California State Auditor Report 2010-124

September 2011

achieving certain goals, such as looking for a job, attending an anger
management program, or practicing money management. However,
the pilot program operations manual does not specify whether
parole agents are required to verify this reported information.
The pilot project appears to be a positive step in that it establishes
a monitoring process to better ensure that parolees address their
criminal risk factors. However, the ultimate success of COMPAS
reentry assessments relies on Corrections’ staff ’s willingness to
use the assessment tool and their belief that it is a worthwhile
instrument. Similar to our observations of Corrections’ use of
COMPAS core assessments in the reception centers, it appears
that Corrections still has more work remaining to foster greater
acceptance and use of the COMPAS reentry assessment. One way in
which it could have obtained greater acceptance and understanding
of this assessment tool, by both parole agents and external
stakeholders, would have been to develop regulations, as required
by the California Administrative Procedures Act, and sections in
its operations manual that discuss its use. The process for adopting
regulations requires public input from interested stakeholders, such
as parole agents or prisoner advocacy groups, and is intended to
ensure that any adopted rule or regulation can be clearly understood.
Corrections Cannot Demonstrate How Much It Cost to Fully Deploy
and Administer COMPAS

Corrections cannot demonstrate
that it has provided accurate and
complete cost information to the
Technology Agency regarding
its deployment of COMPAS to its
33 prisons and monitored the project
for cost overruns.

Corrections lacks accounting records indicating how much it cost
to deploy and administer COMPAS to its parole units and reception
centers, and acknowledges that it did not establish accounting
codes to track such costs. Further, even though the California
Technology Agency (Technology Agency), which approves and
oversees state technology projects, required Corrections to report
the actual costs of the COMPAS deployment to its prisons, it has
not done so. When we spoke with the COMPAS project’s controller
and technical project manager, they acknowledged that they were
not reporting actual costs to the Technology Agency, stating that
Corrections’ accounting system was not set up to report this
information. We corroborated Corrections’ statements when we
asked it to provide records from its accounting system to support
a sample of actual expenditure amounts recorded on its tracking
spreadsheet (cost tracking workbook). Corrections was unable
to do so in all cases, including $2.1 million in Corrections’ salary
costs. As a result, Corrections cannot demonstrate that it has
provided accurate and complete cost information to the Technology
Agency regarding its deployment of COMPAS to its 33 prisons
and monitored the project for cost overruns. Further, the lack

California State Auditor Report 2010-124

September 2011

of complete expenditure information for all phases of COMPAS
deployment and administration prevents Corrections from
demonstrating accountability for its spending.
Corrections recognized the importance of tracking actual costs
for the deployment and administration of COMPAS to its prisons
when it established its project management plan in 2010, requiring
that the project’s controller be responsible for managing and
reporting on the project’s costs to the executive steering committee
and the Technology Agency. The project management plan also
stated that actual project costs would be validated via invoices or
other written documentation in Corrections’ accounting system.
With this expectation in mind, we asked Corrections to provide
us with the reports from its accounting system that would support
specific expenditure amounts for the COMPAS project. Specifically,
we asked Corrections to support $2.1 million in reported staff
costs associated with the COMPAS project charged during fiscal
year 2009–10 and $1.1 million in other one-time contract costs
from fiscal year 2010–11, as well as other expenditure amounts
exceeding $500,000. In response to our request, Corrections
indicated that it could not provide accounting reports to support its
$2.1 million in staff costs, stating that its accounting system lacked
the functionality required to provide this information. However,
this lack of functionality raises questions as to how Corrections
developed this expenditure amount for staff costs in the first place.
Instead, Corrections simply provided the cost tracking workbook
that we were attempting to verify. For the remaining $1.6 million in
COMPAS costs that we asked Corrections to support, it provided
accounting records and invoices that did not match the amounts
recorded in its cost tracking workbook for the COMPAS project.
For example, when we asked Corrections to provide the accounting
records supporting the $1.1 million in one-time contract costs
reported for fiscal year 2010–11, it provided accounting records
totaling more than $1.2 million. Although Corrections attempted to
explain the discrepancy by indicating it paid a June 2011 invoice that
was not reflected in the cost tracking workbook, such an explanation
seems unlikely, since Corrections’ accounting unit had no record
of a June 2011 invoice. The last invoice Corrections had paid
to Northpointe Institute for Public Management, Incorporated
(Northpointe) during fiscal year 2010–11 was dated May 31, 2011,
for $20,000 in project management fees, which does not fully
explain the discrepancy.
The lack of accurate information in its cost tracking workbook
also results in Corrections providing the Technology Agency with
suspect or inaccurate information regarding COMPAS’s costs.
For example, in its March 2010 report to the Technology Agency,
Corrections reported that it had incurred $4.1 million in one-time
staff costs; however, this amount includes the previously discussed

Corrections indicated that it could
not provide accounting reports to
support its $2.1 million in reported
COMPAS staff costs, stating that
its accounting system lacked the
functionality required to provide
this information.

39

40

California State Auditor Report 2010-124

September 2011

$2.1 million in Corrections’ salary costs that it cannot support.
Further, Corrections also reported to the Technology Agency in the
same month that it had incurred $606,000 in one-time contractor
costs associated with Northpointe. However, although Corrections
had received Northpointe invoices for $952,000 through
March 30, 2010, it had yet to pay any of these invoices. When we
asked Corrections’ technical project manager to explain the source
for the $606,000, she stated that the amounts that Corrections
reports to the Technology Agency include actual expenditures
and amounts it plans to spend in the future based on the invoices
received. However, as evidenced by the previous example, it does
not appear that Corrections has been following this practice.

Corrections reported actual
COMPAS costs of $5.7 million for
fiscal year 2008–09 and $8.9 million
for fiscal year 2009–10, for a total of
$14.6 million in actual COMPAS costs
that it could not explain.

In addition to not being able to report accurate cost information
to the Technology Agency, Corrections also lacks verifiable
information that would indicate how much it has spent to deploy
and administer COMPAS in its reception centers and its various
parole units. As stated in the Introduction, Corrections began
deployment of COMPAS to Parole Operations and its reception
centers in 2006 and 2007, respectively. The associate director of
Parole Operations confirmed that Corrections did not establish
accounting codes to specifically track COMPAS’s implementation,
explaining that these costs were absorbed in other parole-related
budget areas. Additionally, staff from Corrections’ Budget and
Management Branch indicated that Corrections did not establish
accounting codes for costs related to implementing COMPAS
at reception centers because reporting this information was not
necessary for presentation in the annual budget. However, providing
this information was necessary for COMPAS’s deployment to
Corrections’ 33 prison institutions. In the governor’s proposed
budgets for fiscal years 2010–11 and 2011–12, Corrections reported
actual COMPAS costs of $5.7 million for fiscal year 2008–09 and
$8.9 million for fiscal year 2009–10, for a total of $14.6 million in
actual COMPAS costs. When we asked Corrections’ accounting and
budget staff to explain how it arrived at these amounts, the associate
director of the Budget and Management Branch told us that they
were unaware of any COMPAS-specific accounting codes and did
not specifically answer our question regarding how the branch
arrived at the expenditure amounts reported.
Although we were not able to verify Corrections’ costs to deploy
and administer COMPAS in its various reception centers, prison
institutions, and parole units because it does not track these costs,
Table 8 provides Corrections’ budget for deploying COMPAS to
its 33 prison institutions as approved by the Technology Agency.
Given that this budget focuses only on deploying COMPAS to
Corrections’ 33 prison institutions, Table 8 does not include
the potentially significant continuing staffing costs associated
with COMPAS, such as the numerous PSAs that administer

California State Auditor Report 2010-124

September 2011

the COMPAS reentry assessments in Corrections’ 33 prisons.
During fiscal year 2009–10, Corrections had more than 140 PSAs,
who, according to Corrections, spent 72 percent of their time
administering COMPAS reentry assessments. With the midpoint
salary of a PSA being roughly $52,116 per year, this equates to more
than $5.2 million in annual continuing COMPAS costs.
Table 8
Department of Corrections and Rehabilitation’s Current Budget for
Deploying the Correctional Offenders Management Profiling for Alternative
Sanctions to Its Prison Institutions
INFORMATION TECHNOLOGY COSTS

PROJECTED COSTS

One-Time Costs
Staff

$4,339,381

Hardware

687,948

Software purchase/license

177,979

Other contract services

606,000

Data center services
Other
Total one-time costs

2,380
253,400
$6,067,088

Annual Maintenance and Operations Costs
Staff

$761,647

Software maintenance/licenses

409,958

Contract services

428,108

Data center services

132,763

Other

361,500

Total continuing annual maintenance
and operations costs

$2,093,976

Source:  Department of Corrections and Rehabilitation’s special project report submitted to the
California Technology Agency and approved October 2010.

The majority of Corrections’ proposed one-time costs for
deployment of COMPAS to its prisons, as shown in Table 8, are for
its own staff, totaling $4.3 million out of the nearly $6.1 million in
one-time costs. Of these one-time staffing costs, nearly two‑thirds
are for limited-term project management and training staff. The
remaining one-third is for staff within Corrections’ enterprise
information service and information technology divisions. Hardware
for the COMPAS implementation makes up the next largest
portion of one-time costs in the COMPAS budget, representing
11 percent of the budget, or nearly $688,000, which included
additional servers, laptops, and network printers. Following the
deployment of COMPAS to all prisons, Corrections estimates
continuing annual maintenance and operating costs of nearly

41

42

California State Auditor Report 2010-124

September 2011

$2.1 million, which includes $762,000 in staff costs for Corrections’
enterprise information service and information technology
personnel; however, as we mentioned earlier, this amount excludes
costs associated with correctional staff at the general population
prisons and reception centers who will be administering COMPAS
core assessments to inmates. Annual continuing costs also included
$410,000 for software maintenance and licenses provided by
Northpointe. Finally, the budget shown in Table 8 does not reflect
additional software customization costs that Corrections intends
for Northpointe to make to COMPAS. These customizations were
originally budgeted at an additional $549,000 in one-time costs,
which includes activities such as creating a shortened version of
COMPAS for use during inmates’ annual reviews in prison.
Although we could not verify Corrections’ total costs for
deploying COMPAS to its reception centers, prisons, and
parole units, Corrections’ accounting staff were able to provide
verifiable accounting records indicating that Northpointe has
received $4.9 million related to its work on COMPAS, of which
$1.2 million pertained to Corrections’ efforts to deploy COMPAS
to its prison institutions. Further, Corrections provided records
indicating that it had paid $480,000 to the University of California
for a validation study on COMPAS. Although we obtained
evidence of Corrections’ spending on COMPAS amounting to
only $5.4 million, Corrections’ actual spending on COMPAS is
significantly higher due to staffing costs such as for its information
technology personnel and its more than 140 PSAs.
Recommendations
To ensure that the State does not spend additional resources on
COMPAS while its usefulness is uncertain, Corrections should
suspend its use of the COMPAS core and reentry assessments until
it has done the following:
•	 Issued regulations and updated its operations manual to define
how Corrections’ use of COMPAS will affect decision making
regarding inmates, such as clarifying how COMPAS results will
be considered when sending inmates to different prison facilities,
enrolling them in rehabilitative programs to address their criminal
risk factors, and developing expectations for those on parole.
•	 Demonstrated to the Legislature that it has a plan to measure
and report COMPAS’s effect on reducing recidivism. Such a
plan could consider whether inmates enrolled in a rehabilitative
program based on a COMPAS assessment had lower recidivism
rates than those provided rehabilitative programming as a result
of non‑COMPAS factors.

California State Auditor Report 2010-124

September 2011

Once Corrections resumes its use of the COMPAS core and reentry
assessments, it should take the following steps to better ensure that
COMPAS is a valuable inmate assessment and planning tool:
•	 Provide ongoing training to classification staff representatives,
parole agents, and others that may administer or interpret
COMPAS assessment results.
•	 Develop practices or procedures to periodically determine
whether its staff are using COMPAS core or reentry assessments
as intended. Such a process might include performing periodic
site visits to corroborate that COMPAS is being used as required.
•	 Develop practices or procedures to periodically compare the
demand for certain rehabilitative programs, as suggested by
a COMPAS core assessment, to the existing capacity to treat
such needs.
To ensure transparency and accountability for costs associated with
information technology projects such as COMPAS, Corrections
should take the following actions:
•	 Disclose that it lacks accounting records to support certain
COMPAS expenditure amounts it reported to the Technology
Agency and seek guidance on how to proceed with future reporting
requirements for its deployment of the COMPAS core to its
adult institutions.
•	 Develop policies to ensure that accounting or budget management
personnel are involved in the project planning phase of future
information technology projects so that appropriate accounting
codes are established for reporting actual project costs.

43

44

California State Auditor Report 2010-124

September 2011

We conducted this audit under the authority vested in the California State Auditor by Section 8543
et seq. of the California Government Code and according to generally accepted government
auditing standards. Those standards require that we plan and perform the audit to obtain sufficient,
appropriate evidence to provide a reasonable basis for our findings and conclusions based on our
audit objectives specified in the scope section of the report. We believe that the evidence obtained
provides a reasonable basis for our findings and conclusions based on our audit objectives.
Respectfully submitted,

ELAINE M. HOWLE, CPA
State Auditor
Date:	

September 6, 2011

Staff:	

Grant Parks, Audit Principal
John Billington
Alicia Anne Beveridge, MPA
Meghann K. Leonard, MPPA

Legal Counsel:	

Scott A. Baxter, JD

ITAS Staff:	

Michelle J. Baur, CISA, Audit Principal
Sarah Rachael Black, MBA

For questions regarding the contents of this report, please contact
Margarita Fernández, Chief of Public Affairs, at 916.445.0255.

California State Auditor Report 2010-124

45

September 2011

(Agency comments provided as text only.)
August 18, 2011
State of California–Department of Corrections and Rehabilitation
Office of the Secretary
P.O. Box 942883
Sacramento, CA 94283-0001
Ms. Elaine M. Howle*
State Auditor
Bureau of State Audits
555 Capitol Mall, Suite 300
Sacramento, CA 95814
Dear Ms. Howle:
The California Department of Corrections and Rehabilitation (CDCR) is submitting this letter in response to the
Bureau of State Audits’ (BSA) report entitled, “Department of Corrections and Rehabilitation: The Benefits of
Its Correctional Offender Management Profiling for Alternative Sanctions (COMPAS) Program Are Uncertain.”
While CDCR appreciates the effort and resources BSA committed to this report, as well as the professionalism of
BSA’s staff in conducting the review, we are profoundly disappointed in the report’s ultimate recommendation.
We do not believe the audit’s findings support the recommendation to suspend the use of COMPAS.
Furthermore, we are convinced that doing so would severely jeopardize CDCR’s rehabilitative mission.
Following is our initial response to the audit findings and recommendations.

1

2

As the report notes, in 2006, as part of the Budget Act, the Legislature instructed CDCR to convene a panel
of  correctional program experts to make recommendations specific to improving rehabilitative outcomes
and reducing recidivism. As a result of this effort, in June 2007, the Expert Panel on Adult Offender
Reentry and Recidivism Reduction Programs (Expert Panel) released its report to the Legislature entitled, “A
Roadmap for Effective Offender Programming in California.” As part of this report, the Expert Panel recommended
CDCR adopt what they coined the “California Logic Model” to “operationalize these [evidence‑based] principles
and practices so California can deliver a core set of effective rehabilitation programs to its adult offenders.” The
California Logic Model includes eight basic steps – the second of which is to “assess criminogenic needs.”
COMPAS is CDCR’s method of meeting this objective, specifically by assessing an offender’s criminogenic
needs, i.e., the dynamic risk factors that put an offender at risk to reoffend.
CDCR acknowledges COMPAS has not yet been fully implemented. Due to the State’s fiscal crisis and several
recent budget reductions, CDCR has been forced to adjust its implementation plan periodically and readjust
priorities numerous times. However, despite these constraints and challenges, CDCR has continued to make
significant progress in its implementation of COMPAS. For example, since CDCR began implementation of
COMPAS in 2006, as part of the prerelease planning process, the percentage of active parolees with a COMPAS
assessment has increased to approximately 96 percent of active parolees as of November 2010. Since CDCR

*  California State Auditor’s comments begin on page 49.

3
4

46

California State Auditor Report 2010-124

September 2011

Ms. Elaine M. Howle
Page 2

3

5

6

7

8

9

9

10

began implementing COMPAS at the reception centers in 2007, the percent of inmates with a core COMPAS
assessment increased to 20 percent in November 2010. In March 2011, CDCR began implementation of
COMPAS at the general population prisons and in the few short months since then, the percentage of inmates
assessed with a core COMPAS has grown to 25 percent. This progress should not be minimized. Currently,
California is only 1 of 19 states conducting both risk and needs assessments of adult offenders on a statewide
level. If CDCR were to discontinue implementing COMPAS now, this progress, including five years of training,
would be lost and it would take years to regain our momentum.
BSA’s report minimizes the challenges CDCR faces in bringing about a culture change toward evidence‑based
corrections. We do not doubt there are staff throughout our institutions and parole regions who question the
value of COMPAS and believe they can do their jobs just as well without it. We agree CDCR can and should do a
better job of communicating with our staff and others about the value of COMPAS and its role in evidence‑based
efforts to reduce recidivism. We are currently working on training plans for parole and institution staff. We
concur with BSA’s recommendations to continue to provide training, perform periodic site visits, and reassess
capacity based on the information gathered from COMPAS. However, we do not agree that we should suspend
the use of COMPAS. By doing so, CDCR risks losing any ground we have made in our cultural change toward
effective rehabilitative practices. We refuse to return to the method of simply placing an offender in the next slot
available—regardless of their criminogenic needs.
BSA’s report questions COMPAS’s value because CDCR also takes other factors into consideration when
assigning inmates to prisons or placing them into programs. CDCR disagrees with this notion and argues that
any correctional system must consider other factors, such as security risk, and medical/mental health needs
when placing offenders. Doing so does not minimize the importance of considering criminogenic needs as
well. As to why CDCR has not been using COMPAS specifically for assignment into education programs,
Northpointe was only able to distinguish between academic and vocational needs in our most recent release.
This distinction will allow us the flexibility necessary to incorporate these needs into our priority placement
procedures. Ultimately, the point that BSA’s report fails to acknowledge is that until all inmates are assessed,
CDCR will be limited in our ability to fully integrate COMPAS into the program assignment process. Suspending
COMPAS at this point, as BSA recommends, would make it impossible for CDCR to achieve this goal in the near
future and could possibly devastate any future efforts in this area.
In addition, the BSA report appears to confuse inmate needs for needed capacity. Just because a certain number
of offenders have a particular need does not mean CDCR must have enough slots for each of those inmates to
receive services. Determining capacity for particular programs requires taking into consideration the length
of the program (i.e., if a program lasts 90 days, 4 inmates can participate in 1 slot per year), other exclusionary
criteria (i.e., certain inmates may have a need, but are not able to access the program for other reasons), and
the appropriate timeframe for the program (i.e., in order to support a seamless continuum to  aftercare or
post‑releases services, some programs are best offered immediately prior to release). By failing to acknowledge
these other factors, BSA has oversimplified its data analysis and come to a misleading conclusion.
CDCR strongly disagrees with BSA’s assertion that the value of COMPAS is lessened because of CDCR’s limited
rehabilitative programs, and feels strongly that the opposite is true. Our limited funding makes it even more

California State Auditor Report 2010-124

47

September 2011

Ms. Elaine M. Howle
Page 3

critical that we do not waste precious resources by placing inmates into the wrong program or even into the
right program at the wrong time. Furthermore, our COMPAS assessments will provide valuable information in
helping us to determine how to best allocate additional funding when it becomes available in the future. If we
suspend COMPAS now, we may lose this ability.
BSA’s claim that CDCR does not intend to use COMPAS to assess all inmates is also misleading. In order to
maximize the effectiveness of COMPAS while phasing in implementation (as noted earlier, CDCR only began
conducting COMPAS assessments at the general population prisons in March 2011), CDCR is currently
prioritizing the assessment of those offenders who are most likely to need programs first, i.e., those with
shorter lengths of stay who do not meet other exclusionary criteria. Once this population is assessed, CDCR
intends to broaden its use of COMPAS until eventually all inmates will be assessed by COMPAS.
CDCR is pleased BSA acknowledged the value of our new parole model, the California Parole Supervision and
Reintegration Model, which began implementation in late 2010 and includes new supervision strategies as
well as the integration of COMPAS into field level operations. This new model is scheduled to be expanded
to an additional 20 parole units statewide beginning in October 2011, with the next roll-outs scheduled
every six months thereafter. Given BSA’s endorsement, we do not understand why BSA would recommend
suspending COMPAS, which would put this new model at risk.
BSA recommends CDCR should not implement COMPAS until it has a plan to measure COMPAS’s affect on
recidivism reduction. COMPAS is an assessment tool; it is not a treatment modality, as characterized in the
report. By design, this tool, in and of itself, will never be able to alter offender behavior. To impact behavior,
appropriate treatment must be provided.  The COMPAS assessment is a critical step to identifying treatment
needs, but only the treatment, and not the assessment, will impact recidivism. CDCR plans to use the COMPAS
assessment data as one component of many within an evaluation framework to assess the effectiveness of
CDCR rehabilitative programs in future recidivism reports. BSA’s recommendation to suspend use of COMPAS
would make such future evaluation plans impossible.
Regarding the transparency and accountability for costs associated with COMPAS, BSA suggests CDCR has not
appropriately tracked COMPAS related expenses. The primary justification for this assumption appears to be
that specific accounting codes were not initially established for COMPAS and that CDCR is unable to determine
all of the staffing costs associated with COMPAS. Given the large number of projects CDCR has going at any
given point in time and the fact projects like COMPAS cross many different program areas and incorporate
various efforts, it is not surprising a system as large and (until recently) as antiquated as CDCR’s was not set up
to track all expenditures or portions of expenditures across program areas for every single project. That being
said, however, CDCR does agree a different approach for reporting project-related expenditures is necessary
and we are working on enhancements to establish proper codes to track all future expenditures for projects,
such as COMPAS. With respect to the staffing costs associated with COMPAS, CDCR believes BSA has drawn
a misleading conclusion. Specifically, CDCR does not track the portion of staff time, particularly among the
correctional counselors, parole service associates, and subject matter experts, dedicated to COMPAS because
this duty is integrated within their broader job duties.

11

12

13

13

14

15

48

California State Auditor Report 2010-124

September 2011

Ms. Elaine M. Howle
Page 4

1

While we disagree with the report’s ultimate conclusion, we appreciate BSA’s work on this report and will
address the specific recommendations in a corrective action plan at 60-day, 6-month, and 1-year intervals. If
you have questions regarding CDCR’s response or would like to further address our above concerns, please
contact me at (916) 323-6001.
Sincerely,
(Signed by: Scott Kernan)
SCOTT KERNAN
Undersecretary (A), Operations

California State Auditor Report 2010-124

September 2011

Comments
CALIFORNIA STATE AUDITOR’S COMMENTS ON THE
RESPONSE FROM THE DEPARTMENT OF CORRECTIONS
AND REHABILITATION
To provide clarity and perspective, we are commenting on the
Department of Corrections and Rehabilitation’s (Corrections)
response to our audit. The numbers below correspond to the
numbers we have placed in the margin of Corrections’ response.
We are disappointed that Corrections did not respond to our specific
recommendations, instead stating only that it believes the audit’s
findings do not support a recommendation to suspend the use of
the Correctional Offender Management Profiling for Alternative
Sanctions (COMPAS). Corrections indicated that it plans to provide
more detailed responses regarding our specific recommendations
in its 60-day, six-month, and one-year responses. We look forward
to providing the Legislature with an update on Corrections’ efforts
in early 2012. In the meantime, we will provide the Joint Legislative
Audit Committee with Corrections’ 60‑day and six-month response
to our audit report’s recommendations.

1

We stand by the findings and conclusions reached in our audit
report. Our recommendation on page 42 that Corrections
temporarily suspend COMPAS is based on Corrections’ failure
to follow the public rulemaking process and issue regulations
defining how COMPAS’s use will affect inmates and those serving
parole. Corrections has a legal obligation to issue such regulations
under the Administrative Procedures Act as we state on page 21
of the audit report. Furthermore, as we state on that same page,
had Corrections solicited public input from its stakeholders and
adopted regulations, it might have been in a better position to
ensure that its application of COMPAS was legally enforceable,
understood by its employees, and would have a demonstrable effect
on inmate decision making.

2

We question Corrections’ contention that it made significant
progress in implementing COMPAS. As we describe on pages 26
through 29 of this report, although Corrections administers
COMPAS assessments to inmates, it does not consistently use those
assessments to make program placement decisions. Further, as we
describe on page 36, paroling inmates are not required to adhere
to the tasks and goals outlined in their COMPAS case plans and
parole agents’ use of COMPAS have been inconsistent. Although
Corrections began using COMPAS in 2006, the fact that much
more remains to be done calls into question how much progress
Corrections has actually made. Simply administering COMPAS

3

49

50

California State Auditor Report 2010-124

September 2011

assessments is different than actually using them to place inmates in
rehabilitative programs that affect positive outcomes in inmates
and parolees.
4

In its response, Corrections indicated that as of November 2010,
96 percent of its active parolees had received COMPAS assessments,
which varies significantly from the assessment data we present in
Table 7 on page 33 of this report. When we asked Corrections how
it arrived at 96 percent, the associate director of parole operations
explained that Corrections limited its analysis to a subset of its
parolees, such as those parolees with active parole status who were
not excluded from receiving a COMPAS assessment, as described
in Table 4 on page 30. In contrast, our methodology focused on all
offenders that were released to parole. Furthermore, Corrections
considered any COMPAS assessment the offender had received
regardless of when it was conducted, whereas we only considered
COMPAS assessments conducted within 12 months of the offender’s
first release to parole for each of the fiscal years we analyzed. We
believe our approach is more reasonable and we stand by the data
presented in Table 7.

5

Corrections asserts that our audit report minimizes the challenges
it faces in bringing about culture change towards evidence-based
corrections. To clarify, our audit focused on Corrections’ practices
for using COMPAS and not its culture.

6

Corrections’ response incorrectly implies that our audit
recommendation to temporarily suspend COMPAS assessments
will cause it to return to its old methods of placing an offender in
the next slot available for rehabilitative programs, regardless of
an inmate’s needs. Corrections has other assessment tools to identify
inmate needs for certain rehabilitative programs. For example,
Corrections currently uses the Test of Adult Basic Education for
placement into its academic programs, while inmates’ problems
with substance abuse could be identified through a review of their
arrest histories. Finally, it is questionable whether Corrections
has substantially changed from the “old methods” it describes in
its response since, as we state on page 18, COMPAS does not play
a significant role in reception centers when determining where
inmates will be housed, and by extension the rehabilitative programs
they might access at their assigned prison facilities. Figure 4 on
page 27 also shows a significant number of inmates accessing the
in-prison substance abuse program either without a COMPAS
assessment or demonstrating a low need for this program.

7

Corrections has misread our report. We do not question the value
of assessing an inmate’s criminal risk factors (i.e. criminogenic
needs) for placement into rehabilitative programs. Instead, as
we state on pages 17 and 18 of the audit report, state regulations

California State Auditor Report 2010-124

September 2011

already require Corrections’ staff to consider many factors—
such as an inmate’s security level or escape risk—when making
prison placement decisions, and by extension the rehabilitative
programs inmates may access at those prisons. By recommending
that Corrections suspend COMPAS until it issues regulations, we
intend for Corrections to determine both how best to implement
COMPAS given these other factors, and to ultimately decide
whether COMPAS can be implemented in a meaningful way.
Corrections’ response states the audit report failed to acknowledge
that until all inmates are assessed, Corrections will be limited in its
ability to fully integrate COMPAS into its rehabilitative program
assignment process.  However, it is unclear what Corrections
believes we failed to acknowledge, because it did not explain what
it means by “limited in its ability to fully integrate COMPAS.”
Nevertheless, our audit report correctly points out that Corrections
lacks regulations defining its use of COMPAS on inmates and
parolees, lacks capacity in its in-prison rehabilitative programs
that reflect COMPAS-determined needs, and lacks a timeline for
assessing all inmates and parolees with COMPAS. Furthermore,
our audit report acknowledges that Corrections excludes numerous
types of inmates and parolees from receiving COMPAS assessments
in order to accommodate Corrections’ workload.

8

Corrections is wrong. Its response states that the audit report
appears to confuse inmate need for rehabilitative programs with
treatment capacity, stating that just because a certain number of
inmates have a particular need does not mean Corrections must
have treatment slots for each inmate. Our report makes no such
conclusion or implication. Instead, our analysis of the data in Table 3
on page 23 considered inmates with certain rehabilitative needs
based on their amount of time left to serve given Corrections’ own
program assignment policies and our consideration of the length
of these programs. Further, Table 3 reflects the annual capacity of
these programs, not the number of program slots. Finally, although
Corrections incorrectly claims that we have oversimplified our
analysis and reached a misleading conclusion, we note that
Corrections’ own response neither specifically critiques our analysis
nor offers its own perspective on what a correct conclusion would be.

9

We stand by our conclusion that the value of COMPAS is
diminished when there are few rehabilitative programs to
address COMPAS‑identified needs. Although Corrections claims
that COMPAS can help it determine where to invest resources in
its rehabilitative programs, it appears that it has not been using
COMPAS for this purpose based on Table 3 on page 23, which shows
that the annual capacity for academic and vocational education
programs is higher than the annual capacity for substance abuse
treatment, even though more inmates have COMPAS‑identified

10

51

52

California State Auditor Report 2010-124

September 2011

substance abuse treatment needs. Further, as evidenced by Figure 4
on page 27, Corrections assigns inmates with low need according to
COMPAS into its substance abuse treatment programs instead of
only assigning inmates that COMPAS assesses as having a moderate
to high need. This practice stands in contrast to Corrections’
assertion that COMPAS assessments allow staff to assign the right
inmates to the right programs at the right time.
11

We disagree with Corrections’ assertion that our conclusion
is misleading. As noted on page 33 of our report, Corrections
expressed a plan as of April 2011 to assess the majority of inmates
not already excluded from COMPAS assessment by approximately
September 2012. As we show in tables 4 and 5 on pages 30 and
31 of our report, Corrections has progressively increased the
number of inmate subgroups that are excluded from receiving
COMPAS assessments. In addition, Corrections indicated on
page 33 of our report that it intends to exclude even more inmates
from receiving a COMPAS assessment.

12

Corrections questions why it should temporarily suspend
COMPAS reentry assessments given our positive comments
about its pilot project in its Division of Adult Parole Operations
(Parole Operations). Our reasons for recommending a temporary
suspension of COMPAS reentry assessments include Corrections’
lack of regulations as explained in comment #2. Furthermore,
for context, the pilot project is being followed by relatively
few parole offices and does not represent a systemwide process for
Parole Operations. Moreover, Parole Operations pilot project has
not concluded. We look forward to hearing about the results of the
pilot project in Corrections’ periodic responses to our report.

13

Corrections’ response takes a narrow view of how it can assess the
value of COMPAS and incorrectly asserts that the audit report
characterizes COMPAS as a treatment program. On page 1 of the
audit report, we state that COMPAS is a software tool used to
help identify factors that cause inmates to commit crimes, so they
can participate in certain rehabilitative programs to reduce their
likelihood of reoffending. Furthermore, our recommendation on
page 42 that Corrections develop a methodology for measuring
COMPAS’s effect on recidivism, prior to resuming COMPAS,
is intended to ensure that Corrections can demonstrate to the
Legislature and the public that COMPAS’s effectiveness will be
measured. On page 29 of the audit report, we describe one potential
methodology Corrections could follow.

14

Corrections incorrectly states that we based our conclusion that
it has not appropriately tracked COMPAS‑related expenses based
on assumptions. Our conclusion is based on assertions from the
COMPAS project controller and technical project manager that

California State Auditor Report 2010-124

September 2011

they were not reporting actual costs of COMPAS to the California
Technology Agency as required because Corrections’ accounting
system was not set up to report this information as stated on
page 38 of the report. Further, we corroborated Corrections’
assertion by asking for accounting records supporting a sample of
COMPAS expenditures and reviewing those documents. As we
describe on pages 39 and 40, Corrections was unable to support
some of these amounts.
It is not clear why Corrections believes we provided a misleading
conclusion regarding its failure to track staffing costs associated with
COMPAS. As we indicate in our report on page 39, Corrections
reported $2.1 million in staff salaries; however, by its own
admission, Corrections was unable to provide any accounting records
or other support demonstrating how it arrived at this amount.

15

53

54

California State Auditor Report 2010-124

September 2011

cc:	

Members of the Legislature
Office of the Lieutenant Governor
Milton Marks Commission on California State
Government Organization and Economy
Department of Finance
Attorney General
State Controller
State Treasurer
Legislative Analyst
Senate Office of Research
California Research Bureau
Capitol Press