Skip navigation

Adult Male Svori Participants Reentry Evaluation Dec 2009

Download original document:
Brief thumbnail
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
Prisoner Reentry Experiences
of Adult Males: Characteristics,
Service Receipt, and Outcomes
of Participants in the SVORI
Multi-site Evaluation
THE MULTI-SITE EVALUATION OF THE SERIOUS AND VIOLENT OFFENDER REENTRY INITIATIVE

December 2009

Pamela K. Lattimore

Danielle M. Steffey

RTI International
3040 East Cornwallis Road
Research Triangle Park, NC 27709
Phone: (919) 485-7759
Fax: (919) 541-5985
Lattimore@rti.org

RTI International
3040 East Cornwallis Road
Research Triangle Park, NC 27709
Phone: (919) 485-7759
Fax: (919) 541-5985
Steffey@rti.org

Christy A. Visher
University of Delaware/The Urban Institute
Center for Drug and Alcohol Studies
77 East Main Street
Newark, DE 19716
Phone: (302) 831-6921
Fax: (302) 831-3307
Visher@udel.edu
This project was supported by Grant No. 2004-RE-CX-002 awarded by the National Institute of Justice, Office of Justice Programs, U.S.
Department of Justice. Points of view in this document are those of the authors and do not necessarily represent the official position or policies
of the U.S. Department of Justice.

Acknowledgments
The Multi-site Evaluation of the Serious and Violent Offender
Reentry Initiative (SVORI) was supported by grant number
2004-RE-CX-002 from the National Institute of Justice (U.S.
Department of Justice) and was conducted by RTI International
and the Urban Institute. Points of view are those of the authors
and do not necessarily represent those of the U.S. Department
of Justice.
Principal Investigators
Pamela K. Lattimore, RTI International
Christy A. Visher, University of Delaware and Urban Institute
Report Authors
Pamela K. Lattimore, RTI International
Danielle M. Steffey, RTI International
Christy A. Visher, University of Delaware and Urban Institute
Staff Contributors
Susan Brumbaugh, RTI International
Alexander Cowell, RTI International
Debbie Dawes, RTI International
Christine Lindquist, RTI International
Mark Pope, RTI International
Laura Winterfield, Urban Institute
We also acknowledge the contributions of the site liaisons from
RTI and the Urban Institute, who documented the
implementation of SVORI programming across the sites and
facilitated data collection for the impact study. In addition, we
are grateful for the hard work and dedication shown by our field
interviewers, supervisors, and data collection task leader
throughout the data collection period.

iii

RTI and the Urban Institute thank the SVORI project directors,
other program and research staff from the SVORI sites, and
staff at the facilities where interviews were conducted. We
greatly appreciate the assistance and support received from
these individuals.
Finally, RTI and the Urban Institute acknowledge the invaluable
assistance and direction provided by the members of our
external advisory group.
For more information about the SVORI Multi-site Evaluation,
please visit our Web site at http://www.svori-evaluation.org/.

iv

Abstract
Statement of Purpose
The Serious and Violent Offender Reentry Initiative (SVORI)
funded 69 agencies in 2003 to develop programs to improve
criminal justice, employment, education, health, and housing
outcomes for released prisoners. These programs were to
conduct assessments and provide participants programs and
services during and after incarceration. The SVORI Multi-site
Evaluation was funded by the National Institute of Justice to
examine the extent to which SVORI program participation
improved access to appropriate, comprehensive, integrated
services and resulted in better outcomes.
Research Subjects
This report presents findings for the adult male participants in
12 programs selected for the impact evaluation (863 SVORI
participants; 834 comparison men). The study participants had
extensive criminal and substance use histories, low levels of
education and employment skills, and high levels of need
across a range of services (e.g., education, driver’s license,
substance abuse treatment, job training and employment).
Study Methods
The impact evaluation included interviews 30 days pre-release
and 3, 9, and 15 months post-release. Data from state
agencies and the National Criminal Information Center
documented post-release recidivism. Propensity score
techniques were used to improve the comparability between
the SVORI and non-SVORI groups. Weighted analyses
examined the treatment effects of SVORI program
participation.
Major Findings
The report provides evidence that SVORI program participation
increased receipt of services and programming. Program
participants were significantly more likely, e.g., to have reentry

v

plans, met with someone and participated in programs to
prepare for release, participated in programs to reduce criminal
thinking, participated in employment and education programs,
and received a needs assessment—although levels of provision
for most services fell short of 100% and declined substantially
following release.
Modest improvements were observed for some outcomes in the
housing, employment, substance use, and criminal behavior
domains. SVORI appeared to have a positive impact on
abstinence from drug use, although drug use was quite high for
both groups and increased across the follow-up periods. The
men enrolled in SVORI programs were less likely to report
perpetrating violence and engaging in criminal behavior and to
have an officially recorded rearrest, although these differences
were not statistically significant. There was no difference in the
proportions reporting being booked in jail or reincarcerated.
Administrative data confirmed no difference in reincarceration
rates—with about 40% of both groups reincarcerated within 24
months.
Conclusions
Although SVORI programs were successful in increasing the
types and amounts of needs-related services provided prior to
and after release from prison, the proportion of individuals who
reported receiving services was less than reported need and,
generally, less than the expectations of the SVORI program
directors. This finding is consistent with SVORI programs that
were developing and implementing their programs and provides
a reminder that starting complex programs may require
sustained effort over several years to reach full
implementation.
Service delivery declined following release. Thus, overall, the
programs were unable to sustain support to individuals during
the critical, high-risk period immediately following release. This
decline may be due to the programs’ difficulty identifying and
coordinating services for individuals released across wide
geographic areas and, again, suggests the need for sustained
effort to reach full implementation.
SVORI program participation resulted in modest improvements
in intermediate outcomes at levels consistent with findings from
meta-analyses of single-program efforts (e.g., 10% to 20%). If

vi

the underlying model that links services to improved
intermediate outcomes that in turn improve recidivism is
correct, the level of improvement in these intermediate
outcomes may have been insufficient to result in observable
reductions in recidivism.
Additional analyses are planned to determine whether there are
specific programs or subgroups associated with positive
outcomes and to examine the relationship between receipt of
specific services and outcomes.

vii

Contents
Section

Page

Executive Summary

ES-1 

Introduction

1 

Research on Prisoner Reentry ..................................... 4 
SVORI ....................................................................16 
Adult Impact Site SVORI Programs .............................19 
Multi-site Evaluation design .......................................22 
Characteristics of the SVORI and Non-SVORI
Comparison Respondents

35 

Demographic Characteristics ......................................35 
Housing ..................................................................39 
Family and Children ..................................................39 
Substance Use and Physical and Mental Health .............42 
Employment History and Financial Support ..................50 
Criminal History, Violence, Victimization, and Gang
Involvement .......................................................54 
In-Prison Experiences ...............................................56 
Summary ................................................................62 
Self-reported Service Needs

65 

Weighted Service Need Bundle Scores across
Waves 1 through 4 ..............................................68 
Employment/Education/Skills Services ........................71 
Transition Services ...................................................72 
Health Services ........................................................76 
Domestic Violence Services ........................................76 
Child Services ..........................................................78 
Levels of Need Across Services ...................................80 
Self-reported Service Receipt

83 

Weighted Service Need Bundle Scores Across
Waves 1 through 4 ..............................................86 
Coordination Services ...............................................89 

ix

Employment/Education/Skills Services ........................92 
Transition Services ...................................................94 
Health Services ........................................................97 
Domestic Violence Services ........................................99 
Child Services ..........................................................99 
Summary and Conclusions ....................................... 100 
Outcomes

105 

Housing ................................................................ 107 
Employment .......................................................... 113 
Family, Peers, and Community Involvement ............... 117 
Substance Use and Physical and Mental Health ........... 121 
Criminal Behavior and Recidivism ............................. 127 
Taking a Broad View: Outcomes Over Domains and
Time ............................................................... 134 

x

Conclusions, Policy Implications, and Future Work

139 

References

147 

Appendix A. Data Tables

A-1 

Exhibits
Exhibit Number
ES-1. Completed interviews by wave, group, and site

ES-6 

ES-2. Use of specific substances during the 30 days
prior to incarceration, by group

ES-9 

ES-3. Criminal history of respondents, by group

ES-9 

ES-4. Weighted average super bundle scores by group
for Waves 1 through 4

ES-11 

ES-5. Weighted average service receipt super bundle
scores by group for Waves 1 through 4

ES-13 

1.

SVORI program logic model and evaluation
framework

17 

2.

Outcome foci among adult impact and non–impact
sites

20 

3.

Mean proportion of SVORI program participants
receiving pre-release and post-release services in
adult program impact sites (as reported by
program directors)

21 

4.

Adult male sample sizes, by state and group

25 

5.

Completed interviews by wave, group, and site

27 

6.

t-statistics comparing means of SVORI and nonSVORI groups

29 

7.

Balance checks for Wave 1 data based on
propensity score weighted regression of the
variable on a SVORI indicator

33 

ˆ distributions for adult male
Characteristics of p
SVORI and non-SVORI evaluation participants

34 

9.

Demographic characteristics of respondents at
time of interview, by group

36 

10.

Age at time of interview, by site and group

37 

11.

Race (white or black), by site and group

38 

12.

Completed 12th grade or obtained a GED, by site
and group

39 

8.

xi

Page

13.

Percentages of fathers reporting on child care or
child support responsibilities, by group

40 

14.

Criminal history and substance use of family and
peers, by group

41 

15.

Lifetime substance use, by group

43 

16.

Lifetime use of cocaine, heroin, and hallucinogens,
by site and group

44 

17.

Substance use during the 30 days prior to
incarceration, by site and group

45 

18.

Use of specific substances during the 30 days
prior to incarceration, by group

45 

19.

Any substance use treatment prior to current
incarceration, by site and group

46 

20.

Lifetime health problems, by group

47 

21.

Current health problems, by group

48 

22.

Average scores on Brief Symptom Inventory
subscales, by group

49 

23.

Employment prior to incarceration, by group

51 

24.

Employment during the 6 months prior to
incarceration, by site and group

51 

25.

Characteristics of respondents’ jobs prior to
incarceration, by groupa

52 

26.

Sources of income during the 6 months prior to
incarceration, by employment status and group

54 

27.

Criminal history of respondents, by group

55 

28.

Conviction offenses for current incarceration, by
group

56 

29.

Average duration of incarceration at time of
interview, by site and group

57 

30.

Disciplinary infractions and administrative
segregations during current incarceration, by
group

58 

31.

Institutional employment, by site and group

59 

32.

Work-release participation, by site and group

60 

33.

Frequency of in-prison contact with family
members and friends, by group

61 

34.

Amount of contact with family members and
friends at time of interview compared with contact
when first incarcerated

62 

35.

Pre-release service need bundle scores across
service bundles, by group

67 

36.

Most commonly reported service needs prerelease, by group

67 

Weighted average super bundle scores by group
for Waves 1 through 4

69 

37.

xii

38.

Weighted average service bundle scores by type
(Employment/Education/Skills, Transition
Services, Health Services), group (SVORI, nonSVORI), and wave (1, 2, 3, 4)

70 

Weighted average super bundle scores by type
(Child Services, Domestic Violence, group (SVORI,
non-SVORI), and wave (1, 2, 3, 4)

71 

40.

Weighted means for employment/education/skills
bundles and items, by group and wave

73 

41.

Weighted means for transition services bundles
and items, by group and wave

74 

42.

Weighted means for health services bundles and
items, by group and wave

77 

43.

Weighted means for domestic violence services
bundles and items, by group and wave

78 

44.

Weighted means for child services bundles and
items, by group and wave

79 

45.

Pre-release service receipt bundle scores across
service bundles, by group

85 

46.

Weighted average service receipt super bundle
scores by group for Waves 1 through 4

86 

47.

Weighted average service need and receipt super
bundle scores by group for Waves 1 through 4

87 

48.

Weighted average service receipt bundle scores by
type, group, and wave

88 

49.

Weighted average service receipt bundle scores by
type, group, and wave

89 

50.

Weighted means for coordination services bundles
and items, by group and wave

91 

Weighted means for employment/education/skills
service receipt bundles and items, by group and
wave

93 

Weighted means for transition services receipt
bundles and items, by group and wave

95 

53.

Weighted means for health services receipt
bundles and items, by group and wave

98 

54.

Weighted means for domestic violence services
receipt bundles and items, by group and wave

100 

55.

Weighted means for child services bundles and
items, by group and wave

101 

56.

SVORI Program Model

106 

57.

Self-reported homeless, living in a shelter, or
without a set place to live, by group and wave

108 

58.

Percentages living with mothers or fathers postrelease, by group (SVORI and non-SVORI) and
post-release follow-up wave (2, 3, and 4)

109 

39.

51.

52.

xiii

59.

Percentages living with partners, spouses, or
children post-release, by group (SVORI and nonSVORI) and post-release follow-up wave (2, 3,
and 4)

110 

60.

Weighted means and regression parameter
estimates for housing outcomes

111 

61.

Percentage reporting living in own house or
apartment; weighted means by group and followup wave

112 

Weighted means and regression parameter
estimates for employment outcomes

115 

Weighted means for self-report that job provides
insurance or paid leave, by group (SVORI, nonSVORI) and data collection wave (2, 3, or 4)

116 

Weighted means for self-report of supporting self
with a job, by group and data collection wave

117 

65.

Marital status and intimate partnerships by group
and wave

119 

66.

Percentage reporting civic action since release/last
interview

121 

67.

Weighted means and regression parameter
estimates for substance use outcomes

123 

68.

Percentages of each group who reported no drug
use since release/last interview and had no
positive drug test

124 

Percentages of each group who reported no drug
use in the past 30 days and had no positive drug
test

124 

70.

Weighted means and regression parameter
estimates for mental health outcomes

126 

71.

Weighted means and regression parameter
estimates for core self-report recidivism outcomes

128 

72.

Weighted average reports of committing no crimes
since release/last interview

129 

73.

Weighted average reports of committing no
violent crimes/no weapons since release/last
interview

130 

74.

Weighted average reports of not reincarcerated at
interview

131 

75.

Weighted average reports of not reincarcerated at
interview and no jail/prison stay of more than 24
hours since release/interview

131 

76.

Official measures of recidivism

133 

77.

Cumulative rearrest rates by group

134 

78.

Cumulative reincarceration rates by group

134 

62.
63.

64.

69.

xiv

79.

Odds ratios from propensity score weighted
logistic regressions of 98 Wave 2 (3-month)
outcomes as a function of SVORI program
participation

135 

Odds ratios from propensity score weighted
logistic regressions of 93 Wave 3 (9-month)
outcomes as a function of SVORI program
participation

136 

Odds ratios from propensity score weighted
logistic regressions of 98 Wave 4 (15 month)
outcomes as a function of SVORI program
participation

136 

Summary results of odds ratios from propensity
score weighted logistic regressions of three waves
of outcomes as a function of SVORI program
participation

137 

A-1.

Adult male case disposition—Wave 1 (pre-release)

A-1 

A-2.

Respondent characteristics, by group (prerelease)

A-2 

A-3.

Proportion of respondents who reported needing
specific services, by group (pre-release)

A-11 

A-4.

Proportion of respondents who reported receiving
specific services, by group (pre-release)

A-12 

A-5.

Proportion of respondents who reported living with
the indicated individuals in the period since the
last interview

A-14 

80.

81.

82.

xv

Executive Summary

SVORI responded to
emerging research
findings that suggested
providing individuals
with comprehensive,
coordinated services
based on needs and risk
assessments could result
in improved post-release
outcomes.

In 2003, the U.S. Department of Justice (DOJ), Department of
Labor (DOL), Department of Education (DOEd), Housing and
Urban Development (HUD), and Department of Health and
Human Services (HHS) provided more than $100,000,000 in
grant funds to states to develop, enhance, or expand programs
to facilitate the reentry of adult and juvenile offenders returning
to communities from prisons or juvenile detention facilities. The
Serious and Violent Offender Reentry Initiative (SVORI) funded
agencies to develop programs to improve criminal justice,
employment, education, health, and housing outcomes for
released prisoners. Sixty-nine agencies received federal funds
($500,000 to $2,000,000 over 3 years) to develop 89
programs. The initiative responded to emerging research
findings that suggested that providing individuals with
comprehensive, coordinated services based on needs and risk
assessments could result in improved post-release outcomes.
Grantees were to use their SVORI funding to create a threephase continuum of services for returning serious and/or
violent prisoners that began during the period of incarceration,
intensified just before release and during the early months
post-release, and continued for several years after release as
former inmates took on more productive and independent roles
in the community.
The initiative imposed relatively few restrictions on grantees.
The criteria for programs funded by SVORI grants were the
following:
ƒ

Programs were to improve criminal justice, employment,
education, health (including substance abuse and mental
health), and housing outcomes.

ƒ

Programs were to include collaborative partnerships
between correctional agencies, supervision agencies,

ES-1

Prisoner reentry experiences of adult males

other state and local agencies, and community and
faith-based organizations.
ƒ

Program participants were to be serious and/or violent
offenders.

ƒ

Program participants were to be 35 years of age or
younger.

ƒ

Programs were to encompass three stages of reentry—in
prison, post-release on supervision, and postsupervision.

ƒ

Needs and risk assessments were to be used to guide
the provision of services and programs to participants.

In some cases, grantees asked for and received permission for
exceptions to these criteria. For example, some programs were
primarily post-release programs, and age restrictions were
sometimes lifted (e.g., for programs targeting sex offenders).
The SVORI programs attempted to address the initiative’s goals
and provide a wide range of coordinated services to returning
prisoners. Although SVORI programs shared the common goals
of improving outcomes across various dimensions and
improving service coordination and systems collaboration,
programs differed substantially in their approaches and
implementations (Lindquist, 2005; 2005; Winterfield,
Lattimore, Steffey, Brumbaugh, & Lindquist, 2006; Winterfield
& Lindquist, 2005).
In spring 2003, the National Institute of Justice awarded RTI
International, a nonprofit research organization, a grant to
evaluate programs funded by SVORI. The Urban Institute, a
nonpartisan economic and social policy research organization,
collaborated on the project. With data collected from grantee
staff, partnering agencies, and returning prisoners, the 6-year
evaluation involved an implementation evaluation of all 89
SVORI programs, an intensive impact evaluation of 12 adult
and 4 juvenile programs, and an economic analysis of a subset
of the impact sites (see Lattimore, Visher, Winterfield,
Lindquist, & Brumbaugh, 2005). The goal of the SVORI
evaluation was to document the implementation of SVORI
programs and determine whether they accomplished SVORI’s
overall goal of increasing public safety by reducing recidivism
among the populations served. The SVORI evaluation was
designed to answer the following research questions:
ƒ

ES-2

To what extent did SVORI lead to more coordinated and
integrated services among partner agencies?

Executive Summary

The evaluation was
designed to determine
whether individuals who
participated in enhanced
reentry programming, as
measured by their
enrollment in SVORI
programs, had improved
post-release outcomes.

ƒ

To what extent did SVORI participants receive more
individualized and comprehensive services than
comparable non-SVORI offenders?

ƒ

To what extent did reentry participants demonstrate
better recidivism, employment, health, and personal
functioning outcomes than comparable non-SVORI
offenders?

ƒ

To what extent did the benefits derived from SVORI
programming exceed the costs?

The local nature of the SVORI programs and the expectation
that programs would tailor services to meet individual needs
meant that the intervention to be evaluated was not a program
in the typical conceptualization of the term (e.g., a residential
drug program or a cognitive behavior program). Instead,
SVORI was a funding stream that agencies used to expand and
enhance existing programs or to develop and implement new
programs. Further, individuals not in SVORI programs also
generally received some services. Thus, although the
components of the individual programs were identified and the
extent of service receipt was measured, the SVORI Multi-site
Evaluation was not designed to examine the impact of specific
services or combinations of services. The evaluation was
designed to determine whether individuals who participated in
enhanced reentry programming, as measured by their
enrollment in SVORI programs, had improved post-release
outcomes.
This report presents findings for the adult male participants in
the impact evaluation. Other results from the impact and
economic evaluations are presented in the following reports:
ƒ

Lattimore, P. K., & Visher, C. A. (2009). The Multi-site
Evaluation of SVORI: Summary and synthesis. Research
Triangle Park: RTI International.

ƒ

Lattimore, P. K., & Steffey, D. M. (2009). The Multi-Site
Evaluation of SVORI: Methodology and analytic
approach. Research Triangle Park, NC: RTI
International.

ƒ

Lindquist, C. H., Barrick, K., Lattimore, P. K., & Visher,
C. A. (2009). Prisoner reentry experiences of adult
females: Characteristics, service receipt, and outcomes
of participants in the SVORI Multi-site Evaluation.
Research Triangle Park, NC: RTI International.

ES-3

Prisoner reentry experiences of adult males

ƒ

Lattimore, P. K., Steffey, D. M., & Visher, C. A. (2009).
Prisoner reentry experiences of adult males:
Characteristics, service receipt, and outcomes of
participants in the SVORI Multi-site Evaluation. Research
Triangle Park, NC: RTI International.

ƒ

Hawkins, S., Dawes, D., Lattimore, P. K., & Visher, C. A.
(2009). Reentry experiences of confined juvenile
offenders: Characteristics, service receipt, and outcomes
of juvenile male participants in the SVORI Multi-site
Evaluation. Research Triangle Park, NC: RTI
International.

ƒ

Cowell, A., Roman, J., & Lattimore, P. K. (2009). An
economic evaluation of the Serious and Violent Offender
Reentry Initiative. Research Triangle Park, NC: RTI
International.

The following section summarizes the research design.
Subsequent sections present key findings and conclusions.

RESEARCH DESIGN
The SVORI Multi-site Evaluation included an implementation
assessment (to document the programming delivered across
the SVORI programs) and an impact evaluation (to determine
the effectiveness of programming). Sixteen programs were
included in the impact evaluation, comprising 12 adult
programs and 4 juvenile programs located in 14 states (adult
only unless specified): Colorado (juveniles only), Florida
(juveniles only), Indiana, Iowa, Kansas (adults and juveniles),
Maine, Maryland, Missouri, Nevada, Ohio, Oklahoma,
Pennsylvania, South Carolina (adults and juveniles), and
Washington. 1 The impact evaluation included pre-release
interviews (conducted approximately 30 days before release
from prison) and a series of follow-up interviews (conducted at
3, 9, and 15 months post-release). Nearly 2,400 prisoners
returning to society—some of whom received SVORI
programming and some of whom received “treatment as usual”
in their respective states—were included in the impact
evaluation. An economic analysis was also conducted in five of
the impact sites to assess the extent to which program benefits
exceeded costs; findings from this study are reported
separately (see Cowell et al., 2009).
1

ES-4

Site selection and other methodological aspects of the study are
described in The Multi-site Evaluation of SVORI: Methodology and
Analytic Approach (Lattimore & Steffey, 2009).

Executive Summary

A site-specific research design was developed for each impact
site (see Lattimore & Steffey, 2009). In two sites (Iowa and
Ohio), individuals were randomly assigned to SVORI programs.
In the remaining sites, comparison groups were developed by
isolating the criteria that local site staff used to identify
individuals eligible for enrollment in their SVORI program
(these included factors such as age, criminal history, risk level,
post-release supervision, transfer to pre-release facilities, and
county of release) and replicating the selection procedures on a
different population. Where possible, the comparison
participants came from the same pre-release facilities and were
returning to the same post-release geographic areas as the
SVORI participants.
Data collection consisted of four waves of in-person, computerassisted interviews: the pre-release interview (Wave 1)
conducted about 1 month before expected release and three
follow-up interviews (Waves 2 through 4) conducted 3, 9, and
15 months after release. In addition, oral swab drug tests were
conducted during the 3- and 15-month interviews for
respondents who were interviewed in a community setting. For
examination of recidivism outcomes, the interview and drug
test data were supplemented with arrest data obtained from
the FBI National Crime Information Center (NCIC) and with
administrative records obtained from state correctional
agencies. 2 All interviews were conducted in private settings by
experienced RTI field interviewers using computer-assisted
personal interviewing. Pre-release interviews were conducted
from July 2004 through November 2005 in more than 150
prisons and juvenile detention facilities and were designed to
obtain data on the respondents’ characteristics and pre-prison
experiences, as well as incarceration experiences and services
received since admission to prison. Post-release interviews
were conducted from January 2005 through May 2007. The
post-release interviews were similar in content across waves
and obtained data on reentry experiences, housing,
employment, family and community integration, substance
abuse, physical and mental health, supervision and criminal
history, service needs, and service receipt.

2

Note that in some instances these administrative records were
supplemented with data obtained from online criminal history
databases. Readers are referred to Lattimore and Steffey (2009) for
details.

ES-5

Prisoner reentry experiences of adult males

A total of 2,564 cases were fielded of adult men eligible for
inclusion in the multi-site evaluation. Wave 1 (pre-release)
interviews were obtained with 1,697 (66%) of these men.
Among eligible subjects approached for interviews, refusal rates
were reasonably low—11.5% across the 12 sites. Most of the
noninterviews (21% of fielded cases) were due to the men
being released before their Wave 1 pre-release interview could
be scheduled and completed. Nearly 80% of the men who were
interviewed at Wave 1 responded to at least one of the followup interviews. All cases were fielded for each follow-up wave.
Overall, the response rate for follow-up interviews increased
over time. Response rates for the Wave 2, 3, and 4 interviews
were 58%, 61%, and 66%, respectively. All three follow-up
interviews were obtained for 43% of the adult male samples.
Exhibit ES-1 shows the number of interviews conducted at each
wave, by group and site.
Exhibit ES-1. Completed interviews by wave, group, and site

State
IA
IN
KS
MD
ME
MO
NV
OH
OK
PA
SC
WA
Total

Wave 1
(Pre-release)
NonSVORI
SVORI
114
55
64
94
23
48
130
124
35
44
36
50
107
50
47
38
42
51
57
66
179
166
29
48
863
834

Wave 2
(3 Months Post)
NonSVORI
SVORI
59
29
49
53
11
15
58
63
20
21
26
31
77
31
25
26
26
12
43
50
123
104
12
20
529
455

Wave 3
(9 Months Post)
NonSVORI
SVORI
82
39
41
56
14
15
64
56
24
26
27
24
81
31
28
27
29
17
44
50
119
95
12
34
565
470

Wave 4
(15 Months Post)
NonSVORI
SVORI
87
46
45
59
15
24
65
65
25
30
26
35
82
29
28
26
24
27
46
48
126
109
13
33
582
531

Although the response rates were reasonable, the possibility
remains that respondents who “dropped out” of subsequent
waves of interviews differed from those who completed the
follow-up interviews. As preliminary evidence that the attrition
was random or affected the SVORI and non-SVORI groups
similarly, the SVORI and comparison groups were compared

ES-6

Executive Summary

and were found to be similar at each wave on a range of
characteristics. Results from models that examined for
differences between groups with respect to response also
suggested that SVORI program participation was not related to
whether a participant responded.
Propensity score techniques were used to improve the
comparability between the SVORI and non-SVORI groups.
Weighted analyses were used to examine the treatment effects
of SVORI program participation with respect to outcomes in
housing, employment, family/peer/community involvement,
substance use, physical and mental health, and criminal
behavior and recidivism.

KEY FINDINGS
This section summarizes key findings from the evaluation.
Characteristics of study participants are described next,
followed by descriptions of expressed service needs, reported
service receipt, and post-release outcomes.
Research Subject Characteristics

The study participants
were high-risk offenders
who had extensive
criminal and substance
use histories, low levels
of education and
employment skills, and
families and peers who
were substance and
criminal justice system
involved.

The study enrolled 863 SVORI program participants and 834
comparison men. The study participants were high-risk
offenders who had extensive criminal and substance use
histories, low levels of education and employment skills, and
families and peers who were substance and criminal justice
system involved. There were few statistically significant
differences in the characteristics of the groups.
More than half (57%) of the SVORI respondents were black and
32% were white. Only 4% of both groups identified themselves
as Hispanic. The average age of respondents in both samples
was about 29 years at the time of the pre-release interview.
Respondents in both groups had substantial educational
deficiencies—over one third (39% SVORI and 42% non-SVORI)
had not completed 12th grade or earned a GED. Most subjects
reported having worked at some time prior to incarceration—
89% of SVORI versus 92% of non-SVORI—and about two thirds
of both groups reported having a job during the 6 months prior
to incarceration (64% and 68%, SVORI and non-SVORI,
respectively).
More than 1 in 10 of both SVORI and non-SVORI respondents
reported that they were homeless, living in a shelter, or had no

ES-7

Prisoner reentry experiences of adult males

set place to live prior to their current incarceration. About 40%
of both groups reported that they were either currently married
or in a steady relationship (39% SVORI, 40% non-SVORI) at
the time of the pre-release interview. Most study participants
from both groups (59% SVORI and 61% non-SVORI) reported
having children under age 18. Large majorities of both groups
reported having family members and friends who had been
convicted of a crime or incarcerated, and who had problems
with drugs or alcohol.

Nearly all of the
respondents reported
having used alcohol and
drugs during their
lifetimes.

About two thirds of both
groups reported having
used one or more illicit
drugs during the 30 days
prior to their
imprisonment.

Nearly all of the respondents reported having used alcohol and
drugs during their lifetimes. The majority of both groups
reported using alcohol (96% SVORI and 97% non-SVORI), and
the average age of first use was about 14 years (13.7 and 13.6
for the SVORI and non-SVORI respondents, respectively).
Similarly, nearly all respondents in both groups reported having
used marijuana (92% SVORI and 94% non-SVORI), again
reporting a young age of first use (13.9 and 14.1 for the SVORI
and non-SVORI respondents, respectively). More than half of all
respondents reported having used cocaine (53% and 58% of
the SVORI and non-SVORI respondents, respectively), and
nearly one half reported having used hallucinogens (43% and
49%, SVORI and non-SVORI, respectively). Fewer respondents
reported using other substances.
There were few differences between the two groups with
respect to reported drug use during the 30 days prior to their
current incarceration. About two thirds of both groups reported
having used one or more illicit drugs during the 30 days prior to
their imprisonment (66% and 69% for the SVORI and nonSVORI respondents, respectively). Exhibit ES-2 shows the two
groups’ reported drug use during the 30 days prior to
incarceration for the most commonly reported drugs. More than
half of both SVORI and non-SVORI respondents reported using
marijuana; approximately one quarter of all respondents
reported using cocaine. More than half of SVORI and nonSVORI respondents had received treatment for a substance use
or mental health problem at some point during their lifetime
(56% and 55% of SVORI and non-SVORI, respectively).
SVORI and non-SVORI respondents reported considerable
involvement with the criminal justice system prior to their
current incarceration (Exhibit ES-3). On average, the
respondents were 16 years old at the time of their first arrest

ES-8

Executive Summary

Exhibit ES-2. Use of
specific substances
during the 30 days prior
to incarceration, by
group

52
53

Marijuana
22
26

Cocaine
Amphetamines

13
14

Pain relievers

11
14

SVORI
Non-SVORI

9
9

Hallucinogens
0%

20%

40%

60%

80%

100%

Note: Differences between SVORI and non-SVORI are not significant at the 0.05
level.

Exhibit ES-3. Criminal
history of respondents,
by group

Criminal History
Age at first arrest (mean)
Times arrested (mean)
Times convicted (mean)
Ever been previously incarcerated*
Times previously incarcerated (mean)*

SVORI
15.92
12.42
5.48
83%
1.20

Non-SVORI
16.03
13.14
5.70
87%
1.47

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

and had been arrested more than 12 times. 3 In addition to their
current term of incarceration, most respondents had served a
previous prison term, with the non-SVORI group being
significantly more likely to report a prior prison term (83% of
SVORI, 87% of non-SVORI). Also, the non-SVORI respondents
reported significantly more incarcerations, on average, than the
SVORI group (1.20 for SVORI, 1.47 for non-SVORI).
The two groups were similar in self-reported juvenile
detentions. Overall, about half reported that they had spent
time in a juvenile correctional facility for committing a crime. Of

3

This measure of prior arrest recoded extreme values to the 95th
percentile of reported arrests. The uncapped means were 14.48
(standard deviation 23.25) and 14.56 (standard deviation 17.49)
for SVORI and non-SVORI groups, respectively.

ES-9

Prisoner reentry experiences of adult males

those who reported a juvenile detention, they reported having
been detained, on average, 3.5 times.

SVORI respondents had
been incarcerated
significantly longer than
non-SVORI
respondents…and were
more likely to report
infractions and
segregation.

At the time of the pre-release interview, SVORI respondents
had been incarcerated significantly longer than non-SVORI
respondents (an average of 2.8 years and 2.3 years,
respectively). SVORI respondents also were more likely than
the non-SVORI respondents to report disciplinary infractions
and administrative segregations during their current
incarceration. Nearly two thirds (64%) of SVORI respondents
reported at least one disciplinary infraction, compared with
57% of non-SVORI respondents. Less than half reported
administrative segregation during the current term of
incarceration (45% of SVORI and 40% of non-SVORI). These
differences between groups are statistically significant but may
simply reflect the longer lengths of stay reported by the SVORI
respondents. 4
Service Needs
The findings substantiate previous research that male prisoners
returning to their communities after serving more than 2 years
in prison comprise a population with extremely high needs and
that their expressed needs remained high (if somewhat
diminished from pre-release) up to 15 months following release
from prison. Overall, there was little difference in reported
needs between the two groups. Key findings are summarized
below.

4

ES-10

ƒ

Expressed need for services post-release were lower
overall than those expressed 30 days prior to release.

ƒ

There was little difference in reported service needs at
3, 9, and 15 months following release.

ƒ

Levels of expressed need for educational/education/skills
and transition services and programming were high and
similar overall for men participating in SVORI programs
and the comparison subjects.

ƒ

More than 85% of respondents at all waves reported
needing more education, the highest need of six
employment/education/skills services

ƒ

Financial assistance, transportation and a driver’s license
were the most commonly reported of 10 transition

Longer lengths of stay expose subjects to greater opportunity to
commit infractions and receive administrative segregation; in other
words, the period at risk is longer.

Executive Summary

service needs pre-release and at 3, 9 and 15 months
post-release.
ƒ

SVORI and non-SVORI
subjects were similar on
the level of reported need.
Overall need dropped
following release.
Average need was similar
at 3, 9, and 15 months
post-release.

Exhibit ES-4. Weighted
average super bundle
scores by group for
Waves 1 through 4

Public health care insurance and financial assistance
were also consistently reported as needs by majorities of
both groups.

Service need bundle scores were developed to summarize
needs across specific domain areas (transition services; health
care services; employment, education and skills services;
domestic violence-related services; and child-related services)
and overall. There were a total of 28 needs across the bundles,
including 5 in the child-related services that were only relevant
for the men who reported having children. Scores for each
individual were generated by summing one/zero indicators for
whether the individual reported or did not report needing each
of the items; this sum was then divided by the number of items
in the bundle. At the individual respondent level, this score can
be interpreted as the proportion of the services in the bundle
that the individual reported needing. The “super bundle” need
scores are shown in Exhibit ES-4, which suggests three
findings: (1) SVORI and non-SVORI subjects were similar on
the level of reported need across all waves; (2) overall need
dropped following release; and (3) there was little difference in
average need across the three follow-up waves.

100

SVORI

90

Non-SVORI

80
70
60

54

54

50

42

43

43

43

44

45

40
30
20
10
0
Wave 1

Wave 2

Wave 3

Wave 4

Note: Differences between groups were not significant at the 0.05 level. Wave 1
= 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months
post-release; Wave 4 = 15 months post-release.

ES-11

Prisoner reentry experiences of adult males

Service Receipt
The report provides evidence that SVORI program participation
increased receipt of services and programming, including
programs to prepare for release, meeting with a case manager,
and receiving a needs assessment—although levels of provision
for most services fell far short of 100%, were substantially
below expressed needs for services, and declined substantially
following release. Key findings are summarized below:

SVORI program
participants received
substantially more services
pre-release.
Service receipt dropped
following release.
Service receipt continued
to decline as the time since
release increased.
By 15 months following
release, there was little
difference in receipt, with
both groups reporting
receipt of less than 10% of
service items.

ƒ

SVORI program participants reported receiving
significantly more coordination,
employment/education/skills, and transition services
than comparison subjects at all interviews.

ƒ

Overall levels of reported service receipt declined
substantially between the pre-release and the first postrelease interview and the differences between SVORI
and non-SVORI groups diminished over time.

ƒ

The percentages of SVORI participants who reported
receiving any employment-related services was 37%
pre-release—a small proportion that dropped to 34% at
3 months, 21% at 9 months, and 14% at 15 months
post-release; although far less than 100%, these
proportions were significantly higher at all waves than
reported by the non-SVORI group—20% pre-release,
declining to 10% at 15 months.

ƒ

Aggregate levels of service receipt were substantially
lower than comparable measures of service need (across
all bundles and time periods and among both groups),
indicating that most men had unmet needs.

ƒ

The services that men were most likely to report
receiving after release were similar across the postrelease waves and included post-release supervision,
case management, and needs assessments.

In addition to the domain areas identified for expressed needs,
a coordination services bundle was identified for service receipt.
Exhibit ES-5 summarizes information on average reported
service receipt for each group across the four waves of
interview data. For Wave 1, the results suggest that, on
average, about 30 days before release the SVORI subjects
reported receiving about a third (34%) of the various service
items and the non-SVORI comparison subjects reported
receiving about one quarter (24%) of the items. Exhibit ES-5
suggests four findings: (1) SVORI program participants

ES-12

Executive Summary

Exhibit ES-5. Weighted
average service receipt
super bundle scores by
group for Waves 1
through 4

100

SVORI

90

Non-SVORI

80
70
60
50
34

40
30

22

20

18

12

10

13

10

9

8

0
Wave 1

Wave 2

Wave 3

Wave 4

Note: Differences between groups at Waves 1, 2, and 3 were significant at the
0.0001 level. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release;
Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.

received substantially more services pre-release than nonSVORI subjects; (2) service receipt dropped substantially at
release; (3) reported service receipt continued to drop as the
time since release lengthened; and (4) there was little
difference in average receipt at 15 months following release.
Post-Release Outcomes
The significant—albeit less-than-universal—increase in service
receipt associated with participation in SVORI programs was
associated with moderately better outcomes with respect to
housing, employment, substance use, and self-reported
criminal behavior—although these improvements were not
associated with reductions in official measures of
reincarceration. As many of the previous evaluations of reentry
programs have focused primarily on recidivism and substance
use, this evaluation provided an opportunity to examine the
impacts of reentry programming on an array of other important
indicators of successful reintegration, including housing and
employment. Key findings are summarized below.
Housing
ƒ

SVORI programming did not appear to affect core
housing outcomes, including housing independence,
stability, and the extent to which housing challenges
were encountered. Although SVORI participants were
significantly more likely to have achieved housing
independence at 15 months, they were also less likely

ES-13

Prisoner reentry experiences of adult males

(albeit not significantly) to report that housing had been
stable and that they had not encountered housing
challenges during the 6 months prior to the 15-month
interview.
ƒ

For both the SVORI and non-SVORI groups, housing
situations improved gradually over the 3-, 9-, and 15month post-release time periods; for example, 72% of
SVORI participants and 70% of the non-SVORI group
were “housing independent” at the 3-month interview
compared to 86% of SVORI and 80% of non-SVORI 15
months following release.

ƒ

SVORI program participants were more likely to report
living in their own house or apartment at each follow-up,
although differences were not statistically significant at
the 0.05 level.

Employment
ƒ

SVORI programming appeared to positively affect many
dimensions of employment, with effects being strongest
for the time period prior to the 15-month post-release
interview. Compared with the non-SVORI group, SVORI
program participants were more likely to report
–

supporting themselves with a job (at 15 months
post-release),

–

working at a job that provided health insurance and
paid-leave benefits (at 3 and 15 months postrelease),

–

working at a job that provided formal pay (at
3 months)

ƒ

Members of both groups reported working about the
same number of months during the reference period—
about two thirds of the 3 or 6 months prior to interview.

ƒ

SVORI program participants were much more likely
post-release to report having supported themselves with
a job during the reference period in comparison to the
period immediately preceding their incarceration—
increasing from 59% prior to incarceration to 64% at
3 months post-release to 71% at 15 months; there was
little difference over time for those in the comparison
group.

Family, Peers, and Community Involvement

ES-14

ƒ

SVORI programming did not have an impact on familial
or peer relationships.

ƒ

For both groups, the levels of family emotional support,
family instrumental support, and the quality of intimate-

Executive Summary

partner relationships declined over the three postrelease time periods.
ƒ

The SVORI and non-SVORI groups reported similar
levels of negative peer exposure.

ƒ

Participation in civic activities, such as performing
volunteer work or participating in local organization, was
low for both groups; however, reports of civic action
increased over the three follow-up periods such that by
15 months post-release SVORI program participants
were significantly more likely to report civic activity.

Substance Use
ƒ

SVORI appeared to have a positive impact on abstinence
from drug use. Self-reported substance use was
generally lower for the SVORI group than the nonSVORI group, in several cases significantly so.

ƒ

Members of the SVORI group were also more likely to
report not using drugs and to not test positive on oral
swab drug tests that were conducted at 3 and 15
months post-release although these differences were not
statistically significant at the 0.05 level.

ƒ

Overall, substance use was quite high (primarily
marijuana) and increased across the post-release followup periods.

Physical and Mental Health
ƒ

SVORI did not appear to influence physical outcomes.
No differences were observed, at any time period,
between the SVORI and non-SVORI subsamples on the
12-Item Short Form Health Survey (SF-12; Ware et al.,
2002) physical health scale or the number of physical
conditions experienced.

ƒ

SVORI group members appeared to have slightly better
mental functioning as indicated by the SF-12 mental
health scale (at 3 and 15 months post-release) and
mental health status as measured by the Global Severity
Index (GSI-45; at 15 months post-release).

Criminal Behavior and Recidivism
ƒ

The men enrolled in SVORI were less likely to report
having perpetrated violence and to have engaged in any
criminal behavior during the 3 or 6 months prior to the
follow-up interviews, although differences were
statistically significant only for no criminal behavior at
3 months post-release.

ES-15

Prisoner reentry experiences of adult males

ƒ

There was little difference between the two groups in a
measure that combined reports of carrying a weapon
and committing a violent crime.

ƒ

SVORI program participants were more likely than nonSVORI respondents to report complying with conditions
of supervision at 15 months post-release.

ƒ

The proportion of members of both groups who either
(1) reported having been booked into jail since the last
interview or (2) were incarcerated at the time of the
follow-up interview was high and similar. Only about
80% of both groups were either not incarcerated during
the 3 month interview or reported having not been
booked into jail. This proportion declined to about 50%
at the 15 month interview.

ƒ

Members of the SVORI group were less likely to have an
officially recorded arrest during the 24 months following
release, but the differences were not statistically
significant.

ƒ

Reincarceration rates for both groups were high with
about 40% of both groups reincarcerated (official
measure) within 24 months of release.

Conclusions
Although SVORI programs were successful in increasing the
types and amounts of needs-related services provided prior to
and after release from prison, the proportion of individuals who
reported receiving services was less than reported need and,
generally, less than the expectations of the SVORI program
directors. This finding is consistent with SVORI programs that
were developing and implementing their programs and provides
a reminder that starting complex programs may require
sustained effort over several years to reach full
implementation.
Service delivery declined following release. Thus, overall, the
programs were unable to sustain support to individuals during
the critical, high-risk period immediately following release. This
decline may be due to the programs’ difficulty identifying and
coordinating services for individuals released across wide
geographic areas and, again, suggests the need for sustained
effort to reach full implementation.
SVORI program participation resulted in modest improvements
in intermediate outcomes at levels consistent with findings from
meta-analyses of single-program efforts (e.g., 10% to 20%). If

ES-16

Executive Summary

the underlying model that links services to improved
intermediate outcomes that in turn improve recidivism is
correct, the level of improvement in these intermediate
outcomes may have been insufficient to result in observable
reductions in recidivism.
Additional analyses are planned to determine whether there are
specific programs or subgroups associated with positive
outcomes and to examine the relationship between receipt of
specific services and outcomes.

ES-17

Introduction
SVORI was a collaborative federal effort to improve outcomes
for adults and juveniles returning to their communities after a
period of incarceration. The initiative sought to help states
better utilize their correctional resources to address outcomes
along criminal justice, employment, education, health, and
housing dimensions by providing grant funds in 2003 to state
agencies to establish or enhance prisoner reentry
programming. 5 Funded by the U.S. Departments of Justice
(DOJ), Labor (DOL), Education (DOEd), Housing and Urban
Development (HUD), and Health and Human Services (HHS),
SVORI was an unprecedented national response to the
challenges of prisoner reentry. Sixty-nine state and local
grantees (corrections and juvenile justice agencies) received
SVORI funding, representing all 50 states, the District of
Columbia, and the U.S. Virgin Islands. These grantees
developed 89 programs that targeted adult and juvenile
correctional populations.
The initiative responded to emerging research findings that
suggested that providing individuals with comprehensive,
coordinated services based on needs and risk assessments
could result in improved post-release outcomes. Grantees were
to use their SVORI funding to create a three-phase continuum
of services for returning prisoners that began during the period
of incarceration, intensified just before release and during the
early months post-release, and continued for several years
after release as former inmates took on more productive and
5

Although grant awards were announced by the federal partners in
2002, grantees were required to complete planning and other
activities prior to having access to full funding. Program directors
reported in a survey conducted in 2005 that access to full grant
funds was provided to grantees over a substantial time frame—
February 2003 to May 2004.

1

Prisoner reentry experiences of adult males

independent roles in the community. The SVORI programs
attempted to address the initiative’s goals and provide a wide
range of well-coordinated services to returning prisoners.
Although SVORI programs shared the common goals of
improving outcomes across various dimensions and improving
service coordination and systems collaboration, programs
differed substantially in their approaches and implementations
(Lindquist, 2005; Winterfield & Brumbaugh, 2005; Winterfield
et al., 2006; Winterfield & Lindquist, 2005).
The SVORI evaluation
was designed to answer
the following research
questions:
● To what extent did
SVORI lead to more
coordinated and
integrated services
among partner
agencies?
● To what extent did
SVORI participants
receive more
individualized and
comprehensive
services than
comparable, nonSVORI offenders?
● To what extent did
reentry participants
demonstrate better
recidivism,
employment, health,
and personal
functioning outcomes
than comparable,
non-SVORI
offenders?
● To what extent did
the benefits derived
from SVORI
programming exceed
the costs?

In spring 2003 the National Institute of Justice awarded RTI
International, a nonprofit research organization, a grant to
evaluate programs funded by SVORI. The Urban Institute, a
nonpartisan economic and social policy research organization,
collaborated on this project, which was one of the largest
evaluation studies ever funded by the National Institute of
Justice. With data collected from grantee staff, partnering
agencies, and returning prisoners, the 6-year evaluation
involved an implementation evaluation of all 89 SVORI
programs, an intensive impact evaluation of 16 selected
programs, and an economic analysis on a subset of the impact
sites (Lattimore et al., 2005). The goal of the SVORI evaluation
was to document the implementation of SVORI programs and
determine whether they accomplished SVORI’s overall goal of
increasing public safety by reducing recidivism among the
populations served.
The SVORI evaluation was designed to answer the following
research questions:
ƒ

To what extent did SVORI lead to more coordinated and
integrated services among partner agencies?

ƒ

To what extent did SVORI participants receive more
individualized and comprehensive services than
comparable, non-SVORI offenders?

ƒ

To what extent did reentry participants demonstrate
better recidivism, employment, health, and personal
functioning outcomes than comparable, non-SVORI
offenders?

ƒ

To what extent did the benefits derived from SVORI
programming exceed the costs?

This report presents findings for the adult male participants in
the impact evaluation, which included 863 SVORI participants
and 834 comparison subjects returning from prison in 12

2

Introduction

states. 6 Other results from the impact and economic
evaluations are presented in the following reports:
ƒ

Lattimore, P. K., & Visher, C. A. (2009). The Multi-site
Evaluation of SVORI: Summary and synthesis. Research
Triangle Park: RTI International.

ƒ

Lattimore, P. K., & Steffey, D. M. (2009). The Multi-Site
Evaluation of SVORI: Methodology and analytic
approach. Research Triangle Park, NC: RTI
International.

ƒ

Lindquist, C. H., Barrick, K., Lattimore, P. K., & Visher,
C. A. (2009). Prisoner reentry experiences of adult
females: Characteristics, service receipt, and outcomes
of participants in the SVORI Multi-site Evaluation.
Research Triangle Park, NC: RTI International.

ƒ

Hawkins, S., Dawes, D., Lattimore, P. K., & Visher, C. A.
(2009). Reentry experiences of confined juvenile
offenders: Characteristics, service receipt, and outcomes
of juvenile male participants in the SVORI Multi-site
Evaluation. Research Triangle Park, NC: RTI
International.

ƒ

Cowell, A., Roman, J., & Lattimore, P. K. (2009). An
economic evaluation of the Serious and Violent Offender
Reentry Initiative. Research Triangle Park, NC: RTI
International.

In the remainder of this chapter, previous research on prisoner
reentry is summarized, the SVORI and the evaluation design
are described, and the SVORI programs provided in the 12
impact evaluation sites are characterized. Subsequent sections
provide detailed information on characteristics of the evaluation
participants, self-reported service needs and receipt at each of
the four interviews, post-release outcomes, and conclusions
and policy recommendations.

6

Findings from the analysis of the pre-release interview data collected,
on average, 30 days prior to release from prison were published in
Pre-release Characteristics and Service Receipt among Adult Male
Participants in the SVORI Multi-site Evaluation (Lattimore et al.,
2008). These findings were primarily descriptive and conveyed
characteristics of the respondents, as well as their pre-prison and
incarceration experiences. Differences between the SVORI and nonSVORI groups were also identified and discussed with respect to
assessing the comparability of the groups and implications for
outcome analyses.

3

Prisoner reentry experiences of adult males

RESEARCH ON PRISONER REENTRY
In 2008, more than 735,000 prisoners were released from state
and federal prisons across the country (West, Sabol, & Cooper,
2009). This number represents a greater than four-fold
increase over the nearly 170,000 released in 1980 (Harrison,
2000). With the exception of those who die while in prison, all
prisoners will eventually “re-enter” the community. Prisoner
reentry has sweeping consequences for the individual prisoners
themselves, their families, and the communities to which they
return (Petersilia, 2003; Travis, 2005). Nationwide, over half of
individuals who are released from prison are reincarcerated
within three years. Programs and services for men and women
leaving prison are designed to stop this revolving door and
encourage individuals to desist from offending. Imprisonment
without such preparation for community reintegration may
reduce human capital and impede the acquisition of pro-social
skills and behaviors, thus lessening the probability of a
successful transition from prison to the community (Visher &
Travis, 2003; Western, 2007). However, in comparison to
twenty years ago, men and women leaving prison are less
prepared for reintegration, less connected to community-based
social structures, and more likely to have health or substance
abuse problems than prior cohorts (Lynch & Sabol, 2001;
Petersilia, 2005).
In recent years, significant attention has been focused on the
impact of these increases in rates of incarceration and rates of
return from jail or prison (Bonczar & Beck, 1997; Clear, Rose, &
Ryder, 2001; Hagan & Coleman, 2001; Mauer, 2000; Travis,
2005). The geographic clustering of former prisoners by socioeconomic characteristics has led to disproportionate rates of
removal from, and return to, already distressed communities
(Clear et al., 2001; Lynch & Sabol, 2001). As a result, current
research on the social and economic impacts of incarceration is
increasingly focused on local effects of incarceration and
prisoner reintegration, and the concurrent effects on family
structure, intergenerational offending, and general community
well-being (Clear et al., 2001; Hagan & Dinovitzer, 1999).
Prisoner reentry programs that have emerged since the late
1990s seek to address the effects of incarceration by more
successfully reintegrating former prisoners, thereby reducing
subsequent offending.

4

Introduction

Reentry programming is designed to break the cycle between
offending and incarceration. Incarcerating offenders generally
has two purposes: incapacitation and deterrence. Incapacitation
leads to temporary instrumental desistance, and specific
deterrence may lead to future deterrence. However, desistance
is mainly achieved through rehabilitative programming.
Predictors of desistance generally do not vary by the pattern of
past criminal behavior or by the antecedent characteristics of
the offender (Laub & Sampson, 2001). Processes that
consistently are identified as leading to desistance include
marriage and stable families (Laub & Sampson, 1993; Sampson
& Laub, 1993), aging (Glueck & Glueck, 1974; Gottfredson &
Hirschi, 1990; Laub & Sampson, 2003), stable employment
(Laub & Sampson, 1993; Sampson & Laub, 1993) and reduced
exposure to antisocial peers (Laub & Sampson, 2003; Warr,
1998). In addition, all of these outcomes may be dependent
upon cognitive changes in identity which are the precursor to
changes in behavior (Maruna, 2001).
Until recently, the majority of rehabilitation and reentry
strategies have been dominated by service providers who
represent a single domain from among the possible correlates
of desistance. For instance, many reentry programs are
centered on one-stop workforce centers whose main function is
to prepare and place individuals in jobs. Reentry services may
include interventions directly related to skill acquisition to
improve labor market prospects such as job readiness, training,
and placement programs. Other reentry programs may focus on
reducing specific deficits by reducing substance abuse,
addressing physical and mental health disorders, improving
educational attainment through GED or high school
programming, or offering other assistance from the small
(access to official identification and transfer of prescriptions) to
the large (securing transitional and long-term housing).
Reentry initiatives also may assist in the cognitive development
of participants to promote behavioral change through faithbased or classroom-based programming (e.g., anger
management, parenting skills, and life skills).

5

Prisoner reentry experiences of adult males

The complexity of the
disadvantages
confronting prisoners
after release means that
individual offenders often
require more than a
single program or
intervention.

…many reentry
specialists are
encouraging a broader
focus on comprehensive
reentry strategies, not
specific programs…

However, the needs of individuals returning to the community
usually span these domains of problems, and typical service
providers are unlikely to be as effective at providing or
facilitating other services as they are in their primary area of
expertise. For example, it is not unusual for individuals
struggling with mental health and substance abuse disorders to
be denied entry into programs designed to respond to either
but not both of these disorders. The complexity of the
disadvantages confronting prisoners after release means that
individual offenders often require more than a single program
or intervention. To address this dilemma, many reentry
specialists are encouraging a broader focus on comprehensive
reentry strategies, not specific programs (Lattimore, 2007;
National Research Council, 2007; Re-entry Policy Council,
2005; Visher, 2007). Such strategies would involve multiple
levels of government, coordination of efforts across agencies,
and involvement of organizations that are traditionally not part
of the reentry discussion (e.g., public health, local businesses;
community colleges). Moreover, these coordinated efforts may
improve reintegration across a broader range of outcomes
(e.g., employment, substance use, health) than simply
reductions in recidivism.
This emerging focus on the need for comprehensive
programming provided the context within which the federal
government developed the Serious and Violent Offender
Reentry Initiative, resulting in the award of SVORI grants in
2002 and SVORI program start-ups in 2003 and 2004. The brief
review of literature discussed in the remainder of this section
provides a context in which the findings of the Multi-Site
Evaluation of the Serious and Violent Offender Reentry
Initiative can be assessed.
Characteristics and Needs of Former Prisoners
Individuals in the prison population face disadvantages
associated with poor educational and employment histories,
alcohol and other drug misuse, and poor physical and mental
health (e.g., Petersilia, 2005). These disadvantages can result
in serious challenges at the time of release as the individual
attempts to reenter and integrate with the community. Among
the most serious challenges facing former prisoners are finding
employment and addressing health needs, including risk of
renewed substance abuse after release. Most prisoners have

6

Introduction

extensive substance use histories and need substance abuse
treatment. About half of individuals incarcerated in state
prisons report that they used illegal drugs in the month before
the arrest that led to the current prison term (Karberg & James,
2005), and many used drugs on a daily or weekly basis
(Petersilia, 2005; Visher & Mallik-Kane, 2007). Despite high
levels of drug use, half or fewer receive drug treatment while
incarcerated (Petersilia, 2005; Visher & Mallik-Kane, 2007;
Winterfield & Castro, 2005). Prisoners are also more likely to
suffer from physical and mental health problems than the
general population (see Greifinger, Bick, & Goldenson, 2007;
Petersilia, 2005). The Bureau of Justice Statistics reported that
as many as 16 percent of those in state prisons, local jails, or
on probation are mentally ill (James & Glaze, 2006). Finally,
finding employment is the single largest concern reported by
men and women before they are released from prison (Visher,
2007). Although, about two thirds of prisoners nationwide had
worked before incarceration (Petersilia, 2005) , former
prisoners often have difficulty finding and maintaining
employment. Results of a study of men and women leaving
prison in three states revealed that less than one in five had a
job lined up in the month before release (Visher, 2007). Not
surprisingly, nine in ten prisoners also reported that they
needed job training and more education.
Former prisoners are also not a homogeneous group. The need
for services immediately before and soon after release is likely
to vary from individual to individual. For example, although one
in five have served sentences of five years or more, the
average prisoner in the U.S. serves about 28 months, and
about 17 percent are released within one year (Lynch & Sabol,
2001). The average age of men leaving state prison is older
than many realize—33 years—and those in their thirties will
likely have different needs than those in their twenties
(Petersilia, 2005). Some have serious mental or physical health
needs; others are fathers with minor children.
Thus, men and women face formidable challenges in their
personal lives as they leave prison and return to the
community. Many individuals require intensive services and
support to meet post-release needs ranging from transportation
to employment services to health care. Further, these
individuals are often faced with additional obstacles to a
successful transition from prison in the form of collateral effects

7

Prisoner reentry experiences of adult males

(sometimes called “invisible punishments”) that may include
exclusions from certain professions (e.g., realtor or health care)
and access to public benefits (e.g., student loans, public
housing, or food stamps), as well as the loss of parental rights
(Hagan & Dinovitzer, 1999; Petersilia, 2003; Travis, 2002).
These realities, in conjunction with conditions imposed upon
their release (e.g., employment, in-person reporting, and
payment of restitution, fees, and fines), make the first few
weeks and months after release especially difficult.
Risk and Needs Assessment
Research over the past 25 years has led to the development of
a new generation of risk assessment tools that are reasonably
able to predict the probability that an individual will commit
additional crimes (Holsinger, Lurigio, & Latessa, 2001).
Moreover, there is considerable evidence that concentrating
both services and supervision on the higher-risk individuals will
result in greater reductions in crime than a more generalized
approach to service delivery and supervision (Byrnes, Macallair,
& Shorter, 2008; MacKenzie, 2006).

Programs that adhere
more strictly to
assessment results to
generate appropriate
treatment matching may
ultimately achieve better
outcomes, although this
hypothesis remains
generally untested

8

An important first step in linking newly incarcerated men and
women to rehabilitative programming is a thorough assessment
of needs. Intake screening is widely utilized in correctional
institutions for security classification, and prisoners are also
typically screened for literacy and physical and mental health
conditions. However, problems exist with screening instruments
and procedures including choice of instrument, lack of staff
training, and lack of follow-through regarding subsequent
assessment and treatment while in prison (Moore & Mears,
2003). Moreover, if programs are limited (which is a typical
situation in many prisons), staff can feel that screening and
assessment serves no worthwhile purpose. Programs that
adhere more strictly to assessment results to generate
appropriate treatment matching may ultimately achieve better
outcomes, although this hypothesis remains generally untested.
There is also general agreement that promising correctional
programs should be timed so that they are provided close to a
prisoner’s release date, focus on skills applicable to the local
job market, reflect current market demands, and are provided
for an extended period of time, generally three to six months
(Cullen & Gendreau, 2000). However, the duration of in-prison
programming is somewhat arbitrary—one research area that

Introduction

remains unexplored is whether reentry programs would be
more effective if the length of reentry programs varied
according to the magnitude of the individual prisoner’s deficits.
Employment, Education, and Skills Building
Research indicates that ‘work’ is a primary feature of successful
reintegration (Laub, Nagin, & Sampson, 1998; Sampson &
Laub, 1990, 1993) as connections made at work may serve as
informal social controls that help prevent criminal behavior. For
former prisoners, employment is correlated with lower
recidivism (Rossman & Roman, 2003; Visher, Debus, & Yahner,
2008) and rates of return to prison are significantly reduced by
participation in work readiness programs (Buck, 2000; Finn,
1998; Sung, 2001). Although recent studies have indicated that
work programs can have a significant impact on the
employment and recidivism rates of men (see Bushway &
Reuter, 2002), vocational and educational programs often are
unavailable in prisons, and their availability has declined (Lynch
& Sabol, 2001).
Stable employment is one of the key predictors of desistance
that can be directly addressed through policy or programming.
As a result, many reentry initiatives have typically focused on
preparing returning prisoners to re-enter the job market.
Reentry services often include interventions directly related to
skill acquisition to improve labor market prospects such as job
readiness, training and placement programs. Although about
two thirds of prisoners worked prior to incarceration (Beck et
al., 1993), their educational level, work experience, and skills
are well below national averages for the general population
(Andrews & Bonta, 2006), and the stigma associated with
incarceration often makes it difficult for them to secure jobs
following release (Bushway & Reuter, 2002; Holzer, Raphael, &
Stoll, 2006). When former prisoners do find jobs, they tend to
earn less than individuals with similar background
characteristics who have not been incarcerated (Bushway &
Reuter, 2002). Thus, research supports a strong programmatic
emphasis on increasing individual employability through skills
training, job readiness, and work release programs, both during
incarceration and after release.
Few such programs have been studied using a random
assignment research design. One exception is the evaluation of
the Opportunity to Succeed (OPTS) program, which delivered

9

Prisoner reentry experiences of adult males

employment services within a set of comprehensive services for
drug-using former prisoners, and found that participants were
more likely to be employed full-time in the year after release.
However, self-reported arrests and official record measures of
recidivism showed no differences between participants and
controls (Rossman & Roman, 2003). Employed participants,
however, reported fewer arrests and less drug use. Another
study of prisoners in Tennessee, who were required to secure
either employment or enroll in a training program as a
condition of release, found that those who graduated had
marginally better outcomes than a matched comparison, while
those who failed had significantly worse outcomes (Chalfin,
Tereshchenko, Roman, Roman, & Arriola, 2007).
In a meta-analysis examining the impact of employment
training and job assistance in the community for persons with a
criminal record, Aos, Miller, and Drake (Aos, Miller, & Drake,
2006) concluded that these programs have a modest, but
significant, 5 percent impact on recidivism. However, in another
meta-analysis, using a very similar set of studies and methods,
Visher, Winterfield, and Coggeshall (Visher, Winterfield, &
Coggeshall, 2005) concluded that community-based
employment programs do not significantly reduce recidivism for
persons with previous involvement with the criminal justice
system.
Contemporary job assistance and training programs for former
prisoners such as the Center for Employment Opportunities
(New York), Safer Foundation (Chicago), and Project Rio
(Texas) are more holistic in their approach and incorporate
other transition services and reentry support into their
programs (Buck, 2000) while maintaining a primary focus on
job placement. Although several rigorous evaluations are
underway, the impact of these newer types of comprehensive,
employment-focused programs on former prisoners’
employment and recidivism rates is not yet known.
In the U.S., adult corrections has a long history of providing
education and employment training (Gaes, Flanagan, Motiuk, &
Stewart, 1999; Piehl, 1998). Comprehensive reviews of dozens
of individual program evaluations generally conclude that adult
academic and vocational programs lead to modest reductions in
recidivism and increases in employment (Aos, 2006; Cullen &
Gendreau, 2000; Gaes et al., 1999; Gerber & Fritsch, 1994;

10

Introduction

Wilson, Gallagher, & MacKenzie, 2000). However, the majority
of the evaluations have one or more methodological problems
(Wilson et al., 2000).
Numerous studies show that recidivism rates are significantly
reduced for those with more education; however, just getting a
GED—the most prevalent in-prison educational programming—
generally does not show impacts on reentry outcomes (Adams
et al., 1994; Boudin, 1993; Harer, 1995; Stillman, 1999).
Despite the high demand for these programs by inmates,
participation in these programs declined from 42 percent in
1991 to 35 percent in 1997 (Lynch & Sabol, 2001). Reasons for
these declines include the rapid growth in prison populations in
combination with decreased funding for correctional
programming, the frequent transfer of prisoners from one
facility to another, and greater interest in short-term programs
such as substance abuse and cognitive-behavioral programs
(Lawrence, Mears, Dubin, & Travis, 2002).
Research suggests that correctional education programming is
most successful as part of a systematic approach, integrating
employability, social skills training, and other specialized
programming (Holzer & Martinson, 2005). Best-practice
correctional education programs carefully tailor programming to
the needs of individuals and to related vocational and job skills
training. Education and job training for low earners are most
successful when they provide workers with credentials that
meet private sector demands. Thus, comprehensive programs
that provide training, a range of services and supports, job
retention incentives, and access to employers are promising,
but rigorous evaluations are lacking.
Integrated Post-release Service Delivery

Individuals exiting prison
are at the highest risk of
rearrest during the first
few months after release.

Individuals exiting prison are at the highest risk of rearrest
during the first few months after release. The overall probability
of arrest is roughly twice as high in the first month of
supervision as it is in the fifteenth month (National Research
Council, 2007). Just as important, the arrest rate between
months 15 and 36 is uniformly lower than in the earlier period.
Between months 1 and 15, the chance of arrest for property
and drug offenses drops by 40 percent. Thus, the challenge for
policy makers and corrections professions is to respond to this
‘high-risk’ period by frontloading services and supervision
during these early months (Pew Center on the States, 2009).

11

Prisoner reentry experiences of adult males

Developing tailored reentry plans to guide post-release services
begins with an assessment and classification of re-offending
risk, needs, and strengths leading to individualized and unified
case planning and management that addresses individual
variation in likelihood to re-offend (Andrews & Bonta, 2006;
Wiebush, McNulty, & Le, 2000). These assessments differ from
the initial screens conducted at prison entry and are the
foundation for an individualized reentry plan. Such targeted
assessments not only determine the specific kinds of services
and supervision level that individuals need to succeed, but also
determine who requires documents, medications, or other
immediate transition preparation. The ideal case management
approach incorporates a family and social network perspective,
a mix of surveillance and services focused on risk and
protective factors, the imposition of realistic and enforceable
release conditions coupled with graduated incentives and
sanctions. In addition, this approach uses service brokering to
community service providers and resources and other
supportive community organizations (Healy, Foley, & Walsh,
1999; Petersilia, 2003).
In the immediate period after release from prison, access to
supportive reentry (or transitional) services are critical in those
first few days, weeks, and months for men and women as they
readjust to life in the community. Among the most important
are obtaining photo identification, getting appropriate clothes,
securing stable housing, having reliable transportation, and
signing up for public assistance, if eligible. Unfortunately, often
these types of needs are not addressed before release and it
falls to family and friends to help arrange these for the former
prisoner. There are no evaluations of programs that focus
exclusively on the immediate post-release period, but one
program in Maryland that generally focused on this period was
found to reduce recidivism—particularly the most serious and
violent offenses—and to be cost-effective (Roman, Brooks,
Lagerson, Chalfin, & Tereshchenko, 2007).
Research suggests that an integrated, multi-modal intervention
strategy—often referred to as ‘wraparound’ service delivery—
can be effective in meeting the multiple needs of this
population. This strategy involves “wrapping” a comprehensive
array of individualized services and support networks “around”
clients, rather than forcing them to enroll in pre-determined,
inflexible treatment programs (Franz, 2003). Treatment

12

Introduction

services are usually provided by multiple agencies working
together as part of a collaborative interagency agreement, and
each individual’s treatment plan is determined and managed by
an interdisciplinary team consisting of a caseworker, family and
community members, and several social services and mental
health professionals.
General Evaluations of Reentry Programs
Over many decades of attention, the overwhelming majority of
evaluations of rehabilitative programs for offenders have
focused on programs designed to address specific individual
needs, such as reducing drug and alcohol use, addressing
mental health issues, or finding a job. But as discussed earlier,
the focus of recent reentry services and programs has been to
ease the transition of individuals exiting prison, addressing
needs with individualized approaches. Many, if not most reentry
programs for individuals exiting prison are relatively new,
having begun as a result of federal funding in the early years of
this decade. As a result, there are few impact evaluations of
programs focused specifically on reentry (Petersilia, 2004).
Numerous challenges characterize the extant research
assessing the effectiveness of programs for formerly
incarcerated individuals, whether focused on reentry or general
rehabilitation. Foremost among the challenges is the lack of
theoretical models that articulate behavior change among
former prisoners. Within any particular substantive area, there
are also problems of fidelity in that a particular service
approach may manifest itself in different ways under different
programs and circumstances. As a result, it is often difficult to
generalize research findings from one program to others, and
substantial variability exists among the outcome variables
examined (e.g., employment, homelessness, substance use,
recidivism). The numerous combinations of program types
unique to each study also render comparisons difficult. Finally,
there are problems related to the research itself, as rigorous
experimental designs—including the use of comparison groups
(randomly assigned or otherwise)—are rare in this research
literature (National Research Council, 2007).
Several reviews of reentry program evaluations recently have
examined the available research on what works with regard to
reentry and/or rehabilitative programming (Aos et al., 2006;
Gaes et al., 1999; Lipsey & Cullen, 2007; MacKenzie, 2006;

13

Prisoner reentry experiences of adult males

Petersilia, 2004; Seiter & Kadela, 2003). The evidence has been
very consistent in establishing that contact-driven supervision,
surveillance, and enforcement of supervision conditions have a
limited ability to change offender behavior or to reduce the
likelihood of recidivism (MacKenzie, 2006). However, intensive
supervision programs with a clear treatment component do
show a sizeable impact on recidivism (Aos et al., 2006; Gaes et
al., 1999; Petersilia, 2004).

All of the strategies
MacKenzie identified as
effective focus on
dynamic criminogenic
factors, are skill-oriented,
are based on
cognitive/behavioral
models, and treat multiple
offender deficits
simultaneously.

MacKenzie (MacKenzie, 2006) recently summarized the “what
works” literature in corrections, with specific chapters on
various community corrections programs (e.g., life skills,
cognitive behavioral therapy, education, drug treatment, and
intensive supervision). She concluded that human serviceoriented programs were much more effective than those based
on a control or deterrent philosophy. In particular, there is
growing consensus that practices focusing on individual-level
change, including cognitive change, education, and drug
treatment, are likely to be more effective than other strategies,
such as programs that increase opportunities for work, reunite
families, and provide housing (see also Andrews & Bonta,
2006). All of the strategies MacKenzie identified as effective
focus on dynamic criminogenic factors, are skill-oriented, are
based on cognitive/behavioral models, and treat multiple
offender deficits simultaneously. These conclusions are
consistent with several large meta-analyses of the evaluation
literature (Andrews et al., 1990; Aos et al., 2006; Lipsey &
Cullen, 2007).
Selection of program type may be less important than proper
implementation of the program. Delivering a program in the
wrong context (i.e., intensive substance abuse treatment to
casual drug users) or poor implementation are common and
may explain most of the weak or null findings in the research
studies. Despite advances in knowledge and best practices,
studies of programs for offenders have documented persistent
problems in implementation and adherence to the fidelity of
evidence-based practice models (Lowenkamp, Latessa, &
Smith, 2006; Petersilia, 2004; Young, 2004). Additionally,
improperly implemented programs may be harmful. One recent
reentry program, Project Greenlight, was developed from
research and best practice models to create an evidence-based
reentry initiative which was evaluated with a random
assignment research design (Wilson & Davis, 2006). However,

14

Introduction

the program participants performed significantly worse than a
comparison group on multiple measures of recidivism after one
year. The evaluators concluded that the New York program did
not replicate past best practice. Instead, Project Greenlight
modified past practice to fit institutional requirements, was
delivered ineffectively, did not match individual needs to
services, and failed to implement any post-release continuation
of services and support (Wilson & Davis, 2006; see also Rhine,
Mawhorr, & Parks, 2006; Visher, 2006; Marlowe, 2006). The
evaluators attributed the findings to a combination of
implementation difficulties, program design, and a mismatch
between participant needs and program content. A key
difficulty for Project Greenlight, as with many other communitybased reentry programs, was its lack of integration into an
overall “continuum of care” strategy that linked prison and
community-based treatment.

… effective rehabilitation
strategies have strong
program integrity,
identify criminogenic
factors, employ a multimodal treatment
approach, use an
actuarial risk
classification, and ensure
responsivity between an
offender’s learning style
and mode of program
delivery.

Another line of research has focused on identifying the
principles of effective treatment (as opposed to the substantive
content of the program) in assessing evidence-based practices
(e.g., Andrews & Bonta, 2006; Cullen & Gendreau, 2000).
MacKenzie (2006) summarizes this work, identifying five
principles of effective rehabilitation strategies. Specifically, she
notes that effective rehabilitation strategies have strong
program integrity, identify criminogenic factors, employ a
multimodal treatment approach, use an actuarial risk
classification, and ensure responsivity between an offender’s
learning style and mode of program delivery. One of the failings
of Project Greenlight was poor management of the program
according to these principles that help guide or maximize
program effectiveness (Andrews, 2006).
In her review of what works in reentry programming, Petersilia
(2004) discusses the striking disconnect between the published
‘what works’ literature and the efforts of governmental reentry
task forces to develop programs that are thought to improve
offender transitions from prison to the community. The goal of
most reentry programs is to develop a seamless transition from
prison to the community. However, the challenges in this
regard are enormous. Corrections departments and community
supervision agencies often have conflicting incentives, and
community-justice partnerships linking these organizations with
community groups face even larger hurdles. An important
barrier to effective reentry strategies in many communities is

15

Prisoner reentry experiences of adult males

the lack of information sharing between the criminal justice
system and the community because of institutional barriers and
privacy rules. Effective service delivery after release requires
coordinated actions by government agencies, non-government
service providers, and the community to ensure that returning
prisoners do not fall through service gaps between agencies.
Yet, knowledge about how to develop and manage these
partnerships is lacking (Rossman & Roman, 2003).

… the SVORI Multi-site
Evaluation provided an
opportunity to test
whether coordinated
services provided in
response to assessment to
meet individual needs
could be implemented
and whether these
services would have
positive impacts on
criminal justice,
employment, health,
housing, and substance
abuse outcomes.

SVORI programs were developed and implemented by the
grantee agencies as these strands of research findings were
emerging. The programs were to provide a range of
coordinated services (based on needs and risk assessments)
that spanned incarceration and return to the community,
including services that focused on cognitive development.
Although the programs differed from site to site, as discussed
below, the overall focus of the SVORI initiative was consistent
with emerging recommendations at the time the programs were
developed and implemented. Thus, the SVORI Multi-site
Evaluation provided an opportunity to test whether coordinated
services provided in response to assessment to meet individual
needs could be implemented and whether these services would
have positive impacts on criminal justice, employment, health,
housing, and substance abuse outcomes.

SVORI
Exhibit 1 shows the SVORI program logic model and the SVORI
Multi-site Evaluation design, including the evaluation
components. The SVORI program model identifies SVORI
funding, technical assistance (TA) and requirements as INPUTS
that, in combination with local resources in the sites
(THROUGHPUTS), yield a set of services and programming
(OUTPUTS) that are expected to improve the OUTCOMES for
SVORI participants, as well as to improve the state and local
systems that provide these services and programs. Community
and individual participant characteristics influence these
throughputs, outputs, and outcomes.
The criteria for programs funded by SVORI grants were as
follows:
ƒ

16

Programs were to improve criminal justice, employment,
education, health (including substance abuse and mental
health), and housing outcomes.

Introduction

Exhibit 1. SVORI program logic model and evaluation framework

ƒ

Programs were to include collaborative partnerships
between correctional agencies, supervision agencies,
other state and local agencies, and community and
faith-based organizations.

ƒ

Program participants were to be serious and/or violent
offenders.

ƒ

Program participants were to be 35 years of age or
younger.

ƒ

Programs were to encompass three stages of reentry—in
prison, post-release on supervision, and postsupervision.

ƒ

Needs and risk assessments were to be used to guide
the provision of services and programs to participants.

In some cases, grantees asked for and received permission for
exceptions to these criteria. For example, some programs were

17

Prisoner reentry experiences of adult males

primarily post-release programs, and age restrictions were
sometimes lifted (e.g., for programs targeting sex offenders).

SVORI programs were
locally designed, varying
in terms of what was
provided, when, and to
whom.

The evaluation was
designed to determine
whether individuals who
participated in enhanced
reentry programming, as
measured by their
enrollment in SVORI
programs, had improved
post-release outcomes.

Operating within the broad guidelines of these criteria, each
program was locally designed along a variety of dimensions,
including the types of services offered, the focus on pre-release
and post-release components, and the type(s) of individuals to
be served. In other words, programs varied in terms of what
was being provided, when, and to whom. Grantees also
identified the locations where the program would be provided
both pre- and post-release. Thus, a SVORI program could be
narrowly focused on a single institution pre-release serving
participants who were returning to a single community postrelease or could be implemented throughout the correctional
(or juvenile justice) system serving participants who were to be
released statewide. A combination of multiple (but not all)
institutions and multiple (but not all) communities was the
modal configuration. Finally, because services were to be
delivered to individuals based on their specific needs and risk
factors, individuals participating in a SVORI program could
receive different types and amounts of services depending upon
individual needs. 7 Thus, one challenge for the evaluation was to
attempt to characterize SVORI and the SVORI programs.
The local nature of the SVORI programs and the expectation
that programs would tailor services to meet individual needs
meant that the intervention to be evaluated was not a program
in the typical conceptualization of the term (e.g., a residential
drug program or a cognitive behavior program). Instead,
SVORI was a funding stream that agencies used to expand and
enhance existing programs or to develop and implement new
programs. Further, individuals not in SVORI programs also
generally received some services. Thus, although the
components of the individual programs were identified and the
extent of service receipt was measured, the SVORI Multi-site
Evaluation was not designed to examine the impact of specific
services or combinations of services. The evaluation was
designed to determine whether individuals who participated in

7

18

Specific details on the planned characteristics of individual programs
are available in the National Portrait of SVORI (Lattimore et al.,
2004). Winterfield and Lindquist (2005) and Lindquist (2005)
provide information on the delivery of services and programs by the
SVORI programs, along with information on barriers to
implementation.

Introduction

enhanced reentry programming, as measured by their
enrollment in SVORI programs, had improved post-release
outcomes.
Exhibit 1 shows the evaluation components. The
implementation assessment addressed the extent to which the
89 SVORI programs (69 grantees) increased access to services
and promoted systems change. The impact evaluation assessed
the effectiveness of SVORI by comparing key outcomes among
those who received services as part of SVORI programs with
those of a comparable group of individuals who received
“treatment as usual” in 16 sites. The third component of the
evaluation, an economic analysis, determined the return on
SVORI investment and included both a cost-benefit and a costeffectiveness analysis.
The remainder of this chapter briefly describes the SVORI
programs provided in the 12 adult impact sites, and the
evaluation design. Subsequent chapters provide detailed
information on the characteristics of study participants, selfreported service needs and receipt, post-release outcomes, and
conclusions and policy recommendations.

ADULT IMPACT SITE SVORI PROGRAMS
Information on the SVORI programs was obtained by reviewing
program proposals and implementation plans and through
annual surveys of the local SVORI programs. This section
provides information on the characteristics of the adult
programs in the sites selected for the impact evaluation.
Two thirds of the SVORI program directors (PDs) for the
12 adult impact programs reported that their programs focused
equally on the pre- and post-release phases, while one PD
reported that the pre-release phase and three PDs reported
that the post-release phase was the primary focus.8 Six of the
12 PDs indicated that the SVORI grant funds were used to start
new programs; three indicated the funds were used to fill

8

Source: 2005 SVORI program director survey; see Lattimore and
Steffey (2009) for comparisons between impact and non–impact
sites and for additional information on program characteristics. With
the exception of specific impact site selection criteria including
expected enrollment and likelihood of implementation, the impact
and non–impact sites were similar along most measured
dimensions.

19

Prisoner reentry experiences of adult males

service gaps, and two said they were used to expand existing
services.
Exhibit 2 shows the primary outcome foci for the 12 adult
impact sites, as reported by the PDs. When asked to identify
the top 3 (of 8) outcomes for their program, six or more of the
PDs indicated community integration, employment, improved
decision-making or self-sufficiency, and substance use. Further,
6 or more of the 12 PDs thought that it would be fair to use all
of the eight outcomes as measures of the effectiveness of their
programs. Indeed, all 12 PDs thought that it would be fair to
rate the effectiveness of their programs with respect to
community integration and employment outcomes; and 11
reported that it would be fair to rate effectiveness based on the
programs’ ability to reduce recidivism.
Exhibit 2. Outcome foci
among adult impact and
non–impact sites

%
Outcomes
Outcomes targeted (program director ranked in top 3)
Community integration
66.70%
Employment
58.30%
Improved decision-making or self-sufficiency
58.30%
Reduced substance use
50.00%
Housing
33.30%
Educational attainment
16.70%
Improved physical and/or mental health
8.30%
Family reunification/functioning
8.30%
Outcomes fair to determine program effectiveness
Community integration/connectedness
100.00%
Employment
100.00%
Reduced recidivism
91.70%
Reduced substance use
75.00%
Family reunification/functioning
75.00%
Educational attainment
66.70%
Housing
66.70%
Improved physical and/or mental health
50.00%

N
8
7
7
6
4
2
1
1
12
12
11
9
9
8
8
6

Source: 2006 program director survey of 12 impact site PDs.

Exhibit 3 provides information on the types of services that the
SVORI PDs reported were being provided to SVORI program
participants in 2005. The services are organized into “service
bundles,” which are described in more detail in subsequent
chapters. Of coordination services, the 12 PDs suggested that,
on average, most participants were receiving risk and needs
assessments, as well as treatment/release plan development,
both pre- and post-release. Further, most of the PDs reported

20

Introduction

Exhibit 3. Mean proportion of SVORI program participants receiving pre-release and postrelease services in adult program impact sites (as reported by program directors)

Services
Bundle 1: Coordination Services
Risk assessment
Needs assessment
Treatment/release plan development
Formal post-release supervision
Bundle 2: Transition Services
Legal assistance
Assistance obtaining identification (e.g., driver’s license, Social
Security card)
Assistance obtaining benefits and completing applications (e.g.,
Medicaid, disability)
Financial support/emergency assistance
Peer support groups
One-on-one mentoring
Housing placements or referrals
Transportation
Bundle 3: Health Services
Comprehensive drug treatment programs
Alcoholics Anonymous/Narcotics Anonymous
Counseling sessions
Mental health services
Anger management/violence counseling
Medical services
Dental services
Bundle 4: Employment, Education, and Skills Development Services
Education/GED/tutoring/literacy
Vocational training
Employment referrals/job placement
Resume and interviewing skills development
Work-release program
Cognitive skills development/behavioral programming
Life skills training
Bundle 5: Family Services
Domestic violence services
Parenting skills development
Family reunification
Family counseling

Pre-release
Mean (SD)

Post-release
Mean (SD)

0.88 (0.30)
0.88 (0.30)
0.91 (0.29)
NA

0.86 (0.29)
0.81 (0.35)
0.91 (0.29)
0.82 (0.27)

0.24 (0.25)
0.60 (0.42)

0.21 (0.24)
0.46 (0.35)

0.44 (0.35)

0.47 (0.35)

0.45 (0.43)
0.42 (0.38)
0.50 (0.42)
0.54 (0.39)
NA

0.63 (0.29)
0.37 (0.34)
0.45 (0.38)
0.66 (0.29)
0.60 (0.34)

0.31 (0.20)
0.39 (0.33)
0.60 (0.37)
0.35 (0.31)
0.52 (0.34)
0.70 (0.41)
0.70 (0.41)

0.32 (0.28)
0.44 (0.30)
0.48 (0.31)
0.35 (0.30)
0.35 (0.25)
0.30 (0.32)
0.24 (0.23)

0.67 (0.29)
0.33 (0.29)
0.51 (0.35)
0.68 (0.33)
0.17 (0.30)
0.49 (0.35)
0.71 (0.39)

0.37 (0.29)
0.34 (0.30)
0.65 (0.33)
0.56 (0.37)
0.24 (0.39)
0.46 (0.35)
0.62 (0.37)

0.38 (0.35)
0.40 (0.37)
0.39 (0.34)
0.18 (0.26)

0.32 (0.32)
0.48 (0.34)
0.50 (0.34)
0.24 (0.25)

Note: NA = Not applicable, GED = general educational development, SD = standard deviation. Values were
calculated by taking the midpoint of the response categories (0%, 1–25%, 26–50%, 51–75%, 76–99%, and 100%)
reported by the SVORI adult program directors for each of the services.
Source: 2005 program director survey.

21

Prisoner reentry experiences of adult males

that most SVORI program participants were expected to be on
post-release supervision.
Other services were reported to be less universally provided.
Housing placements/referrals was the only one of seven
transition services that the PDs reported was being provided to
more than half of SVORI program participants both pre- and
post-release. The PDs were also asked about the provision of
post-release transportation assistance and indicated, on
average, that 60% of SVORI participants were receiving help
with transportation post-release.
Although the PDs reported that more than half of program
participants were receiving counseling, anger
management/violence counseling, or medical and dental
services pre-release, this finding did not apply to post-release,
as the PDs reported providing less than half of all participants
with the seven different types of health services.
PDs were asked to report the percentage of SVORI program
participants who were provided each of seven different types of
employment, education, and skills development services. The
PDs reported that about two thirds were receiving education
programming pre-release, but only about one third (0.37) postrelease. The PDs did report that, on average, more than half of
all SVORI program participants were provided employment
referrals/job placement, resume and interviewing skills
development, and life skills training both pre- and post-release.
Of the four types of family services, only family reunification
post-release was provided to half or more of the SVORI
program participants. In a subsequent chapter of this report,
reports on service receipt by SVORI program participants are
compared with those of the comparison subjects who did not
participate in SVORI programs.

MULTI-SITE EVALUATION DESIGN
This section briefly summarizes the evaluation design, data
collection procedures, and issues related to potential bias.
Interested readers are referred to Lattimore and Steffey (2009)
for a full description of the evaluation methodology.
The evaluation components are shown in Exhibit 1. The
implementation assessment addressed the extent to which the
89 SVORI programs (69 grantees) increased access to services

22

Introduction

and promoted systems change. Results were presented in
Lattimore et al. (2005) and Winterfield et al. (2006) and in a
series of Reentry Research in Action brief reports, including
Lindquist and Winterfield (2005) and Lindquist (2005).
The impact evaluation assessed the effectiveness of SVORI by
comparing key outcomes among those who received services as
part of SVORI programs with those of a comparable group of
individuals who received “treatment as usual” in 16 sites. These
impact sites were chosen from the 89 SVORI programs
following an extensive site selection process, with the objective
of achieving diversity in programmatic approach and
geographic representation. Likelihood of implementation and
anticipated program enrollment (case flow) were other criteria
considered during the site selection process. The 16 programs
included 12 adult programs and 4 juvenile programs located in
14 states (adult only unless specified): Colorado (juveniles
only), Florida (juveniles only), Indiana, Iowa, Kansas (adults
and juveniles), Maine, Maryland, Missouri, Nevada, Ohio,
Oklahoma, Pennsylvania, South Carolina (adults and juveniles),
and Washington. The impact evaluation consisted of a
longitudinal study of 2,391 returning prisoners (adult males,
adult females, and juvenile males) who were interviewed
approximately 1 month before release and then again at 3, 9,
and 15 months post-release. 9
As detailed in Lattimore and Steffey (2009), a site-specific
research design was developed for each impact site. In two
sites (Iowa and Ohio), the programs implemented a randomassignment evaluation design. 10 In the remaining sites,
comparison groups were developed by isolating the criteria that
local site staff used to identify individuals eligible for enrollment
in their SVORI program (these included factors such as age,
criminal history, risk level, post-release supervision, transfer to
pre-release facilities, and county of release) and replicating the
selection procedures on a different population. Where possible,
the comparison participants came from the same pre-release
9

Juvenile females were excluded from the impact evaluation because
of the extremely small number of SVORI participants in this
subgroup.
10
Even though random assignment was employed in Iowa,
participants were not evenly allocated to the two conditions.
Program slots were filled first, and then the remaining participants
were assigned to the control condition (which is the reason SVORI
participants exceeded non-SVORI comparisons in Iowa).

23

Prisoner reentry experiences of adult males

The evaluation adopted
an “intent to treat”
approach…

facilities and were returning to the same post-release
geographic areas as the SVORI participants. In some instances,
comparison participants were identified as those who met all
eligibility criteria except pre- or post-release geographic
parameters. When this exception occurred, the comparison
sample consisted of selected individuals from pre-release
facilities that were comparable to facilities in which SVORI was
available, or they consisted of selected individuals from SVORI
facilities who were returning to a separate but similar
geographic area. Eligible respondents (both SVORI and
comparison) were identified monthly during the 17-month
enrollment period for the impact evaluation. The evaluation
adopted an “intent to treat” approach—SVORI participants were
individuals who were enrolled in SVORI programs at the time of
the pre-release interview regardless of the extent of services
received or the length of time in programs.
The third component of the evaluation, an economic analysis,
determined the return on SVORI investment and included both
a cost-benefit and a cost-effectiveness analysis; results are
presented in Cowell, Roman and Lattimore (2009).
Exhibit 4 presents the distribution of adult male pre-release
interview respondents by site and by group. As the exhibit
indicates, there was variation in the numbers of individuals who
enrolled in the study across the sites, with the total number of
enrollees ranging from 71 (Kansas) to 345 (South Carolina).
Although the within-site case distribution was roughly 50:50
(SVORI to non-SVORI) in most sites, there were exceptions—
most notably Iowa, Indiana, and Kansas.
Data Collection Procedures
Data collection consisted of four waves of in-person, computerassisted interviews: the pre-release interview (Wave 1)
conducted about one month before expected release and three
follow-up interviews (Waves 2 through 4) conducted 3, 9, and
15 months after release. In addition, oral swab drug tests were
conducted during the 3- and 15-month interviews for
respondents who were interviewed in a community setting. For
examination of recidivism outcomes, the interview and drug
test data were supplemented with arrest data obtained from
the FBI National Crime Information Center (NCIC) and with

24

Introduction

Exhibit 4. Adult male
sample sizes, by state
and group

State
Iowa
Indiana
Kansas
Maine
Maryland
Missouri
Nevada
Ohio
Oklahoma
Pennsylvania
South Carolina
Washington
Total

SVORI
114
64
23
35
130
36
107
47
42
57
179
29
863

Non-SVORI
55
94
48
44
124
50
50
38
51
66
166
48
834

Total
169
158
71
79
254
86
157
85
93
123
345
77
1697

% of Total
10.0
9.3
4.2
4.7
15.0
5.1
9.2
5.0
5.5
7.2
20.3
4.5
100.0

administrative records obtained from state correctional
agencies. 11

Pre-release interviews
were conducted
approximately 30 days
before release…

The post-release
interviews were
conducted approximately
3, 9, and 15 months
following release …

All interviews were conducted in private settings by
experienced RTI field interviewers using computer-assisted
personal interviewing. Pre-release interviews were conducted
from July 2004 through November 2005 in more than 150
prisons and juvenile detention facilities. Pre-release interviews
were conducted approximately 30 days before release and
obtained data on the respondents’ characteristics and preprison experiences, as well as incarceration experiences and
services received since admission to prison. These interviews
also obtained data on the respondents’ post-release plans and
expectations about reentry.
Post-release interviews were conducted from January 2005
through May 2007. The post-release interviews were conducted
approximately 3, 9, and 15 months following release and were
similar in content across waves and obtained data on reentry
experiences, housing, employment, family and community
integration, substance abuse, physical and mental health,
supervision and criminal history, service needs, and service
receipt. The interview instruments were developed through an

11

Note that in some instances these administrative records were
supplemented with data obtained from online criminal history
databases. Readers are referred to Lattimore and Steffey (2009) for
details.

25

Prisoner reentry experiences of adult males

extensive instrumentation process involving substantive domain
experts and the use of existing, validated measures and scales.
In addition to obtaining approval from and oversight by the
Institutional Review Boards at RTI and the Urban Institute,
memoranda of agreement or formal research agreements were
negotiated with all agencies, and evaluation staff ensured that
study procedures were approved by all facilities in which
interviews were conducted (or by correctional agencies
overseeing the facilities).
The following sections summarize the examination of
nonresponse, attrition, and selection. The concluding section in
this chapter provides information on the propensity score model
and approach to the analyses.
Approach for Addressing Nonresponse and Attrition
A total of 2,564 cases were fielded of adult men eligible for
inclusion in the multi-site evaluation. Wave 1 (pre-release)
interviews were obtained with 1,697 (66%) of these men.
Among eligible subjects approached for interviews, refusal rates
were reasonably low—11.5% across the 12 sites. A breakdown
of the categories of noninterviews and ineligible cases is
provided in Appendix A, Exhibit A-1. As shown in the exhibit,
most of the noninterviews with eligible men (21%) were due to
the men being released before their Wave 1 pre-release
interview could be scheduled and completed. 12
Nearly 80% of the men who were interviewed at Wave 1
responded to at least one of the follow-up interviews. All cases
were fielded for each follow-up wave. Overall, the response rate
for follow-up interviews increased over time. Response rates for
the Wave 2, 3, and 4 interviews were 58%, 61%, and 66%,
respectively. All three follow-up interviews were obtained for
43% of the adult male samples. Exhibit 5 shows the number of
interviews conducted at each wave, by group and site.
12

26

Release before an interview could be completed was particularly
problematic during the initial fielding of the Wave 1 interviews. The
original protocols required identifying eligible participants 30 to 60
days prior to release. Although these cases were processed quickly
(i.e., within 2 weeks to attempt an interview), some subjects were
released before the interview could be scheduled. The protocols
were changed to identify subjects earlier. No information was
uncovered that would suggest that these “early releases” affected
individuals in the two study groups differently. Further, although
the release was “early,” the difference between actual and expected
release dates was only a few days.

Introduction

Exhibit 5. Completed interviews by wave, group, and site

State
IA
IN
KS
MD
ME
MO
NV
OH
OK
PA
SC
WA
Total

Wave 1
(Pre-release)
NonSVORI
SVORI
114
55
64
94
23
48
130
124
35
44
36
50
107
50
47
38
42
51
57
66
179
166
29
48
863
834

Wave 2
(3 Months Post)
NonSVORI
SVORI
59
29
49
53
11
15
58
63
20
21
26
31
77
31
25
26
26
12
43
50
123
104
12
20
529
455

Wave 3
(9 Months Post)
NonSVORI
SVORI
82
39
41
56
14
15
64
56
24
26
27
24
81
31
28
27
29
17
44
50
119
95
12
34
565
470

Wave 4
(15 Months Post)
NonSVORI
SVORI
87
46
45
59
15
24
65
65
25
30
26
35
82
29
28
26
24
27
46
48
126
109
13
33
582
531

Although the response rates were reasonable, the possibility
remains that respondents who “dropped out” of subsequent
waves of interviews differed from those who completed the
follow-up interviews. As preliminary evidence that the attrition
was random or affected the SVORI and non-SVORI groups
similarly, the SVORI and comparison groups were compared
and were found to be similar at each wave on a range of
characteristics. Results from models that examined for
differences between groups with respect to response also
suggested that SVORI program participation was not related to
whether a participant responded.
Approach for Addressing Selection Bias
In most sites, men were not randomly assigned to SVORI or
non-SVORI conditions. Although considerable effort was
expended during the planning phases of the evaluation to work
with each site to identify appropriate comparison populations,
the possibility of differences between the SVORI and nonSVORI groups existed. A multitude of variables were examined,
comparing SVORI to non-SVORI evaluation participants, and
few differences between the two adult male groups were
observed.

27

Prisoner reentry experiences of adult males

Exhibit 6 shows the t-statistics for comparisons between SVORI
and non-SVORI participants. As can be seen, for the adult
males, there are statistically significant differences for several
variables, some of which have traditionally been linked to
criminal behavior. In particular, those in SVORI programs were
younger on average than non-SVORI participants at the time of
the instant incarceration (26.1 years versus 27.1 years), more
likely to be black (57% versus 50% black), less likely to be
white (32% versus 37% white), and less likely to have been
employed either in the 6 months prior to the current
incarceration (64% versus 68%) or ever (89% versus 92%)—
although the latter differences are relatively small. Although
there were no significant differences in self-reported drug use
immediately prior to the current incarceration, those in the
non-SVORI group were more likely to report ever using cocaine
(58% versus 53%) and heroin (23% versus 18%). SVORI
participants were more likely than non-SVORI participants to be
serving time for a drug crime (36% versus 31%), while nonSVORI participants were more likely to be serving time for a
public order crime (22% versus 17%). This last finding is
consistent with non-SVORI participants being more likely than
SVORI participants to report that they were currently
incarcerated for a parole violation (which was coded as a public
order crime; 35% versus 27%) and for non-SVORI participants
to report more prior prison incarcerations on average (1.33
versus 1.12 incarcerations). 13
Propensity score matching techniques were used to improve the
comparability between the SVORI and non-SVORI groups. The
propensity score model is discussed in the following section.

13

28

Subjects were asked to indicate all crimes for which they were
currently incarcerated, so an individual could have reported serving
time, for example, for both a violent crime and a parole violation.

Introduction

Exhibit 6. t-statistics comparing means of SVORI and non-SVORI groups

Variable
Age at incarceration
Race_white
Race_black
Race_other
Homeless/shelter/no set place to live prior to incarceration
Intimate relationship 6 months prior to incarceration
Employed during 6 months prior to incarceration
Ever held a job?
Received substance use treatment prior to incarceration
Received treatment for MH problem prior to incarceration
Any victimization 6 months prior to incarceration
Any violence perpetration 6 months prior to incarceration
Used alcohol 30 days prior to incarceration
Used marijuana 30 days prior to incarceration
Used drugs other than marijuana 30 days prior to incarceration
Ever use marijuana
Ever use cocaine
Ever use heroin
Conviction offense: person/violent crime
Conviction offense: property crime
Conviction offense: drug crime
Conviction offense: public order/other crime
Currently serving time for parole violation
Age at first arrest (minimum set at 7 years)
Arrest rate
Conviction rate
Times in juvenile lockup
Incarceration rate
Number of previous prison incarcerations

Adult Males
N = 1,691
−2.57*
−2.30*
2.74*
−0.83
−0.12
−0.28
−2.04*
−2.21*
0.42
−0.52
0.61
0.80
0.43
−0.76
−1.92
−1.25
−2.09*
−2.59*
0.92
−1.35
2.36*
−2.58*
−3.18*
−0.47
−0.50
−0.25
0.86
−3.28*
−2.97*

Note: MH = mental health.
*p value < 0.05, two-tailed test.

Propensity Score Model
A propensity score model was used to address potential
selection bias due to the quasi-experimental design (see Rubin
[2006] for a collection of seminal papers in propensity score
modeling; see D’Agostino [1998] for an accessible tutorial).
Propensity score models use observed characteristics to model
the likelihood that an individual with those characteristics will

29

Prisoner reentry experiences of adult males

be selected (or assigned) to the intervention. 14 The purpose is
to identify a set of parameters that are then used to estimate
the probability of assignment to the intervention for each
ˆ ) are then
individual in a study. These probabilities (p-hats or p
used either (1) to establish probability strata (or bins) within
which subjects have similar probabilities of receiving the
intervention; (2) as weights in the outcome models; or (3) as
matching variables. The success of the estimation is judged by
the effectiveness of the strata or weights in reducing
differences between the treatment and control groups on
observed characteristics or, in the common terminology,
achieving balance between the two groups.
The stratification or binning approach was used during
preliminary assessments of the impact of SVORI on multiple
outcomes. The final outcome models were estimated using the
weighting approach, since it greatly simplifies the presentation
of findings. Results showed that population average treatment
effects estimated by combining results from the analyses based
on strata were nearly identical to those derived from the
weighted models—as would be expected.
The adult male sample included 1,697 observations, 1,500
(88.4%) of which had no missing values on any of the variables
14

30

Propensity scoring methods are not without limitations. For
example, use of propensity scores can only adjust for included
covariates (Glynn, Schneeweiss, & Sturmer, 2006; Rosenbaum &
Rubin, 1983). Unlike randomization, which tends to balance
treatment and control groups on observed and unobserved
covariates, use of propensity scores only balances on observed
confounding covariates. The failure to include unobserved
covariates can lead to biased estimates of treatment effects.
However, if many of the covariates believed to be related to
treatment assignment are measured, propensity score approaches
(i.e., matching, stratification, regression adjustment) should yield
consistent and approximately unbiased estimates of treatment
effects (D'Agostino, 1998; Rosenbaum & Rubin, 1983). A second
limitation is that propensity score approaches work better in larger
samples; in studies with small samples, substantial imbalances of
covariates may be unavoidable (Rubin, 1997). However, this is also
true of randomized experiments and is not limited to propensity
score methods. A third possible limitation is that included covariates
that are strongly related to treatment assignment and only weakly
correlated with the outcome are treated the same as covariates that
are strongly related to both treatment assignment and outcome
(Rubin, 1997). This might be considered a limitation because
including irrelevant covariates can reduce efficiency. Rubin (1997)
notes, however, that the potential biasing effects of failing to
control for weakly correlated covariates are worse than the
potential loss of efficiency from including them.

Introduction

included in the propensity score model. However, to take
advantage of the full sample (and minimize selection issues
with respect to item missingness), imputation procedures were
used to generate values for missing items. In particular, the
logit model to generate the probability of assignment to SVORI
[p(SVORI) or p(S)] was estimated within the framework of
SAS® 9.1 PROC MI ANALYZE. This SAS procedure allowed
accommodation of item missingness by imputing values for
missing values.
As detailed in Lattimore and Steffey (2009), the propensity
score model included a total of 24 variables that reflected preassignment characteristics. These variables were the following:
ƒ

age at incarceration;

ƒ

race (white and other; black was reference);

ƒ

homeless in the 6 months prior to incarceration;

ƒ

employed in the 6 months prior to incarceration;

ƒ

steady relationship in the 6 months prior to
incarceration;

ƒ

experienced victimization in the 6 months prior to
incarceration;

ƒ

family deviance scale;

ƒ

peer deviance scale;

ƒ

used drugs other than marijuana in the 30 days prior to
incarceration;

ƒ

used marijuana in the 30 days prior to incarceration;

ƒ

drank alcohol in the 30 days prior to incarceration;

ƒ

mental health treatment prior to incarceration;

ƒ

substance abuse treatment prior to incarceration;

ƒ

age at first arrest;

ƒ

times locked up in juvenile detention;

ƒ

arrest rate;

ƒ

conviction rate;

ƒ

incarceration rate;

ƒ

currently serving time for a parole violation;

ƒ

currently serving time for a person/violent crime;

ƒ

currently serving time for a property crime;

ƒ

currently serving time for a drug crime; and

ƒ

currently serving time for a public order or other crime.

31

Prisoner reentry experiences of adult males

The coefficients from the logit model were then used to
ˆ for each individual (i.e., the 1,697 adult male
generate a p
15
subjects).

Assessment of the validity of the model is attained by
examining the extent to which the propensity scores allow
attainment of balance between the SVORI and non-SVORI
groups. After this issue is addressed, information is presented
on the distribution of the propensity scores that the models
generated.
The purpose of the propensity score is to achieve greater
comparability between treatment and comparison groups. Two
ways of checking for balance are to examine t statistics
comparing SVORI and non-SVORI means or standardized
differences (see, e.g., Rosenbaum & Rubin, 1985). The data
were examined using both approaches; results suggested
substantial improvement with respect to balance within bins.
These results are presented in Lattimore and Steffey (2009).
ˆ values were used as weights in
Additionally, the estimated p
models that provided checks to determine the extent to which
these weights resulted in comparability between the two
groups. Exhibit 7 provides one indication in the improvement in
balance attained with the propensity scores. This exhibit
presents the results of regressing individually the variables that
were included in the SVORI propensity score model on the
SVORI indicator using Proc Survey Logistic Regression. These
analyses tested whether knowledge of a particular
characteristic (e.g., race) resulted in better prediction of SVORI
program participation. As can be seen in Exhibit 7, the results
suggest that this hypothesis was rejected in every case,

15

Within the PROC MI procedure, each of the five data sets was used
to generate the five sets of coefficients shown above. Each set of
coefficients was then applied to the dataset from which it was
generated, resulting in five p
ˆ s for each individual. The five
estimates were then averaged to produce one estimate for each
individual. Because the range of parameter estimates is small for all
of the variables included in the model, the difference between the
minimum and maximum p
ˆ values is also small for most
observations. The difference between the minimum and maximum
ˆ values ranged from 0.0008 to 0.1271 with a mean of 0.012
p
(standard deviation = 0.012) and median of 0.008. Extreme values
in the maximum difference between estimated p-hats were rare.
The 99th, 95th, 90th, and 75th percentiles were 0.063, 0.032,
0.023, and 0.012, respectively

32

Introduction

Exhibit 7. Balance checks for Wave 1 data based on propensity score weighted regression of
the variable on a SVORI indicator

Variable

Estimate

SE

Wald
Chi Sq

Age at incarceration
Race_white
Race_black
Race_other
No home
Employed
Steady relationship
Drug treatment prior
MH treatment prior
Victim_prior
Perpetration
rBSUA1b
rBSUA10b
Other_drug
Person crime
Property crime
Drug crime
Public order/other crime
Parole violation
Age first arrest
arrest_rate
convict_rate
# juvenile detentions
inc_rate

−0.00006
−0.00280
0.00206
0.00107
0.00096
−0.00077
0.00225
0.00696
−0.00293
−0.00636
0.00000
0.00645
−0.00112
−0.00211
0.00217
0.00226
−0.00050
−0.00151
0.00150
0.00048
−0.00760
0.01522
−0.00024
−0.07839

0.007
0.104
0.099
0.149
0.148
0.104
0.106
0.100
0.114
0.100
0.106
0.106
0.099
0.099
0.101
0.114
0.105
0.123
0.114
0.010
0.108
0.206
0.016
0.874

0.000
0.001
0.000
0.000
0.000
0.000
0.000
0.005
0.001
0.004
0.000
0.004
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.005
0.005
0.000
0.008

Prob
Chi Sq

Odds
Ratio
Estimate

Lower
CL

Upper
CL

0.993
0.979
0.983
0.994
0.995
0.994
0.983
0.945
0.980
0.949
1.000
0.951
0.991
0.983
0.983
0.984
0.996
0.990
0.990
0.962
0.944
0.941
0.988
0.929

1.000
0.997
1.002
1.001
1.001
0.999
1.002
1.007
0.997
0.994
1.000
1.006
0.999
0.998
1.002
1.002
0.999
0.998
1.002
1.000
0.992
1.015
1.000
0.925

0.987
0.813
0.825
0.748
0.749
0.814
0.814
0.827
0.797
0.816
0.813
0.818
0.823
0.821
0.823
0.802
0.814
0.784
0.800
0.981
0.803
0.678
0.970
0.167

1.013
1.223
1.217
1.340
1.337
1.226
1.234
1.226
1.248
1.209
1.230
1.238
1.213
1.213
1.221
1.253
1.228
1.272
1.253
1.020
1.226
1.520
1.031
5.129

Note: CL = confidence level. SE = standard error.

implying no differences between the groups with respect to
these observed characteristics.
Similar balance checks were conducted on the data from
Waves 2 through 4 to determine whether the balance attained
with the propensity scores persisted with the reduced samples
available for the follow-up data. Results were similar to those
presented above for Wave 1 data. No parameter estimate was
significantly different from zero at the alpha = 0.05 level.
ˆ are shown in Exhibit 8. As can
The distributional findings for p
ˆ ranges from a low of 0.1806 to a high of 0.7412.
be seen, p
The means of the distributions of the SVORI and non-SVORI
groups are similar—0.5232 for SVORI and 0.4934 for nonSVORI. It is also clear that the distributions have substantial
overlap.

33

Prisoner reentry experiences of adult males

ˆ distributions for adult male SVORI and non-SVORI evaluation
Exhibit 8. Characteristics of p
participants

N
ˆ mean
p
ˆ standard deviation
p
ˆ minimum
p
ˆ maximum
p

All

SVORI

Non-SVORI

1,697
0.5085
0.0862
0.1806
0.7412

863
0.5232
0.0823
0.1933
0.7412

834
0.4934
0.0876
0.1806
0.7020

To estimate the impact of SVORI program participation on a
variety of outcomes, weighted regression models were
ˆ . For these models,
estimated that used weights based on the p
ˆ were used to produce weights to estimate population
the p
average treatment effects. Specifically, the following weights,
wi for each subject i, were generated:
If subject i was a SVORI participant,
wi =

1
ˆ
pi

else
wi =

1
.
ˆi
1 − p

Outcome analyses are presented later in this report. In the next
chapter, the characteristics of the evaluation subjects are
presented, followed by examination of self-reported service
needs and service receipt across the four waves of interviews.

34

Characteristics of
the SVORI and NonSVORI Comparison
Respondents
This chapter provides descriptive information about the 1,697
adult male SVORI and non-SVORI respondents interviewed in
the 12 adult impact sites. Exhibit A-2 in Appendix A provides
the means, standard deviations, and t-statistics for the
variables discussed here. 16

DEMOGRAPHIC CHARACTERISTICS
The men in the SVORI and non-SVORI samples were almost
exclusively U.S. born (100% and 98% of the SVORI and nonSVORI respondents, respectively) and spoke English as a first
language (98% and 97%, SVORI and non-SVORI, respectively).
In addition, as shown in Exhibit 9, more than half (57%) of the
SVORI respondents were black and 32% were white. 17 The
SVORI sample included a higher percentage of black men and a
lower percentage of white men than the non-SVORI comparison
sample, which was 50% black and 37% white. Only 4% of both
groups identified themselves as Hispanic. 18
16

17

18

The results in this chapter are updated from material presented in
Lattimore et al. (2008). In a few cases, results differ slightly from
those presented earlier. The small differences result from decisions
made during the final cleaning of the data.
Respondents were allowed to select all that applied. Individuals who
reported more than one race are coded here as “other,” which also
includes American Indian or Alaska Native, Asian or East Indian,
and Native Hawaiian or other Pacific Islander.
Individuals are coded Hispanic if they chose “Hispanic, Latino or
Spanish,” regardless of whether they chose a race category.

35

Prisoner reentry experiences of adult males

Exhibit 9. Demographic
characteristics of
respondents at time of
interview, by group

Variable
Race

SVORI

Non-SVORI

Black*

57%

50%

White*

32%

37%

Hispanic

4%

4%

Other race
Age

8%

9%

28.9

29.3

61%

58%

Age at interview (mean)
Education
12th grade/GED

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

The average age of respondents in both samples was about 29
years at the time of the pre-release interview. As is evident
from Exhibit 9, respondents in both groups had substantial
educational deficiencies. Well over one third (39% SVORI and
42% non-SVORI) had not completed 12th grade or earned a
GED.
Given the diversity in the states selected for the impact
evaluation, it is not surprising that demographic characteristics
varied among the 12 sites. For example, Exhibit 10 shows the
average age at the time of the pre-release interview for
respondents by group and site. The overall mean age was 29
years; however, average age ranged from a low of 22.6 years
for Maine respondents to a high of 35.1 years for Indiana nonSVORI respondents. 19 Only the average age difference between
groups for the Iowa respondents was statistically significant
(27.0 years for SVORI, 28.9 years for non-SVORI).

There were racial and
ethnic differences among
the state samples.

Race and ethnic differences across the state samples (and,
within a state, between SVORI and non-SVORI samples) were
more substantial than was observed for age. As shown in
Exhibit 9, SVORI respondents were significantly more likely
than non-SVORI respondents to report being black (57% versus
50%) and significantly less likely to report being white (32%
versus 37%). Exhibit 11 shows the percentages of each group

19

36

Although the SVORI funding guidelines mandated that funds be used
for individuals 35 years or younger, many states requested and
received waivers of this requirement.

Respondent Characteristics

Exhibit 10. Age at time
of interview, by site and
group
34.6
32.5

27.5
27.1

28.9
30.0

27.2
27.4

26.2
26.0

28.8
28.4

26.4
27.1

Non-SVORI
28.4
27.8

22.7
22.6

25
Years

27.8
26.8

30

27.0
28.9

35

32.8
35.1

SVORI

40

20
15
10
5
0
IA*

IN KS ME MD MO NV OH OK PA SC WA

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

by site who reported that they were white or black.20 There
were considerable variations among sites, however. For
example, in Maryland, only 2% of the SVORI respondents were
white, whereas in Maine, 69% of the SVORI respondents and
73% of the non-SVORI respondents were white. Overall, where
there were statistically significant differences within a state,
more SVORI respondents than non-SVORI respondents
reported that they were black. This was true for 5 of the 12
sites—Kansas, Maryland, Missouri, Oklahoma, and South
Carolina. Furthermore, in three sites—Maryland, Missouri, and
Oklahoma—the proportion of white SVORI respondents was
significantly less than the proportion of white non-SVORI
respondents.

20

Respondents were also coded as Hispanic or other/multiracial—see
footnote 18.

37

Prisoner reentry experiences of adult males

Exhibit 11. Race (white or black), by site and group

White (SVORI)

95

100%

White (Non-SVORI)
Black (Non-SVORI)

78

81

Black (SVORI)

66

31

39

19

24

30

25
27

24
26
2

5

9

11

19

20
20

23

20%

39

37

39

36

35
25

34
32

40%

44

45

49

52

55

59

56

54
52
43

52

60

69
61

66
58

60%

70
66

73

80%

0%
IA

IN

KS*

ME

MD**

MO**

NV

OH

OK**

PA

SC*

WA

*p < 0.05 for test of significant difference between SVORI and non-SVORI within site in the proportion of black
respondents.
**p < 0.05 for test of significant difference between SVORI and non-SVORI within site in the proportion of black
respondents and in the proportion of white respondents.

There was also considerable state-level variation in educational
attainment, as can be seen in Exhibit 12. 21 In Iowa, more than
80% of respondents had either finished high school or obtained
a GED. Pennsylvania, Indiana, and Kansas also had high rates
of high school or GED completion. But in South Carolina,
Maryland, and Washington, less than half of the respondents
reported that they had a high school degree or GED. In only
one state was educational attainment significantly different
between SVORI and non-SVORI sample members: in Nevada,
significantly more SVORI respondents (79%) than non-SVORI
respondents (52%) reported that they had completed 12th
grade or earned a GED.

21

38

Respondents could have completed the GED during their current
incarceration. The respondents were asked whether they had
completed 12th grade or had received a GED at the time of the prerelease interview.

Respondent Characteristics

Exhibit 12. Completed
12th grade or obtained a
GED, by site and group

SVORI

100%

45
48

51

52

46
47

61

66

40%

SC

WA

39

40

48

50

60%

58

58

66

65

70
65

80%

79
74

79

79

84
89

Non-SVORI

20%
0%
IA

IN

KS

ME

MD

MO

NV*

OH

OK

PA

*p < 0.05 for test of significant difference between SVORI and non-SVORI within
site.

HOUSING
More than 1 in 10
respondents reported that
they were primarily
homeless, living in a
shelter, or had no set
place to live during the 6
months prior to
incarceration.

During the 6 months prior to incarceration, the most common
housing situation reported by the respondents was living in a
house or apartment that belonged to someone else. Just under
half (46%) of both SVORI and non-SVORI respondents reported
primarily living in a house or apartment that belonged to
someone else. About one third (35% SVORI and 32% nonSVORI) reported living primarily in their own house or
apartment. Finally, more than 1 in 10 (12%) of both SVORI and
non-SVORI respondents reported as their primary housing
situation that they were homeless, living in a shelter, or had no
set place to live.

FAMILY AND CHILDREN
Although about 40% of both groups reported that they were
either currently married or in a steady relationship (39%
SVORI, 40% non-SVORI), only small proportions reported being
married (9% and 10%, SVORI and non-SVORI, respectively).
Of those who reported that they were currently married or in a
steady relationship, 59% of SVORI respondents and 67% of
non-SVORI respondents said that they lived with that person
before incarceration.

39

Prisoner reentry experiences of adult males

About 60% of
respondents reported that
they were fathers of
minor children.
About three quarters of
these fathers reported
that they were married or
in a steady relationship at
the time of the interview.
Nearly all fathers
required to pay child
support reported that they
owed back child support.

Exhibit 13. Percentages
of fathers reporting on
child care or child
support responsibilities,
by group

Most study participants from both groups (59% SVORI and
61% non-SVORI) reported having children under age 18. On
average, respondents with children had more than two (2.22
SVORI and 2.29 non-SVORI). About three quarters of these
fathers reported that they were currently married or in a steady
relationship (77% SVORI and 74% non-SVORI). Furthermore,
as can be seen in Exhibit 13, about half of those with children
under 18 indicated that they had primary care responsibilities
for their children (either with or without a partner) during the 6
months prior to incarceration (47% of SVORI respondents and
49% of non-SVORI respondents). Nearly one third of the
fathers (30% SVORI and 32% non-SVORI) reported that they
were required to pay child support during the 6 months prior to
incarceration, and, of those, more than half reported that they
had made court-ordered payments (59% SVORI and 56% nonSVORI). Nearly all fathers required to pay child support
reported that they owed back child support (93% SVORI and
91% non-SVORI), and most of these respondents reported that
they owed more than $5,000 (62% and 55%, SVORI and nonSVORI, respectively). As is evident in Exhibit 13, SVORI and
non-SVORI respondents were similar on these family
background characteristics.

100%
80%
60%

93

SVORI

91

Non-SVORI

59
47

49

40%

30

56

32

20%
0%
Primary care
f or children
under 18

Required to
pay child
support a

Made
required
child
support b

Owed back
child
support b

Note: Differences between SVORI and non-SVORI are not statistically significant
at the 0.05 level.
a
Of those with children under 18 years of age.
b
Of those required to pay child support.

40

Respondent Characteristics

Nearly all SVORI and non-SVORI respondents (97% of both
groups) reported having people in their lives they considered to
be family. Respondents also reported that their family provided
an important source of emotional support (data not shown).
Nearly all respondents (88% of SVORI and 91% of non-SVORI)
agreed or strongly agreed that they felt close to their family
and wanted their family to be involved in their life (95% SVORI
and 96% non-SVORI).
Although they provided a substantial source of emotional
support for these men, family members also may have served
as a negative influence. As shown in Exhibit 14, about three
quarters of both SVORI and non-SVORI respondents reported
having family members who had been convicted of a crime or
incarcerated, and nearly three quarters (72% SVORI and 74%
non-SVORI) reported having family members who had
problems with drugs or alcohol.

About three quarters of
respondents reported
having family members
who had been convicted
of a crime or
incarcerated.

Exhibit 14. Criminal history and substance use of family and peers, by group

SVORI
100%

Non-SVORI
83

80%

75

76

75

74

72

83

81

81

82

83

74

60%

40%

20%

0%
Family who
have been
convicted

Family who
have been
incarcerated

Family with
drug or
alcohol
problems

Friends who
have been
convicted

Friends who
have been
incarcerated

Friends with
drug or
alcohol
problems

Note: Differences between SVORI and non-SVORI are not statistically significant at the 0.05 level.

41

Prisoner reentry experiences of adult males

A large majority of
respondents reported
having criminally
involved friends prior to
incarceration.

Similarly, the reported prevalence of illegal behavior and
problems with substance use among friends was also high. A
large majority of respondents reported having criminally
involved friends prior to incarceration. The majority of both
SVORI and non-SVORI respondents reported having friends
prior to incarceration who had been convicted of a crime (83%
of both groups) or incarcerated (81% of both groups). The
respondents also reported that, prior to incarceration, they had
friends who had problems with drugs or alcohol (82% SVORI
and 83% non-SVORI).

SUBSTANCE USE AND PHYSICAL AND
MENTAL HEALTH
Respondents were asked a variety of questions about their preprison alcohol and drug use, as well as their substance abuse
treatment experiences. They were also asked about their
lifetime and current experiences with a variety of physical
illnesses. In addition, they were asked to respond to a series of
items that comprise three well-known scales—the SF-12
physical health scale, the SF-12 mental health scale, and the
SA-45 Global Severity Index (GSI; Ware et al., 2002; Strategic
Advantages, 2000).
Substance Use and Treatment

Nearly all of the
respondents reported
having used alcohol and
drugs during their
lifetimes.

Self-reports on “ever
using” indicate somewhat
higher usage among the
non-SVORI respondents
for most drugs.

42

Nearly all of the respondents reported having used alcohol and
drugs during their lifetimes. The majority of both groups
reported using alcohol (96% SVORI and 97% non-SVORI), and
the average age of first use was about 14 years (13.7 and 13.6
for the SVORI and non-SVORI respondents, respectively).
Similarly, nearly all respondents in both groups reported having
used marijuana (92% SVORI and 94% non-SVORI), again
reporting a young age of first use (13.9 and 14.1 for the SVORI
and non-SVORI respondents, respectively). Exhibit 15 shows
responses for lifetime use for the most common drugs.
As can be seen, self-reports on “ever using” indicate somewhat
higher usage among the non-SVORI respondents for most
drugs. More than half of all respondents reported having used
cocaine (53% and 58% of the SVORI and non-SVORI
respondents, respectively), and nearly one half reported having
used hallucinogens (43% and 49%, SVORI and non-SVORI,

Respondent Characteristics

Exhibit 15. Lifetime substance use, by group

92
94

Marijuana
53

Cocaine*

58
43

Hallucinogens*

49
26

Amphetamines

30
25

Tranquilizers*

31
24

Pain relievers*

30
18

Heroin*

23
18

Sedatives

21
16

Stimulants*

SVORI

20

Non-SVORI

15
16

Inhalants
0%

20%

40%

60%

80%

100%

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

respectively). Fewer respondents reported using other
substances. 22
There was considerable variability among the states with
respect to self-reports of ever using specific drugs. Exhibit 16
presents the percentages of respondents in each site and group
who reported ever using cocaine, heroin, and hallucinogens.
Only 22% of the Missouri SVORI respondents reported ever
using cocaine in comparison with 82% of the non-SVORI
respondents from Maine. Self-reported heroin use ranged from
a low of 3% (Missouri SVORI) to a high of 64% (Maine nonSVORI), whereas self-reported hallucinogen use ranged from
21% (Maryland SVORI) to 86% (Maine non-SVORI).

22

Less than 10% reported ever using methadone (6% and 9% for the
SVORI and non-SVORI respondents, respectively) or anabolic
steroids (2% for both the SVORI and non-SVORI respondents).

43

Prisoner reentry experiences of adult males

Exhibit 16. Lifetime use of cocaine, heroin, and hallucinogens, by site and group

Cocaine
Site
IA
IN
KS
ME
MD
MO
NV
OH
OK
PA
SC
WA

SVORI
75%
72%
30%*
69%
48%
22%*
36%
34%
55%
49%
56%
66%

Heroin

Non-SVORI
65%
67%
62%*
82%
52%
58%*
50%
50%
54%
59%
53%
60%

SVORI
14%
17%
9%
49%
49%
3%*
5%*
11%
7%
12%
7%
38%

Non-SVORI
13%
17%
21%
64%
49%
26%*
16%*
13%
12%
17%
9%
23%

Hallucinogens
SVORI
68%
47%
48%
83%
21%*
56%
48%
38%
62%
39%
22%*
76%

Non-SVORI
65%
49%
64%
86%
36%*
62%
48%
32%
63%
53%
31%*
63%

*p < 0.05 for test of significant difference between SVORI and non-SVORI within site.

In some sites, more nonSVORI than SVORI
respondents reported ever
using various types of
drugs.

About two thirds of
respondents reported
having used one or more
illicit drugs during the 30
days prior to their
imprisonment.

44

There were only a few statistically significant differences
between SVORI and non-SVORI groups within site; in each of
these cases more non-SVORI than SVORI respondents reported
ever using various types of drugs. Specifically, non-SVORI
respondents in Kansas and Missouri were much more likely
than SVORI respondents in those states to report having used
cocaine, and non-SVORI respondents in Missouri and Nevada
were more likely than SVORI respondents in those states to
report heroin use. Finally, in Maryland and South Carolina, nonSVORI respondents were more likely than SVORI respondents
to report hallucinogen use.
There were few differences between the two groups with
respect to reported drug use during the 30 days prior to their
current incarceration. About two thirds of both groups reported
having used one or more illicit drugs during the 30 days prior to
their imprisonment (66% and 69% for the SVORI and nonSVORI respondents, respectively). Exhibit 17 shows that there
were SVORI/non-SVORI differences among the sites on this
measure although the differences between groups within site
are not statistically significant at the 0.05 level. Reported use
ranged from a high of 84% of non-SVORI respondents in Maine
to a low of 46% of SVORI respondents in Pennsylvania.

Respondent Characteristics

SVORI
Non-SVORI
83
68
62
46

52

60

58
62

58

60%

72

79

82
69

77
65
68

73
64

74
76

80%

83
82

100%
84

Exhibit 17. Substance
use during the 30 days
prior to incarceration,
by site and group

40%
20%
0%
IA

IN

KS

ME MD MO NV

OH OK PA SC WA

Note: Within-site differences between SVORI and non-SVORI are not significant
at the 0.05 level.

Exhibit 18 compares the two groups’ reported use during the 30
days prior to incarceration for the most commonly reported
drugs. More than half of both SVORI and non-SVORI
respondents reported using marijuana; approximately one
quarter of all respondents reported using cocaine.
Exhibit 18. Use of
specific substances
during the 30 days prior
to incarceration, by
group

52
53

Marijuana
22
26

Cocaine
Amphetamines

13
14

Pain relievers

11
14

SVORI
Non-SVORI

9
9

Hallucinogens
0%

20%

40%

60%

80%

100%

Note: Within-site differences between SVORI and non-SVORI are not significant
at the 0.05 level.

45

Prisoner reentry experiences of adult males

More than half of
respondents had received
treatment for a substance
use or mental health
problem at some point
during their lifetime.

More than half of SVORI and non-SVORI respondents had
received treatment for a substance use or mental health
problem at some point during their lifetime (56% and 55% of
SVORI and non-SVORI, respectively). Of these, about one
quarter had received treatment for alcohol abuse or
dependency (25% of SVORI respondents and 28% of nonSVORI respondents), and more than one third reported that
they had received treatment for drug abuse or dependence
(42% SVORI and 41% non-SVORI). On average, those who had
received treatment had started a treatment program on more
than two separate occasions.
As shown in Exhibit 19, the percentage of respondents
reporting receiving treatment prior to prison varied
considerably across sites (but not within). Whereas less than
30% of Nevada respondents reported having previously
received treatment for substance use, about two-thirds of those
in Iowa reported that they had participated in substance use
treatment prior to their current incarceration.

100%

SVORI
Non-SVORI

80%

41
42

36

OK

28

33
37

OH

21

28

36
33

34
32

KS

44

48
50

IN

40%

45

47
48

60%

51
50

57

58

67
64

Exhibit 19. Any
substance use treatment
prior to current
incarceration, by site
and group

20%

0%
IA

ME MD MO

NV

PA

SC WA

Note: Within-site differences between SVORI and non-SVORI are not significant
at the 0.05 level.

46

Respondent Characteristics

Physical Health

Overall, the study
participants reported
currently experiencing
few physical health
problems.

Overall, the study participants reported currently experiencing
few physical health problems. Most respondents rated their
current physical health as excellent or very good (65% of
SVORI and 63% of non-SVORI). The percentages of subjects in
each group who reported ever or currently having specific
diseases are shown in Exhibits 20 and 21.

Exhibit 20. Lifetime health problems, by group

20

Asthma

19
17
16

High blood pressure
15

Chronic back pain

16
6

Tuberculosis

7
5
5

Heart trouble

5

Arthritis

6
3

Hepatitis B or C

5
2
2

Diabetes

SVORI
Non-SVORI

1
1

HIV positive/AIDS
0%

5%

10%

15%

20%

25%

Note: Differences between SVORI and non-SVORI are not significant at the 0.05 level.

Asthma, high blood pressure, and chronic back pain were the
most commonly reported. Only 1% of the respondents reported
that they were HIV positive or had been diagnosed with AIDS,
whereas about 4% reported that they had been diagnosed with
hepatitis B or C. There were no statistically significant
differences in the reports of physical illnesses between the two
groups.

47

Prisoner reentry experiences of adult males

Exhibit 21. Current health problems, by group

11

Asthma

10
9

High blood pressure

8
11

Chronic back pain

13
0

Tuberculosis 0
3
3

Heart trouble

5
5

Arthritis
3

Hepatitis B or C*

5

SVORI

1

Diabetes

2

Non-SVORI

1
1

HIV positive/AIDS
0%

5%

10%

15%

20%

25%

Note: Differences between SVORI and non-SVORI are not significant at the 0.05 level.

Mental Health

There were no differences
between SVORI and nonSVORI respondents in
their general measures of
physical and mental
functioning and mental
health.

48

There were also no differences between SVORI and non-SVORI
respondents in their scores on the four scales measuring
physical and mental functioning (the SF-12 scales) and mental
health (the SA-45 GSI and Positive Symptom Total [PST]).
Scores on the SF-12 physical health scale were above 50
(53.63 for SVORI respondents, 53.34 for non-SVORI
respondents). Furthermore, more than half of each group
responded that they had no limitations with respect to each of
the five items that constitute the physical health scale (59% of
SVORI respondents and 56% of non-SVORI respondents).
Scores on the SF-12 mental health scale were nearly 50 (48.93
for SVORI respondents, 48.51 for non-SVORI respondents).
Both groups scored less than 70 on the GSI, which has a range
of 45 to 225; higher scores indicate more psychopathology
(66.64 and 68.09 for the SVORI and non-SVORI respondents,
respectively). Average scores on the PST index were 13 for
both SVORI and non-SVORI respondents, meaning that
respondents reported experiencing, on average, 13 of the 45

Respondent Characteristics

symptoms included in the SA-45 during the 7 days prior to the
interview.

Non-SVORI respondents
were significantly more
likely than SVORI
respondents to indicate
symptoms of hostility and
psychoticism.

Exhibit 22. Average
scores on Brief
Symptom Inventory
subscales, by group

In addition to the GSI, the SA-45 includes subscales indicating
symptoms of specific psychopathologies. Of the nine subscales,
there were statistically significant differences for two
measures—in each case indicating that the non-SVORI
respondents were slightly worse on these measures than the
SVORI respondents. Results are shown in Exhibit 22. Scores on
these subscales range from a low of 5 to a high of 25, and all
results for the respondents were on the lower end of the range.
Scores were similar between groups for anxiety, depression,
interpersonal sensitivity, obsessive-compulsive disorder,
paranoid ideation, phobic anxiety, and somatization. NonSVORI respondents were significantly more likely than SVORI
respondents to indicate symptoms of hostility (6.41 for SVORI
respondents, 6.69 for non-SVORI respondents) and
psychoticism (6.58 for SVORI respondents, 6.89 for non-SVORI
respondents).

Measure
Anxiety scale
Depression scale
Hostility scale*
Interpersonal sensitivity scale
Obsessive-compulsive scale
Paranoid ideation scale
Phobic anxiety scale
Psychoticism scale*
Somatization scale

SVORI
7.42
8.31
6.41
7.50
8.12
8.84
6.42
6.58
7.05

Non-SVORI
7.67
8.45
6.69
7.60
8.17
8.85
6.56
6.89
7.16

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

Depression was cited as
the most common reason
for the treatment.

As reported previously, more than half of SVORI and nonSVORI respondents had received treatment for a substance use
or mental health problem at some point during their lifetime
(56% and 55% of SVORI and non-SVORI, respectively). Of
those who reported that they had ever received mental health
treatment, depression was cited as the most common reason
for the treatment. About 20% of each group reported that they
had received care for depression or dysthymia (19% SVORI and
20% non-SVORI). Ten percent or more reported that they had
received treatment for attention deficit hyperactivity disorder
(12% of SVORI respondents and 13% of non-SVORI

49

Prisoner reentry experiences of adult males

respondents) or bipolar disorder (10% SVORI and 12% nonSVORI). Less than 10% reported that they were currently
receiving treatment for any mental health problem. Of those
who reported that they were currently receiving treatment, the
most common diagnoses were depression or dysthymia (6%
and 10%, SVORI and non-SVORI, respectively) and bipolar
disorder (5% and 6%, SVORI and non-SVORI, respectively).

Most respondents
described their mental
health status at the time
of the pre-release
interview as excellent or
very good.

Most respondents described their mental health status at the
time of the pre-release interview as excellent or very good
(52% SVORI and 49% non-SVORI). During their current period
of incarceration, 13% of SVORI respondents were prescribed
medication for emotional problems, and 22% felt they needed
treatment for mental health problems. The non-SVORI
respondents were significantly more likely to have been
prescribed medication for a mental or emotional problem while
incarcerated (19%) and to feel in need of treatment for mental
health problems (29%).

EMPLOYMENT HISTORY AND FINANCIAL
SUPPORT
This subsection covers the respondents’ employment history
prior to incarceration and describes additional sources of
financial support.
Employment History

Most subjects reported
having worked at some
time prior to
incarceration.

As shown in Exhibit 23, most subjects reported having worked
at some time prior to incarceration—89% of SVORI versus 92%
of non-SVORI—and about two thirds of both groups reported
having a job during the 6 months prior to incarceration (64%
and 68%, SVORI and non-SVORI, respectively). Although these
differences are statistically significant (at 0.05 levels), they are
relatively small in magnitude.
Some variation in the percentage of respondents who had
worked during the 6 months prior to entering prison was evident
across the 12 sites (Exhibit 24). More than 70% of SVORI
respondents in Iowa, Maine, and South Carolina reported
working during the 6 months prior to their incarceration. In
contrast, only about 40% of all respondents in Washington
reported working immediately prior to incarceration. Differences
between SVORI and non-SVORI respondents were not
statistically significant at the 0.05 level in any state.

50

Respondent Characteristics

Exhibit 23. Employment
prior to incarceration,
by group

100%

SVORI

92

89

Non-SVORI

80%

68

64

60%
40%
20%
0%
Ever held job*

Held job in 6 months before
prison*

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

SVORI

100%

72
73
63
63

57

38

44

47

48

60%

71

66

64
68

56
57

67
72

74
70

67

67

80%

81

82

Non-SVORI
72

Exhibit 24. Employment
during the 6 months
prior to incarceration,
by site and group

40%

20%

0%
IA

IN

KS

ME MD MO NV

OH

OK PA

SC WA

Note: Within-site differences between SVORI and non-SVORI are not significant
at the 0.05 level.

For those who worked during the 6 months prior to
incarceration, about three quarters of respondents described
their most recent job as a permanent job (75% SVORI and
73% non-SVORI) for which they received formal pay
(Exhibit 25).

51

Prisoner reentry experiences of adult males

Exhibit 25.
Characteristics of
respondents’ jobs prior
to incarceration, by
groupa

SVORI

100%
80%

93

Non-SVORI
75

73

74

94

72

60%
40%
20%
0%
Held a permanent Worked for formal Worked more than
job
pay
20 hrs/wk
Note: Differences between SVORI and non-SVORI are not significant at the 0.05
level.
a
Among respondents who worked during the 6 months prior to incarceration.

Almost all who worked reported that they had worked more
than 20 hours a week, working an average of about 42 hours
(41.7 hours per week for SVORI respondents and 41.8 hours
per week for non-SVORI respondents). The SVORI respondents
reported a slightly higher average hourly rate of $10.91
compared with the average $10.13 reported by the non-SVORI
respondents.

When asked about the
longest they had ever
worked at one job since
they were 18, most
respondents reported less
than 2 years.

Although the majority described their most recent job as a
permanent job, many of the respondents who had worked
reported having had more than one job during the 6 months
prior to incarceration. More than one third of the sample (35%
SVORI, 36% non-SVORI) reported having had two or more jobs
during the 6 months prior to incarceration. Furthermore, well
over one third (35% SVORI, 38% non-SVORI) reported that
they worked at the job for 3 months or less. When asked about
the longest they had ever worked at one job since they were
18, most respondents reported less than 2 years (61% SVORI,
62% non-SVORI).
The jobs that respondents typically held were blue-collar jobs.
More than one third of the respondents in both groups who had
been employed during the 6 months prior to incarceration

52

Respondent Characteristics

reported that the last job they had was as a laborer, which
includes construction workers, day laborers, landscapers, and
roofers (35% SVORI, 36% non-SVORI). About one fifth of
respondents (22% of each group) had worked in the service
industry as cooks, waiters, janitors, cashiers, and dishwashers.
Many respondents also reported working as skilled craftsmen
(15% SVORI, 17% non-SVORI) or equipment operators (16%
SVORI, 13% non-SVORI). Few respondents reported having
professional or technical occupations or jobs as managers or
administrators (4% of each group).
Financial Support

Nearly half of the
respondents reported
supporting themselves
with income from illegal
activities during the 6
months prior to
incarceration.

The respondents were asked how they had supported
themselves, in addition to legal employment, during the 6
months prior to incarceration. Nearly half of the respondents
reported supporting themselves with income from illegal
activities (46% and 43% of SVORI and non-SVORI,
respectively). Another one third received support from family
(32% of both groups). Fewer reported receiving financial help
from friends (17% of SVORI respondents, 14% of non-SVORI
respondents) or the government (11% of SVORI respondents,
10% of non-SVORI respondents).
Exhibit 26 shows the sources of financial support for SVORI and
non-SVORI respondents, disaggregated by their employment
status during the 6 months prior to incarceration. As shown in
the exhibit, within employment status there were relatively few
differences between SVORI and non-SVORI respondents with
respect to whether they reported receiving financial support
from each of the four sources.
The most substantial difference between the reports of those
working and not working was in reports of support from illegal
activities. More than 60% of those who were not employed
during the 6 months prior to incarceration reported financial
support from illegal activities, compared with less than 40% of
those who reported working during that period. For both SVORI
and non-SVORI respondents, those who held a job prior to
incarceration were somewhat less likely than those who had no
job to receive financial support from friends, the government,
or other sources.

53

Prisoner reentry experiences of adult males

Exhibit 26. Sources of income during the 6 months prior to incarceration, by employment
status and group

100%

Held job (SVORI)
Held job (Non-SVORI)
No job (SVORI)

80%

63

63

No job (Non-SVORI)

14
3

6

19

34

35
8

8

14

16

18

14

20%

12

20

32

34

31

40%

31

60%

0%
Family

Friends

Government

Illegal

Other*

*p < 0.05 for test of significant difference between SVORI (Held Job) and non-SVORI (Held Job).

CRIMINAL HISTORY, VIOLENCE,
VICTIMIZATION, AND GANG INVOLVEMENT
This subsection describes respondents’ involvement with the
criminal and juvenile justice systems prior to incarceration and
outlines preincarceration perpetration of violence and
victimization. Respondents’ reports of gang involvement are
also briefly described.
Criminal History

Respondents reported
considerable involvement
with the criminal justice
system prior to their
current incarceration.

54

SVORI and non-SVORI respondents reported considerable
involvement with the criminal justice system prior to their
current incarceration (Exhibit 27). On average, the respondents
were 16 years old at the time of their first arrest and had been
arrested more than 12 times. In addition to their current term
of incarceration, most respondents had served a previous
prison term, with the non-SVORI group being significantly more
likely to report a prior prison term (83% of SVORI, 87% of nonSVORI). Also, the non-SVORI respondents reported significantly
more incarcerations, on average, than the SVORI group (1.20
for SVORI, 1.47 for non-SVORI).

Respondent Characteristics

Exhibit 27. Criminal
history of respondents,
by group

Criminal History
Age at first arrest (mean)
Times arrested (mean)
Times convicted (mean)
Ever been previously incarcerated*
Times previously incarcerated (mean)*

SVORI
15.92
12.42
5.48
83%
1.20

Non-SVORI
16.03
13.14
5.70
87%
1.47

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

The two groups were similar in self-reported juvenile
detentions. Overall, about half (51% and 49% of the SVORI
and non-SVORI respondents, respectively) reported that they
had spent time in a juvenile correctional facility for committing
a crime. Of those who reported a juvenile detention, they had
been detained, on average, 3.5 times (3.58 times for SVORI,
3.49 times for non-SVORI).

About 40% of
respondents reported that
they were currently
serving time for a violent
crime.

Exhibit 28 shows the conviction offense(s) that were reported
by the respondents. 23 About 40% of respondents reported that
they were currently serving time for a person/violent crime
(42% SVORI and 40% non-SVORI)—including 19% of SVORI
and 16% of non-SVORI respondents and 15% of SVORI and
13% of non-SVORI who reported that they were currently
serving time for assault and robbery, respectively. About 25%
reported a property crime (24% and 27% of the SVORI and
non-SVORI respondents, respectively), most commonly
burglary. SVORI respondents were significantly more likely than
non-SVORI respondents to report that their current
incarceration was for a drug crime (36% SVORI, 31% nonSVORI)—in particular for drug dealing/manufacturing (21% of
SVORI, 15% of non-SVORI). SVORI respondents were
significantly less likely to report that their current incarceration
was for a public order crime (17% SVORI, 22% non-SVORI).
Public order offenses include probation and parole violations;
members of the non-SVORI group were more likely to report
that their current incarceration was for a violation of probation
or parole (27% and 35% of SVORI and non-SVORI
respondents, respectively).

23

Two percent of the SVORI and 1% of the non-SVORI respondents
reported that their conviction offense was “other.” This category
includes unspecified felonies, gang activity, and habitual offender
violations.

55

Prisoner reentry experiences of adult males

Exhibit 28. Conviction
offenses for current
incarceration, by group

40

Person/Violent

42
27

Property

24
31

Drug*

36

Non-SVORI
22

Public order*

SVORI

17

0%

20%

40%

60%

80%

100%

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

More than two thirds of
respondents reported
violent behavior prior to
incarceration.
Most also reported being
victims of violence.

Perpetration of Violence

During the 6 months prior to incarceration, more than two
thirds of both SVORI and non-SVORI respondents (69% and
67%, respectively) reported violent behavior (including threats
of violence).
Victimization

Most respondents also reported being victims of violence. More
than half of the respondents (59% SVORI and 58% non-SVORI)
reported being victimized either through threats or use of
violence during the 6 months prior to incarceration.
Gang Membership

Very few respondents in both groups (5% of SVORI and 6% of
non-SVORI) reported being a member of a gang. Of the small
number of respondents in a gang, about half (53% of SVORI,
52% of non-SVORI) considered their gang to be family and
about half reported that they had relatives who were members
of the gang (55% SVORI, 58% non-SVORI).

IN-PRISON EXPERIENCES
This subsection describes respondents’ in-prison experiences on
several dimensions, including sentence length, disciplinary
infractions, and in-prison victimization. This is followed by a

56

Respondent Characteristics

description of in-prison work and a discussion of interaction
with family during prison.
Sentence Length

At the time of the pre-release interview, SVORI respondents
had been incarcerated significantly longer than non-SVORI
respondents (an average of 2.8 years and 2.3 years,
respectively). The difference between these is due, primarily, to
statistically significant differences in 5 of the 12 sites, as can be
seen in Exhibit 29. In particular, in Kansas, Missouri, Nevada,
and Oklahoma, SVORI respondents had served, on average,
about 2 years longer than the non-SVORI respondents. In
Washington, SVORI respondents had been incarcerated for 1
year longer than non-SVORI respondents, on average.
Respondents in Maine reported the shortest lengths of stay of
slightly more than a year, whereas stays of about 2 years were
reported by most respondents in the remainder of sites, without
statistically significant differences in length of stay.

SVORI

4.6

3.8

2.6

2.7
1.5

2.3

2.1
2.3

2.6
2.1
1.8

2.2
1.3
1.4

2

2.0
2.2

Years

3

3.1

3.4

4

3.9

3.9

Non-SVORI

1.7

5

3.7

Exhibit 29. Average
duration of incarceration
at time of interview, by
site and group

2.1
1.9

SVORI respondents had
been incarcerated
significantly longer than
non-SVORI respondents.

1
0
IA

IN KS* ME MD MO* NV* OH OK* PA SC WA*

*p < 0.05 for test of significant difference between SVORI and non-SVORI within
site.

Disciplinary Infractions and Administrative Segregations

SVORI respondents also reported more disciplinary infractions
and administrative segregations than were reported by the
non-SVORI respondents. As shown in Exhibit 30, 64% of SVORI
respondents reported at least one disciplinary infraction,

57

Prisoner reentry experiences of adult males

Exhibit 30. Disciplinary
infractions and
administrative
segregations during
current incarceration, by
group

Infractions and Segregations
Disciplinary Infractions
None
One
More than one
Administrative Segregations
None
One
More than one

SVORI

Non-SVORI

35%
17%
47%

43%
17%
40%

55%
19%
26%

60%
18%
22%

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

compared with 57% of non-SVORI respondents. Less than half
reported administrative segregation during the current term of
incarceration (45% of SVORI and 40% of non-SVORI). These
differences between groups are statistically significant but may
simply reflect the longer lengths of stay reported by the SVORI
respondents. 24
In-prison Victimization

Slightly more than half of
all respondents reported
being victimized during
the current incarceration.

Reported in-prison victimization was similar for the two groups.
Slightly more than half of all respondents (55% and 54% of
SVORI and non-SVORI respondents, respectively) reported
being victimized during the current incarceration—somewhat
fewer than reported being victimized during the 6 months prior
to incarceration (59% SVORI and 58% non-SVORI). This
measure includes both threat of violence (including someone
threatening to hit the respondent with a fist or anything else
that could hurt him or someone threatening to use a weapon on
him) and perpetration of violence (including someone throwing
anything at the respondent; pushing, grabbing, shoving,
slapping, kicking, biting, hitting with a fist, or using a weapon
on him; or the respondent needing medical attention for violent
acts directed at him). The reported severity of victimization was
low. On a 36-point victimization scale, SVORI and non-SVORI
respondents scored an average of 2.7 and 2.9, respectively. 25
This victimization severity was somewhat lower than was

24

25

58

Longer lengths of stay expose subjects to greater opportunity to
commit infractions and receive administrative segregation; in other
words, the period at risk is longer.
Responses to six victimization items were coded 0 though 6, with
higher values indicating more frequent victimization. (Response
options ranged from “never” to “daily.”) The six items were
summed to create the in-prison victimization scale.

Respondent Characteristics

reported for the 6 months prior to incarceration (3.85 SVORI,
3.75 non-SVORI).
In-prison Work

SVORI
81
82
69
62

66
58

48

52

40

44

52

56

54

64

71
75

80%
60%

90

Non-SVORI

83

100%

69

Exhibit 31. Institutional
employment, by site and
group

Nearly two thirds of the respondents (63% of SVORI and 61%
of non-SVORI) said that they had a job in the institution where
they were incarcerated. On average, respondents with prison
jobs spent about 23 hours per week working (23.8 and 22.3
hours for SVORI and non-SVORI respondents, respectively).
Most of the jobs were prison service as opposed to prison
industry jobs. About 60% reported having a prison service job
(60% of SVORI, 57% of non-SVORI), but only 4% of both
groups reported having a prison industry job. As can be seen in
Exhibit 31, respondents in South Carolina and Pennsylvania
were most likely to report working while in prison, and those in
Nevada were the least likely. A significant difference between
SVORI and non-SVORI respondents was observed only for
Indiana (56% and 29%, respectively).

65
69

Nearly two thirds of the
respondents said that they
had a job in the
institution where they
were incarcerated.

23

29

30

40%
20%
0%
IA

IN*

KS

ME MD MO NV

OH OK PA

SC WA

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

Very few respondents
reported having a workrelease job.

Very few respondents reported having a work-release job. Only
3% of SVORI and 4% of non-SVORI respondents reported that
they were on work release. Those with work-release jobs
reported working more hours than those with institution jobs.
SVORI respondents reported working significantly more hours
than non-SVORI respondents (40 and 31 hours, respectively).

59

Prisoner reentry experiences of adult males

As shown in Exhibit 32, only in Pennsylvania did more than
10% of the respondents participate in work release. 26 For the
remaining states, less than 10% (and usually many fewer)
reported having a work-release job.
Exhibit 32. Work-release
participation, by site
and group

35%
29

SVORI

30%

Non-SVORI

25%

16

20%
15%

6

IA*

ME

MD

MO

NV

OH

3

2
0

2
2

2
0

2
0

0

0%

3

5
2

4

5%

5

6

8

10%

OK

PA

SC

WA

Note: Values for IN and KS were 0%.
*p < 0.05 for test of significant difference between SVORI and non-SVORI within
site.

Family

Most respondents
indicated that family
members served as an
important source of
emotional support during
incarceration.

Most respondents (97% of both groups) indicated that they had
people in their lives that they considered to be family and that
these family members served as an important source of
emotional support. A scale was created to represent the degree
of family emotional support that respondents felt at the time of
the pre-release interview. Respondents were asked the degree
to which they agreed with 10 statements about their
relationships with their family, such as “I have someone in my
family who understands my problems” and “I have someone in
my family to love me and make me feel wanted.”27 The items
were combined to create a scale with possible values ranging
26
27

60

Most respondents in Pennsylvania were interviewed at a community
corrections center, where work-release jobs were common.
Response categories were “strongly agree,” “agree,” “disagree,” and
“strongly disagree.” Values of 0 through 3 were assigned to
response categories, with higher values representing greater family
emotional support. The values for each of the 10 items were
summed to create the family emotional support scale.

Respondent Characteristics

from 0 to 30 and higher scores indicating higher levels of family
emotional support. There were no significant differences
between SVORI respondents and non-SVORI respondents on
this measure (21.63 for SVORI, 21.35 for non-SVORI).
Respondents were also asked about the frequency of contact
with family members and friends. Response options for each
type of contact ranged from “never” to “daily.” SVORI and nonSVORI respondents reported similar frequencies of contact with
their family members through phone calls or mail (Exhibit 33).
About 40% of both groups reported weekly phone or mail
contact with family members. Both SVORI and non-SVORI
respondents reported less frequent phone and mail contact with
friends than with family. In-prison visits with family members
were less frequent than phone calls and mail. However, on
average, SVORI respondents received more visits from family
and non-family members than the comparison group.
Exhibit 33. Frequency of in-prison contact with family members and friends, by group

Form of Contact
Phone Contact
Never
A few times
Monthly
Weekly
Daily
Mail Contact
Never
A few times
Monthly
Weekly
Daily
In-Person Visits
Never
A few times
Monthly
Weekly
Daily

Contact with
Family Members
SVORI
Non-SVORI

Contact with Friends
SVORI
Non-SVORI

16%
15%
16%
38%
14%

18%
14%
16%
36%
16%

47%
16%
13%
16%
8%

52%
13%
11%
15%
9%

10%
17%
23%
41%
9%

9%
18%
21%
41%
10%

30%
19%
16%
30%
6%

36%
17%
16%
25%
6%

35%*
23%
17%
21%
3%

43%*
21%
18%
17%
2%

64%
16%
8%
10%
2%

71%
13%
6%
8%
1%

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

61

Prisoner reentry experiences of adult males

Respondents were also asked whether the amount of each type
of contact with family and friends was currently more, about
the same, or less than when they were first incarcerated (i.e.,
during the first 6 months of incarceration). Almost half of the
respondents in both groups reported that they had about the
same amount of contact with family and friends as they did
when they were first incarcerated (Exhibit 34). More
respondents reported having less contact, rather than more
contact, with family and friends than when they were first
incarcerated.
Exhibit 34. Amount of contact with family members and friends at time of interview
compared with contact when first incarcerated

More contact (SVORI)

100%

More contact (Non-SVORI)
About the same (SVORI)

80%

About the same (Non-SVORI)
Less contact (Non-SVORI)

30

16

20

17

26

35

36

48

47
32

17

21

20

20%

34

40%

46

46

54

60%

55

Less contact (SVORI)

0%
Phone contact

Mail contact

In-person visits

Note: Differences between SVORI and non-SVORI are not significant at the 0.05 level.

SUMMARY
The SVORI was targeted to serious and violent offenders. The
grantees were allowed to define “serious and violent,” so it was
possible that the programs would “cream” (i.e., select “better”
offenders for the enhanced services to be provided by their
SVORI programs). The criminal histories and circumstances of
their current incarcerations suggest that the adult male
respondents included in the SVORI Multi-site Evaluation fit the
“serious and violent” label. At the time of their pre-release
interview, they were, on average, about 30 years old and had

62

Respondent Characteristics

been in prison more than 2 years. More than 40% were serving
time for a violent/person offense, most commonly assault and
robbery. Other common offenses included drug
dealing/manufacturing and drug possession. Nearly all had a
history of alcohol and drug use and nearly two thirds reported
using drugs in the 30 days prior to their current incarceration.
They reported substantial arrest histories (an average of more
than 12 arrests) and conviction histories (an average of more
than 6). Most also reported having family members and peers
who had crime and substance use involvement.
Consistent with profiles of prisoners found in the literature, the
respondents had low levels of educational attainment and
unstable employment histories. Most had worked at some
point, but only about two thirds reported working in the
6 months prior to their incarceration. Although about 60% said
they had supported themselves with a job, many reported that
they also relied on illegal activities and family for support prior
to their incarceration.
These characteristics suggest that these individuals had high
levels of need that could be addressed with programs and
services to facilitate their transitions back to their communities
at release—a topic that is addressed in the following chapter.

63

Self-reported
Service Needs
SVORI programs were to identify individual service needs and
provide services and programming to respond to those needs.
This chapter focuses on self-reported service needs. In each
interview, respondents were asked about the extent to which
they needed each of 28 specific services. 28 For ease of
presentation and interpretation, the individual services were
grouped into five service categories or “bundles.” The bundles
are

28

ƒ

transition services (10 items: need legal assistance,
financial assistance, public financial assistance, public
health care insurance, mentoring, documents for
employment, place to live, transportation, drivers
license, clothes/food bank);

ƒ

health care services (5 items: need medical treatment,
mental health treatment, substance abuse treatment,
group for abuse victims, anger management services);

ƒ

employment, education, and skills services (6 items:
job, more education, money management skills, life
skills, work on personal relationships, change attitudes
on criminal behavior);

ƒ

domestic violence-related services (2 items: need
batterer intervention program, need domestic violence
support group); and

ƒ

child-related services (5 items: need child support
payments, modifications in child support debt,
modifications in custody, parenting skills, child care).

Response options were “a lot,” “a little,” or “not at all.” Responses
were recoded to “some” and “not at all.”

65

Prisoner reentry experiences of adult males

SVORI service bundle
scores were developed to
summarize service needs
and service receipt by
summing indicators of
needs and receipt in
multiple domain areas.

The service need bundle scores were developed from the
interview data to summarize respondents’ needs in the domains
of transition, health, employment/education/skills, domestic
violence, and child services (which was calculated only for
respondents with children). Scores for each individual were
generated by summing one/zero indicators for whether the
individual reported or did not report needing each of the items
within a bundle; this sum was then divided by the number of
items in the bundle. At the individual respondent level, this
bundle score can be interpreted as the proportion of the
services in the bundle that the individual reported needing
(Winterfield et al., 2006). 29
The pre-release data suggested high levels of expressed need
as can be seen in Exhibit 35. The levels of expressed need for
employment, education, and skills were very high—on average,
respondents reported needing nearly three quarters of all of the
service items in the employment bundle (average bundle scores
of 75 for SVORI and 74 for non-SVORI). Respondents also
expressed a high level of need for the services and assistance
contained in the transition services bundle. On average,
respondents reported needing nearly two thirds of these
services, which include financial assistance, transportation, and
obtaining a driver’s license and other documentation (average
scores of 64 for SVORI and 62 for non-SVORI).

SVORI and non-SVORI
reported high need across
the spectrum of services.

SVORI and non-SVORI respondents were similar on most
measures and reported high need across the spectrum of
services (see Exhibit A-3 in Appendix A). Specifically, as shown
in Exhibit 36, most SVORI respondents reported needing more
education (94%), financial assistance (86%), a driver’s license
(83%), job training (82%), and a job (80%). About three
quarters (75%) also reported needing public health care
insurance and life skills training. Of these services, non-SVORI
respondents were significantly less likely than SVORI
respondents to report needing financial assistance or job
training—perhaps because SVORI participants had an increased
awareness of need as a result of needs assessments conducted
in conjunction with their SVORI program participation.

29

66

Although not presented in this report, program-level bundle scores
of service delivery were also developed from reports provided by
SVORI program directors. These bundle scores are discussed in
Winterfield et al. (2006).

Self-reported Service Needs

Exhibit 35. Pre-release
service need bundle
scores across service
bundles, by group

Employment/education/life skills
services

75
74
64
62

Transitional services
46
48

a
Child servicesa

31
34

Health services*

SVORI

7
8

Domestic violence services
0

Non-SVORI
20

40

60

80

100

a

Among those who reported having minor children.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

Exhibit 36. Most
commonly reported
service needs prerelease, by group

94
92

Education
86
82
83
81

Financial assistance*
Driver’s license

82

Job training*

76
80
76

Job
Public health care insurance

75
73

Life skills

75
73

SVORI
Non-SVORI

0%

25%

50%

75%

100%

*p < 0.05 for test of significant difference between SVORI and non-SVORI.

When asked to identify their top two service needs, more than
one third of respondents mentioned needing a job after release
(38% SVORI, 36% non-SVORI). About one quarter (24%
SVORI, 25% non-SVORI) listed needing a driver’s license as
one of their top two needs. The next four needs mentioned by
the most respondents as one of their top two included more
education (18% of both groups), job training (17% SVORI,
14% non-SVORI), financial assistance (15% SVORI, 16% non-

67

Prisoner reentry experiences of adult males

SVORI), and a place to live when released (15% SVORI, 16%
non-SVORI).

Respondents reported
needing more than half of
all the service items.

In addition to the domain-specific service bundles, an “all
services” bundle was also created to capture the level of overall
need across all services (individual items are in Exhibit A-3 in
Appendix A). 30 On average, the respondents reported needing
more than half of all the service items (average score of 54 for
both SVORI and non-SVORI).
The findings on self-reported service needs are discussed for
each of the five bundles over the four interview waves in the
remainder of this chapter. Needs are first presented and
discussed at the service bundle level. More detailed information
for each bundle and the individual services contained therein
are then presented.
The findings presented in the remainder of this chapter were
generated using SAS procedures that allow use of the
propensity scores to weight the observations (as described
earlier). The mean values were generated using Proc Survey
Means. Tests of significance of differences between SVORI and
non-SVORI were conducted using Proc Survey Logistic or
Regression to regress the SVORI indicator on each of the
service need scores and assessing whether the resulting
parameter estimate was significantly different from zero.

WEIGHTED SERVICE NEED BUNDLE SCORES
ACROSS WAVES 1 THROUGH 4
Four waves of interviews were conducted:
ƒ

Wave 1: 30 days prior to respondent’s release

ƒ

Wave 2: 3 months following respondent’s release

ƒ

Wave 3: 9 months following respondent’s release

ƒ

Wave 4: 15 months following respondent’s release

Respondents were asked about their need for services at each
wave. Exhibit 37 summarizes information on average reported
service needs for each group across the four waves of interview
data. As noted previously, the averages were generated using
30

68

The number of items varied, depending on whether the individual
had children (59% of SVORI and 61% of non-SVORI respondents
reported having children under the age of 18; 62% and 64%
reported having any children).

Self-reported Service Needs

Exhibit 37. Weighted
average super bundle
scores by group for
Waves 1 through 4

100

SVORI

90

Non-SVORI

80
70
60

54

54

50

42

43

43

43

44

45

40
30
20
10
0
Wave 1

Wave 2

Wave 3

Wave 4

Note: Differences between groups were not significant at the 0.05 level. Wave 1
data were collected 30 days prior to release; Wave 2, 3, and 4 data were
collected 3, 9, and 15 months, respectively, post-release.

the propensity-score weights. The “super bundle” scores that
are shown in Exhibit 37 include all of the needs and were
generated by summing across reports of needs for all service
items and dividing by the total number of service items (28
items for those with children and 23 items for those without
children).
For Wave 1, the results suggest that on average about 30 days
before release the subjects reported needing (a little or a lot)
54% of the items. 31

SVORI and non-SVORI
subjects were similar on
the level of reported need.
Overall need dropped
following release.
Average need was similar
at 3, 9, and 15 months
post-release.

Exhibit 37 suggests three findings: (1) SVORI and non-SVORI
subjects were similar on the level of reported need across all
waves; (2) overall need dropped following release; and
(3) there was little difference in average need across the three
follow-up waves (i.e., at 3, 9, and 15 months following
release). As noted, prior to release, the subjects reported
needing about 54% of the various items. Following release,
need dropped to slightly more than 40% of the items.
Exhibit 38 shows the weighted average bundle scores for the
employment/education/skills, transition, and health services
bundles, by group and data collection wave. As can be seen,
the highest levels of expressed need are for employment/
31

It is interesting to note that the weighted and unweighted averages
are the same, reflecting how well balanced (similar) the two groups
were even prior to the development and application of the
propensity score weights.

69

Prisoner reentry experiences of adult males

Exhibit 38. Weighted average service bundle scores by type (Employment/Education/Skills,
Transition Services, Health Services), group (SVORI, non-SVORI), and wave (1, 2, 3, 4)
Employment/education
skills bundle-1
Employment/education
skills bundle-2
Employment/education/
skills bundle-3
Employment/education/
skills bundle-4

74
74
60
60
62
63
64
63
64
62

Transition services bundle-1

49
50
49
49
49
50

Transition services bundle-2
Transition services bundle-3
Transition services bundle-4

32
34

Health services bundle-1

24
25
26
27
28
29

Health services bundle-2
Health services bundle-3
Health services bundle-4

0

10

20

30

SVORI
Non-SVORI

40

50

60

70

80

90

100

Bundle Score

education/skills programming and services. On average, prior
to release, SVORI and non-SVORI respondents reported
needing nearly three quarters (74%) of the seven services
included in this bundle. At the subsequent three time points,
the level of expressed needs remains very high compared to
the other bundles—in particular, at 15 months, respondents are
still reporting needing nearly two thirds (about 63%) of these
services. Finally, these relatively young men (average age of
about 29 at release) expressed lower levels of need for the five
health-related services, reporting on average a need for about
one third of the five types of services included in this bundle
prior to release and about one quarter post-release. For each of
these three bundles, the pattern is similar to that observed in
Exhibit 37 for the super bundle scores—higher levels of need
expressed prior to release than observed at 3 months postrelease, with reports remaining level in subsequent follow-up
periods. It is interesting to note for health care services that
although the differences are small, there appears to be an
upward trend post-release.

70

Self-reported Service Needs

Exhibit 39 presents similar data for the remaining two needs
bundles—child services and domestic violence services. As can
be seen, the levels of expressed need for help with child-related
issues among those with children is fairly high, with
respondents suggesting the need for two or more of the five
services prior to release. Although there is a decline in reported
need post-release, this decline is smaller than for the other
bundles. Few of the men reported needing help with domestic
violence—either through domestic violence support groups or
batterer intervention programs.
Exhibit 39. Weighted
average super bundle
scores by type (Child
Services, Domestic
Violence, group (SVORI,
non-SVORI), and wave
(1, 2, 3, 4)

46
48

Child services-1

Child services-3

39
39
38
41

Child services-4

40
41

Child services-2

7
8

Domestic violence-1
Domestic violence-2

4
4

Domestic violence-3

3
3

Domestic violence-4

4
6

0

SVORI
Non-SVORI

20

40

60

80

100

Bundle Score

The following sections examine the patterns for the specific
items in each of these service bundles.

EMPLOYMENT/EDUCATION/SKILLS
SERVICES
The service area for which the men consistently reported the
greatest need was employment, education, and skills-related
services (E/E/S). On average, as was seen in Exhibit 38, the
men reported needing more than 60% of these six items
throughout and there were no differences between the two
groups.

71

Prisoner reentry experiences of adult males

Exhibit 40 shows the weighted means for the bundle scores and
individual items by group and wave. Included are the weighted
means and the parameter estimate from the Proc Survey
Regression (bundle scores) or Logistic (individual items), the
test statistic (t-statistic for the regression models and Wald chisquare for the logistic regression models), and the associated
p values; estimated odds ratios are also included for the
individual items—indicating the odds of needing a specific item
conditioned on SVORI status. As can be seen, none of the
parameters was significant at the 0.05 level, indicating no
differences in the expressed need for these services between
the SVORI participants and the comparison subjects.
Exhibit 40 also includes entries for the percentage of subjects
reporting at each wave to need ANY of the E/E/S services—
virtually everyone indicated needing help in this domain—99%
at Wave 1, about 95% at Waves 2 and 3, and more than 96%
at Wave 4. Education was the area in which the greatest
proportions of men indicated needing more. More than 90%
said they needed additional education prior to release and
between 80% and 90% stated the need for additional education
post-release. A job followed education as the most frequently
reported need. Majorities also indicated the need for programs
to address money management and other life skills, as well as
personal relationships. Overall, majorities reported a continuing
need for these five services throughout the follow-up period.
Respondents were less likely over time to express the “need to
change attitudes about criminal behavior.” Although about two
thirds of subjects pre-release indicated a need to change their
attitudes, this proportion diminished to 44% at 3 months
following release and remained below 50% throughout the
follow-up period.

TRANSITION SERVICES
As shown in Exhibit 41, the men in the study also reported high
levels of need for transition services, which included 10 forms of
assistance with the reentry process. Self-reported need for these
services was highest at the time of the pre-release interview
(when the weighted bundle scores were 64 for the SVORI group
and 62 for the non-SVORI group) and declined at the 3-month
post-release interview to about 50 (where they remained for the
following interview waves). As noted earlier, there was no

72

Self-reported Service Needs

Exhibit 40. Weighted means for employment/education/skills bundles and items, by group
and wave

Variable

Wave 1
Employ/Ed/Skills Score
Need ANY E/E/S*
Need job
Need education
Need money mgmt skills
Need life skills
Need work personal
relationships
Need change attitudes
criminal behavior
Wave 2
Employ/Ed/Skills Score
Need ANY E/E/S*
Need job
Need education
Need money mgmt skills
Need life skills
Need work personal
relationships
Need change attitudes
criminal behavior
Wave 3
Employ/Ed/Skills Score
Need ANY E/E/S*
Need job
Need education
Need money mgmt skills
Need life skills
Need work personal
relationships
Need change attitudes
criminal behavior
Wave 4
Employ/Ed/Skills Score
Need ANY E/E/S*
Need job
Need education
Need money mgmt skills
Need life skills
Need work personal
relationships
Need change attitudes
criminal behavior

S
N

NonSVORI
N

S
Mean

NonSVORI
Mean

863
863
862
863
863
859
862

834
834
834
834
833
831
832

74.40
0.99
0.79
0.93
0.71
0.75
0.64

73.60
0.99
0.77
0.92
0.68
0.74
0.63

860

833

0.65

529
529
527
529
529
529
529

455
455
455
455
455
455
455

527

Par.
Est.

SE

Test
Stat.

p Val.

OR

0.80
−0.52
−0.16
−0.15
−0.13
−0.04
−0.05

1.25
0.49
0.12
0.19
0.11
0.11
0.10

0.64
1.12
1.76
0.65
1.56
0.12
0.26

0.52
0.29
0.18
0.42
0.21
0.73
0.61

NA
0.60
0.85
0.86
0.88
0.96
0.95

0.69

0.16

0.11

2.41

0.12

1.18

59.73
0.95
0.61
0.85
0.54
0.60
0.54

60.04
0.96
0.63
0.85
0.57
0.58
0.53

−0.31
−0.34
0.06
0.06
0.12
−0.10
−0.04

1.91
0.32
0.13
0.18
0.13
0.13
0.13

−0.16
1.17
0.19
0.13
0.77
0.52
0.12

0.87
0.28
0.67
0.72
0.38
0.47
0.73

NA
0.71
1.06
1.07
1.12
0.91
0.96

454

0.44

0.44

0.00

0.13

0.00

0.99

1.00

564
564
564
564
563
562
563

469
469
469
469
469
468
469

63.17
0.95
0.58
0.87
0.63
0.66
0.58

62.04
0.95
0.62
0.88
0.59
0.64
0.56

1.12
−0.05
0.18
0.05
−0.15
−0.12
−0.08

1.88
0.30
0.13
0.19
0.13
0.13
0.13

0.60
0.02
1.96
0.06
1.41
0.74
0.39

0.55
0.88
0.16
0.80
0.23
0.39
0.53

NA
0.96
1.20
1.05
0.86
0.89
0.92

562

469

0.48

0.44

−0.16

0.13

1.66

0.20

0.85

582
582
582
582
582
581
581

531
531
531
530
530
530
530

63.15
0.96
0.59
0.88
0.62
0.64
0.59

64.16
0.97
0.62
0.89
0.62
0.68
0.56

−1.01
−0.42
0.15
0.10
0.02
0.19
−0.15

1.80
0.34
0.13
0.19
0.13
0.13
0.12

−0.56
1.52
1.48
0.25
0.03
2.26
1.42

0.57
0.22
0.22
0.62
0.85
0.13
0.23

NA
0.66
1.16
1.10
1.02
1.21
0.86

580

531

0.48

0.48

0.02

0.12

0.03

0.87

1.02

Note: NA = Not applicable. E/E/S = employment/education/skills bundle. OR = odds ratio. Par. Est. = parameter
estimate. p Val = probability value. S = SVORI. SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3
months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations.

73

Prisoner reentry experiences of adult males

Exhibit 41. Weighted means for transition services bundles and items, by group and wave

Variable
Wave 1
TransSvcsScore
Need ANY transition
services*
Need legal assistance
Need financial assist
Need public financial assist
Need public health care
insurance
Need mentor
Need docs for employ
Need place to live
Need transportation
Need drivers license
Need cloth/food bank
Wave 2
TransSvcsScore
Need ANY transition
services*
Need legal assistance
Need financial assist
Need public financial assist
Need public health care
insurance
Need mentor
Need docs for employ
Need place to live
Need transportation
Need drivers license
Need cloth/food bank
Wave 3
TransSvcsScore
Need ANY transition
services*
Need legal assistance
Need financial assist
Need public financial assist
Need public health care
insurance
Need mentor
Need docs for employ
Need place to live
Need transportation

74

S
N

NonNonSVORI
S
SVORI
N
Mean Mean

Par.
Est.

SE

Test
Stat.

p Val.

OR

863
863

834
834

63.82
0.99

62.43
0.99

1.39
−0.18

1.21
0.47

1.14
0.16

0.25
0.69

NA
0.83

856
862
861
861

834
834
834
832

0.45
0.86
0.52
0.76

0.48
0.82
0.54
0.73

0.14
−0.36
0.07
−0.14

0.10
0.14
0.10
0.11

1.97
6.89
0.49
1.48

0.16
0.01
0.48
0.22

1.15
0.70
1.07
0.87

862
863
862
862
863
863

833
834
833
834
834
833

0.60
0.55
0.50
0.72
0.82
0.61

0.60
0.56
0.46
0.70
0.81
0.55

0.00
0.03
−0.16
−0.09
−0.05
−0.26

0.10
0.10
0.10
0.11
0.13
0.10

0.00
0.09
2.70
0.69
0.17
6.83

0.99
0.77
0.10
0.40
0.68
0.01

1.00
1.03
0.85
0.91
0.95
0.77

529
529

455
455

48.89
0.94

49.82
0.97

−0.92
−0.56

1.76
0.33

−0.52
2.89

0.60
0.09

NA
0.57

529
529
528
529

455
455
455
455

0.37
0.62
0.38
0.59

0.40
0.61
0.41
0.62

0.12
−0.02
0.12
0.12

0.13
0.13
0.13
0.13

0.77
0.02
0.88
0.86

0.38
0.89
0.35
0.35

1.12
0.98
1.13
1.13

528
529
529
529
529
529

455
454
454
455
455
454

0.44
0.20
0.47
0.70
0.70
0.44

0.39
0.26
0.46
0.70
0.70
0.44

−0.20
0.34
−0.06
0.02
0.01
0.03

0.13
0.16
0.13
0.14
0.14
0.13

2.24
4.76
0.19
0.03
0.00
0.04

0.13
0.03
0.66
0.86
0.96
0.84

0.82
1.40
0.94
1.03
1.01
1.03

564
564

469
469

48.52
0.93

49.19
0.94

−0.67
−0.07

1.80
0.26

−0.37
0.07

0.71
0.78

NA
0.93

563
563
564
564

469
469
469
467

0.43
0.62
0.38
0.58

0.46
0.65
0.37
0.58

0.13
0.16
−0.05
0.02

0.13
0.13
0.13
0.13

1.02
1.41
0.15
0.02

0.31
0.23
0.70
0.88

1.14
1.17
0.95
1.02

564
563
564
564

469
469
469
468

0.48
0.25
0.45
0.62

0.46
0.30
0.45
0.64

−0.09
0.25
−0.02
0.08

0.13
0.14
0.13
0.13

0.50
2.94
0.02
0.40

0.48
0.91
0.09
1.28
0.89
0.98
0.53
1.09
(continued)

Self-reported Service Needs

Exhibit 41. Weighted means for transition services bundles and items, by group and wave
(continued)

Variable
Need drivers license
Need cloth/food bank
Wave 4
TransSvcsScore
Need ANY transition
services*
Need legal assistance
Need financial assist
Need public financial assist
Need public health care
insurance
Need mentor
Need docs for employ
Need place to live
Need transportation
Need drivers license
Need cloth/food bank

S
N
564
564

NonNonS
SVORI
SVORI
N
Mean Mean
468
0.65
0.63
469
0.40
0.38

Par.
Est.
−0.07
−0.07

SE
0.13
0.13

Test
Stat.
0.30
0.27

p Val.
0.58
0.61

OR
0.93
0.94

582
582

531
531

48.89
0.94

50.04
0.92

−1.15
0.32

1.80
0.24

−0.64
1.84

0.52
0.17

NA
1.38

582
582
582
582

528
529
529
528

0.46
0.64
0.36
0.54

0.46
0.64
0.39
0.58

0.03
0.02
0.09
0.17

0.12
0.13
0.13
0.12

0.05
0.02
0.55
1.99

0.82
0.89
0.46
0.16

1.03
1.02
1.10
1.19

582
581
581
582
582
582

530
530
531
531
531
531

0.51
0.31
0.45
0.60
0.64
0.37

0.49
0.34
0.45
0.62
0.65
0.40

−0.10
0.12
0.01
0.06
0.05
0.11

0.12
0.13
0.12
0.13
0.13
0.13

0.71
0.81
0.00
0.22
0.15
0.72

0.40
0.37
0.97
0.64
0.70
0.40

0.90
1.13
1.01
1.06
1.05
1.11

Note: NA = Not applicable. OR = odds ratio. Par. Est. = parameter estimate. p Val. = probability value. S = SVORI.
SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months postrelease; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations.

significant difference in SVORI and non-SVORI subjects with
respect to the transition services bundle scores. Need for at least
one of these services was high prior to release, with 99%
reporting that they needed at least one of the services, and
expressed need remained above 90% throughout the subsequent
interviews.
The most commonly reported needs remained fairly constant
across the waves, although the percentages expressing those
needs diminished some over time. More than half reported
needing all forms of assistance except legal assistance prior to
release, with financial assistance (such as short-term loans or
housing deposits) identified as a need by the largest percentage
and with a driver’s license, transportation, and public health
care identified as needs by more than 70%.

75

Prisoner reentry experiences of adult males

HEALTH SERVICES
Self-reported needs for health services declined slightly
between the pre-release and 3-month post-release interviews
and then increased slowly for both the SVORI and non-SVORI
groups. Exhibit 42 shows the bundle scores for health service
needs at each interview wave. Prior to release, subjects
reported needing about one third of the five health services. At
the time of the 15-month post-release interview, men reported
needing about 28% of the services. The most commonly
reported need was for medical treatment. Unlike items in
previously discussed service bundles, SVORI and non-SVORI
subjects differed on one of the health services items. In
particular, non-SVORI subjects were more likely to report
needing mental health treatment than were SVORI program
participants. Prior to release, 28% of the comparison subjects
reported needing mental health treatment in comparison to
23% of the SVORI program participants. The percentages
expressing need for mental health treatment dropped to 18%
and 23% for SVORI and non-SVORI subjects at 3 months
following release—fewer overall, but still a significant difference
between the two groups. Interestingly, the numbers reporting a
need for mental health treatment increased at 9 and 15 months
following release, such that the proportions reporting a need for
treatment at 15 months was essentially the same as those
observed at the time of the Wave 1 interview.

DOMESTIC VIOLENCE SERVICES
As shown in Exhibit 43, domestic violence services were a lowranked service need among the adult male sample. Only about
10% of the men reported prior to release needing either
domestic violence support groups or batterer intervention
programming. Following release, the percentages dropped to
about 5%. There were no differences between the SVORI and
non-SVORI groups.

76

Self-reported Service Needs

Exhibit 42. Weighted means for health services bundles and items, by group and wave

Variable
Wave 1
Health Services Score
Need ANY health services*
Need medical tx
Need MH tx
Need substance use tx
Need group for abuse
victims
Need anger mgmt program
Wave 2
Health Services Score
Need ANY health services*
Need medical tx
Need MH tx
Need substance use tx
Need group for abuse
victims
Need anger mgmt program
Wave 3
Health Services Score
Need ANY health services*
Need medical tx
Need MH tx
Need substance use tx
Need group for abuse
victims
Need anger mgmt program
Wave 4
Health Services Score
Need ANY health services*
Need medical tx
Need MH tx
Need substance use tx
Need group for abuse
victims
Need anger mgmt program

S
N

NonNonSVORI
S
SVORI
N
Mean Mean

Par.
st.

SE

Test
Stat.

p Val.

OR

863
863
862
861
862
863

834
834
834
832
834
834

31.55
0.80
0.57
0.23
0.38
0.04

33.65
0.80
0.57
0.28
0.42
0.04

−2.10
−0.02
0.01
0.26
0.16
−0.14

1.21
0.12
0.10
0.11
0.10
0.25

−1.73
0.03
0.02
5.36
2.47
0.33

0.08
0.85
0.89
0.02
0.12
0.57

NA
0.98
1.02
1.30
1.17
0.87

861

833

0.36

0.37

0.07

0.10

0.48

0.49

1.07

529
529
529
529
529
529

455
455
455
453
455
455

23.55
0.68
0.53
0.18
0.22
0.03

25.37
0.69
0.56
0.23
0.23
0.02

−1.82
−0.04
0.11
0.33
0.07
−0.08

1.52
0.14
0.13
0.16
0.16
0.41

−1.20
0.07
0.72
3.98
0.22
0.04

0.23
0.80
0.39
0.05
0.64
0.85

NA
0.96
1.12
1.39
1.08
0.93

529

455

0.22

0.22

0.00

0.16

0.00

0.99

1.00

564
564
564
564
563
564

469
469
469
468
469
469

25.57
0.68
0.52
0.19
0.29
0.03

26.56
0.71
0.53
0.24
0.30
0.02

−0.99
−0.17
0.07
0.26
0.04
−0.18

1.55
0.14
0.13
0.16
0.14
0.39

−0.64
1.52
0.31
2.87
0.08
0.21

0.52
0.22
0.58
0.09
0.78
0.65

NA
0.84
1.07
1.30
1.04
0.84

564

469

0.25

0.23

−0.09

0.15

0.33

0.57

0.92

582
582
582
580
582
582

530
530
530
530
530
529

27.86
0.71
0.54
0.23
0.32
0.03

28.65
0.72
0.56
0.27
0.32
0.02

−0.79
−0.06
0.09
0.23
0.02
−0.35

1.53
0.14
0.12
0.14
0.13
0.40

−0.52
0.18
0.48
2.59
0.03
0.78

0.60
0.67
0.49
0.11
0.86
0.38

NA
0.94
1.09
1.26
1.02
0.70

582

530

0.28

0.26

−0.11

0.14

0.64

0.42

0.90

Note: NA = Not applicable. MH = mental health. OR = odds ratio. Par. Est. = parameter estimate. p Val. = probability
value. S = SVORI. SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 =
9 months post-release; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations.

77

Prisoner reentry experiences of adult males

Exhibit 43. Weighted means for domestic violence services bundles and items, by group and
wave

Variable
Wave 1
DomViolScore
Need ANY DV services*
Need batterer intervention
program
Need DV support group
Wave 2
DomViolScore
Need ANY DV services*
Need batterer intervention
program
Need DV support group
Wave 3
DomViolScore
Need ANY DV services*
Need batterer intervention
program
Need DV support group
Wave 4
DomViolScore
Need ANY DV services*
Need batterer intervention
program
Need DV support group

S
N

NonSVORI
N

NonS
SVORI
Mean Mean

862
862
861

834
834
833

6.82
0.09
0.08

862

833

529
529
527

Par.
Est.

SE

Test
Stat.

p Val.

OR

8.17
0.11
0.08

−1.35
−0.20
0.00

1.15
0.16
0.18

−1.17
1.50
0.00

0.24
0.22
0.99

NA
0.82
1.00

0.06

0.09

0.40

0.19

4.42

0.04

1.49

455
455
455

4.35
0.06
0.05

4.37
0.06
0.06

−0.03
−0.05
0.09

1.18
0.27
0.28

−0.02
0.03
0.10

0.98
0.87
0.75

NA
0.96
1.10

529

455

0.04

0.03

−0.14

0.36

0.15

0.70

0.87

564
564
564

469
469
468

3.06
0.04
0.04

3.47
0.05
0.04

−0.41
−0.24
0.19

1.01
0.30
0.33

−0.41
0.62
0.34

0.68
0.43
0.56

NA
0.79
1.21

564

468

0.03

0.03

0.04

0.40

0.01

0.91

1.04

582
582
582

529
529
529

3.81
0.05
0.04

5.76
0.07
0.06

−1.95
−0.43
0.37

1.22
0.26
0.27

−1.60
2.69
1.85

0.11
0.10
0.17

NA
0.65
1.45

582

529

0.03

0.05

0.51

0.31

2.73

0.10

1.67

Note: NA = Not applicable. DV = domestic violence. OR = odds ratio. Par. Est. = parameter estimate. p Val. =
probability value. S = SVORI. SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3 months postrelease; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations

CHILD SERVICES
More than 60% of the men reported having children (62% of
SVORI and 64% of non-SVORI respondents), and nearly all of
those with children reported having children under the age of
18 years (96% SVORI and 95% non-SVORI). Among those who
had minor children, the expressed need for child-related
services (which included instruction in parenting skills, child
care, child support payments, modifications in the child support
debt they owed, and modifications in the custody of their
children) decreased somewhat over the follow-up period. As
shown in Exhibit 44, the men reported needing about half of

78

Self-reported Service Needs

Exhibit 44. Weighted means for child services bundles and items, by group and wave

Variable
Wave 1
ChildSvcsScore
Need ANY child services*
Need child support
payments
Need child support debt
modified
Need custody modified
Need parenting skills
Need child care
Wave 2
ChildSvcsScore
Need ANY child services*
Need child support
payments
Need child support debt
modified
Need custody modified
Need parenting skills
Need child care
Wave 3
ChildSvcsScore
Need ANY child services*
Need child support
payments
Need child support debt
modified
Need custody modified
Need parenting skills
Need child care
Wave 4
ChildSvcsScore
Need ANY child services*
Need child support
payments
Need child support debt
modified
Need custody modified
Need parenting skills
Need child care

S
N

NonSVORI
N

NonS
SVORI
Mean Mean

508
508
500

503
503
495

46.07
0.83
0.46

137

139

502
507
507

Par.
Est.

SE

Test
Stat.

p Val.

OR

47.82
0.85
0.47

−1.75
−0.10
0.07

1.99
0.18
0.13

−0.88
0.31
0.28

0.38
0.58
0.59

NA
0.91
1.07

0.87

0.86

−0.10

0.36

0.07

0.79

0.91

500
502
500

0.35
0.60
0.39

0.37
0.63
0.39

0.11
0.13
0.02

0.13
0.13
0.13

0.70
0.94
0.03

0.40
0.33
0.87

1.12
1.14
1.02

314
314
314

273
273
272

38.82
0.77
0.34

38.99
0.78
0.36

−0.16
−0.03
0.12

2.46
0.20
0.18

−0.07
0.02
0.45

0.95
0.88
0.50

NA
0.97
1.13

99

95

0.87

0.86

−0.04

0.44

0.01

0.92

0.96

312
314
314

272
273
271

0.30
0.53
0.29

0.31
0.49
0.31

0.02
−0.13
0.06

0.18
0.17
0.18

0.01
0.63
0.12

0.92
0.43
0.73

1.02
0.88
1.07

333
333
332

274
274
274

37.85
0.75
0.33

41.15
0.80
0.40

−3.30
−0.28
0.29

2.52
0.20
0.17

−1.31
1.86
2.78

0.19
0.17
0.10

NA
0.76
1.33

109

96

0.79

0.85

0.43

0.38

1.28

0.26

1.53

333
333
332

274
274
274

0.31
0.53
0.28

0.34
0.52
0.30

0.17
−0.06
0.11

0.18
0.17
0.18

0.88
0.13
0.38

0.35
0.72
0.54

1.18
0.94
1.12

354
354
353

342
342
339

40.21
0.80
0.35

41.44
0.81
0.39

−1.24
−0.04
0.17

2.28
0.19
0.16

−0.54
0.05
1.10

0.59
0.82
0.29

NA
0.96
1.18

111

121

0.89

0.90

0.10

0.43

0.06

0.81

1.11

353
354
353

341
342
342

0.31
0.59
0.27

0.33
0.53
0.31

0.06
−0.25
0.23

0.16
0.16
0.17

0.14
2.56
1.87

0.71
0.11
0.17

1.06
0.78
1.26

Note: NA = Not applicable. OR = odds ratio. Par. Est. = parameter estimate. p Val. = probability value. S = SVORI.
SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months postrelease; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations

79

Prisoner reentry experiences of adult males

the child-related services at the time of the pre-release
interview, but only about 40% at the follow-up interviews. The
greatest need across all waves was for modification to child
support debt—a need expressed by most of those with children.
However, a majority also indicated that they needed services to
improve their parenting skills.

LEVELS OF NEED ACROSS SERVICES
Overall, the men reported high needs in the days immediately
before release and somewhat lower needs at each of the followup interviews. Among the various service bundles, services
related to education, employment, and skill building
consistently were reported the most frequently. In terms of
individual services, the ones most consistently identified as a
need by the largest proportion of men were more education,
employment, public health care insurance, and financial
assistance.

The most commonly
reported service needs in
the top two across all
waves were a job (24–
38%) and more education
(18–22%).

At each interview wave, in addition to self-reporting whether a
particular service was needed, respondents were also asked to
report their “top two” service needs. According to this measure,
two employment/education/skills services ranked among the
top five most commonly reported. Specifically, among the
services that were most commonly reported in the top two
across all waves were a job (24–38% of the men) and more
education (18–22%). Other services reported were a driver’s
license (19–28%), financial assistance (11–16%), and a place
to live (12–15%).
Importantly, very few differences in self-reported needs were
observed between the SVORI program participants and the
non-SVORI comparison subjects at any interview wave. Given
that SVORI programming was supposed to identify and provide
services to address individual needs, these findings at first
glance might be considered disappointing and might seems to
suggest either that needs weren’t addressed or that the
services were ineffective. An alternative and plausible
explanation, however, is that many of these needs are
ongoing—regardless of whether they are addressed or not.
Even if someone receives education or medical treatment, the
need for these two types of services can persist—more
education will be helpful, and treatment for a new illness or
injury may be needed. Additionally, former prisoners may

80

Self-reported Service Needs

struggle with low-paying jobs that threaten on an ongoing basis
their ability to attain and retain housing, meet child support
obligations, etc.
Findings presented in the following section show that SVORI
programs were, in fact, successful in increasing the numbers
and types of services provided to program participants.

81

Self-reported
Service Receipt
In addition to collecting extensive information on subjects’
needs, the Multi-site Evaluation of SVORI collected analogous
information on service receipt to determine whether SVORI
programs were successful in increasing participants’ access to
services. In the report Pre-release Characteristics and Service
Receipt among Adult Male Participants in the SVORI Multi-site
Evaluation (Lattimore, Visher, & Steffey, 2008), extensive
findings were presented that suggest that the SVORI programs
substantially increased access to pre-release services for
program participants. For all service areas, men enrolled in
SVORI programs reported significantly higher service receipt
during their periods of incarceration than comparable men not
enrolled in SVORI.
Service receipt bundle scores were calculated analogous to the
calculations of the service need bundle scores: the number of
“yes” responses to items in a bundle was divided by the
number of bundle items and multiplied by 100. Individual
bundle scores were averaged to get site-level scores, which
were averaged to get overall scores. Child services receipt
bundle scores were generated only for those respondents who
reported having children under the age of 18.
In addition to the five service need bundles, a sixth service
receipt bundle was developed consisting of coordination items.
The service receipt bundles measured receipt of specific items
since the last interview 32 and are as follows:
32

Individuals who did not complete the previous interview were asked
about the comparable time period for the 9- and 15-month
interviews (i.e., in the past 6 months rather than since the last
interview).

83

Prisoner reentry experiences of adult males

SVORI programs were
successful in increasing
access to a wide range of
services and
programming.

coordination services (4 items: received a needs
assessment, met with a case manager, worked with
anyone to reintegrate; currently on probation or parole;
Wave 1 did not include currently on probation or parole,
but did include two additional items: received needs
assessment specific for release and reentry plan
developed);

ƒ

transition services (10 items: received legal assistance,
financial assistance, public financial assistance,
assistance with public health care insurance, mentoring,
documents for employment, place to live,
transportation, drivers license, clothes/food bank);

ƒ

health care services (6 items: received any medical
treatment, dental services, mental health treatment,
substance abuse treatment, group for abuse victims,
anger management services);

ƒ

employment, education, and skills services (6 items:
received any employment services, any educational
services, money management assistance, life skills
training, assistance with personal relationships, training
to change attitudes on criminal behavior);

ƒ

domestic violence-related services (2 items: participated
in batterer intervention program, domestic violence
support group); and

ƒ

child-related services (5 items: received assistance
getting child support payments, modifying child support
debt, modifying custody, parenting skills classes, finding
child care).

Exhibit 45 shows the pre-release service receipt bundle scores
for all SVORI and non-SVORI respondents and clearly
demonstrates that SVORI programs were successful in
increasing access to a wide range of services and
programming. 33 In all cases, SVORI respondents were
significantly more likely than non-SVORI respondents to report
receiving more services across the six bundles. However, the
exhibit also shows that, with the exception of the coordination
services bundle that includes assessments and reentry
planning, SVORI respondents reported receiving 40% or less of
the items in the service receipt bundles. Further, the levels of
service receipt are substantially lower than the levels of need
33

84

ƒ

Note that the items included in the health services and
employment/education/skills bundles differ slightly from those that
were included in the bundles reported in Lattimore, Visher, and
Steffey (2008).

Self-reported Service Receipt

Exhibit 45. Pre-release
service receipt bundle
scores across service
bundles, by group

60

Coordination services

33
38

Employment/education/skills
services

24
36
31

Health services

29

Transitional services

17
11

Child services*

6

SVORI

8

Domestic violence services

Non-SVORI

4
0

20

40

60

80

100

Note: DV = domestic violence. All differences between SVORI and non-SVORI
are significant at p < 0.05.
*Among those who reported having minor children.

reported by the respondents and shown earlier in Exhibit 35.
Expressed need was highest pre-release for the services
included in the employment/education/skills and the transition
services bundles. The average employment/ education/skills
service need bundle score was 75 for SVORI and 74 for nonSVORI respondents—substantially higher than the 38 and 24
service receipt scores shown in Exhibit 45. Specifically, the
need score for SVORI respondents was about twice the receipt
score, suggesting that services were provided for only about
one half of needs—substantially better, however, than for nonSVORI respondents whose need score was about three times
the receipt score, suggesting that they received services in
response to only one third of their needs.
Exhibit A-4 in Appendix A shows the proportion of respondents
who reported receiving each item in these six bundles.
This chapter examines whether self-reported service receipt
remained higher for the SVORI participants across each
interview wave (with the weighted service receipt scores shown
for the pre-release time period as a reference point). Weighted
means and logistic parameter estimates to assess the
significance of SVORI program participation on service receipt
are reported.

85

Prisoner reentry experiences of adult males

WEIGHTED SERVICE NEED BUNDLE SCORES
ACROSS WAVES 1 THROUGH 4
Exhibit 46 summarizes information on average reported service
receipt for each group across the four waves of interview data.
As noted previously, the averages were generated using the
propensity-score weights. The “super bundle” scores that are
shown in Exhibit 46 reflect all of the services and were
generated by summing across reports of receipt for all service
items and dividing by the total number of service items.
Exhibit 46. Weighted
average service receipt
super bundle scores by
group for Waves 1
through 4

100

SVORI

90

Non-SVORI

80
70
60
50
40
30

34
22

20

18

12

10

13

10

9

8

0
Wave 1

Wave 2

Wave 3

Wave 4

Note: Differences between groups at Waves 1, 2, and 3 were significant at the
0.0001 level.

SVORI program
participants received
substantially more services
pre-release.
Service receipt dropped
following release.
Service receipt continued to
decline as the time since
release increased.
By 15 months following
release, there was little
difference in receipt with
both groups reporting
receipt of less than 10% of
service items.

86

For Wave 1, the results suggest that on average about 30 days
before release the SVORI subjects reported receiving about a
third (34%) of the various service items and the non-SVORI
comparison subjects reported receiving about one quarter
(24%) of the items. (This difference and the differences at
Waves 2 and 3 are significant at the p < .0001 level.)
Exhibit 46 suggests three findings: (1) SVORI program
participants received substantially more services pre-release
than non-SVORI subjects; (2) service receipt dropped
substantially at release (at 3 months post-release the average
subject reported receiving roughly half as many services as 30
days prior to release); (3) reported service receipt continued to
drop as the time since release lengthened; and (4) there was
little difference in average receipt at 15 months following
release—with both groups reporting receiving less than 10% of
the service items.

Self-reported Service Receipt

Recall that the reported needs also dropped between the prerelease interview (Wave 1) and the 3-month post-release
interview (Wave 2), but the decline was not as dramatic as that
in Exhibit 46. Exhibit 47 combines the data from Exhibits 37
and 46. As can be seen, need exceeds receipt at all waves and
the decline in receipt is steeper than the decline in reported
need.
Exhibit 47. Weighted
average service need
and receipt super
bundle scores by group
for Waves 1 through 4

SVORI-Need

100

non-SVORI-Need

90

Svori-Receipt

80

non-SVORI-Receipt

70
60

54 54

50

42 43

43 43

44 45

40
30
20
10
0
Wave 1

Wave 2

Wave 3

Wave 4

Note: S = SVORI. Differences between groups for service receipt at Waves 1, 2,
and 3 were significant at the 0.0001 level; no other difference is significant at
the 0.05 level. Wave 1 = 30 days pre-release; Wave 2 = 3 months postrelease; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.

Exhibit 48 shows the weighted average bundle scores for the
coordination, employment/education/skills, and transition
services bundles, by group and data collection wave. As can be
seen, the average proportion of services being received
post-release drops rapidly for the employment/education/skills
and the transition services. The coordination services bundle
scores deviate from the pattern of substantial drops between
30 days prior to release (Wave 1) and 3 months following
release (Wave 2) that were observed for the super bundle
scores. One explanation is that for this one bundle the items
contained in the Wave 1 and Wave 2 bundles differ somewhat
(i.e., Wave 1 includes received needs assessment specific for
release and reentry plan developed; Wave 2–4 does not include
these items but includes probation/parole status). But, more
likely the explanation is that most of the study participants
were released to supervision and thus received a minimum of
one of the four coordination services included in the post-

87

Prisoner reentry experiences of adult males

Exhibit 48. Weighted average service receipt bundle scores by type, group, and wave

Coordination bundle-1

60

33

Coordination bundle-2

57

40

Coordination bundle-3

43

31
30

Coordination bundle-4

24

Employment/education/
skills bundle-1

38

24

Employment/education/
skills bundle-2

14

8

Employment/education/
skills bundle-3
Employment/education/
skills bundle-4

6

8
8

11

Transition services bundle-1

17

Transition services bundle-2
Transition services bundle-3

7
6
6

Transition services bundle-4

0

9
9

10

28

13

SVORI
Non-SVORI

20

30

40

50

60

70

80

90

100

Bundle Score

Note: CoS = coordination bundle. E/E/S = employment/education/skills bundle. TS = transition services bundle. All
differences between SVORI and non-SVORI are significant at the .01 level except for the TS-4, where there is no
significant difference between the two groups.

release bundles. The contributions of individual items to the
coordination bundle are explored in more detail in the following
section.
Exhibit 48 clearly shows the dramatic drop in receipt of
employment/education/skills and transition services following
release. The decline occurred for both SVORI and non-SVORI
subjects and continued over time such that by 15 months postrelease there was no difference between SVORI and non-SVORI
in terms of the average number of transition services received
(6%). Since there are 10 items in this bundle, the findings
suggest that at 15 months following release, these men were
receiving very little assistance—less than one of the 10
different services.
Exhibit 49 shows the weighted average bundle scores for the
remaining three bundles—health services, child services and
domestic violence services. As can be seen, there was a
substantial decline between pre-release and post-release

88

Self-reported Service Receipt

Exhibit 49. Weighted average service receipt bundle scores by type, group, and wave

Health services bundle-1

31

35

13
11
13
13
11
12

Health services bundle-2
Health services bundle-3
Health services bundle-4
Child services bundle-1

34

22

Child services bundle-2
Child services bundle-3
Child services bundle-4
Domestic violence services bundle-1
1
1
1
1
1
2

Domestic violence services bundle-2
Domestic violence services bundle-3
Domestic violence services bundle-4
0

5

12
13
10
9
8
7

18

SVORI
Non-SVORI

10

20

30

40

50

60

70

80

90

100

Bundle Score
Note: HS = Health services bundle. CS = Child services bundle. DV = Domestic violence services bundle. Wave 1
differences between SVORI and non-SVORI were significant at the 0.05 level; difference for CS-2 was also
significant at the 0.05 level; other differences were not statistically significant at the 0.05 level.

values. Virtually no one reported receiving domestic violence
treatment or batterer intervention post-release—a finding that
is consistent with very low reports of need.
The following sections examine the patterns for the specific
items in each of these service bundles, presenting the results of
the weighted means and regression analyses for each service
bundle (and individual item) across waves.

COORDINATION SERVICES
The coordination services bundle measures the receipt of
services associated with developing a treatment plan, preparing
for release, and monitoring the offender’s status. Because the
various dimensions of service coordination differed for
incarcerated and released individuals, the coordination services
bundle was scored differently for post-release measures.
Specifically, the pre-release coordination services bundle (as
described previously) included (1) whether a needs assessment

89

Prisoner reentry experiences of adult males

had been conducted, (2) whether a needs assessment
specifically designed to help the individual prepare for release
had been conducted, (3) whether any case management had
been provided, (4) whether a reentry plan had been developed,
and (5) whether the inmate had worked with anyone to help
plan for release. The post-release coordination services
included (1) whether a needs assessment had been conducted,
(2) whether any case management had been provided,
(3) whether the individual had worked with anyone to help
reintegrate him or her back into the community, and
(4) whether the individual was currently on post-release
supervision. Although the bundle scores were standardized on
the basis of the number of items in the score, comparing the
bundle score values from Wave 1 with those from the
subsequent waves was nonetheless difficult since they reflect
different services.

SVORI program
participants were more
likely to report receiving
coordination services
than non-SVORI
comparisons at all four
interviews

Although more likely to
report having a reentry
plan, only 57% of the
SVORI program
participants reported that
they had a reentry plan
about 30 days prior to
their release.

90

Exhibit 50 shows the weighted means for the SVORI and nonSVORI groups, as well as the parameter estimate, standard
error, test statistic, p value, and odds ratio for the coordination
services bundles for each of the four waves. The men reported
higher levels of coordination services than the other service
areas considered in the evaluation, and this pattern was true at
all time periods. Further, higher proportions of the SVORI group
reported receipt of every service item for every wave, but there
was no significant difference in the proportions of the two
groups who reported being on probation/parole.
Focusing first on the Wave 1 data, SVORI program participants
were more likely to report receiving services related to reentry
while they were incarcerated. Specifically, 49% of SVORI
program participants, compared to 23% of the non-SVORI
comparisons, reported receiving a needs assessment
specifically related to their pending release and 66% of SVORI,
compared to only 31% of the non-SVORI comparison subjects,
reported working with anyone to plan for release. A somewhat
surprising finding concerns the proportion of subjects who
reported that they had a reentry plan at the time of the
interview—only 57% of the SVORI program participants
(compared with 25% of the non-SVORI subjects) reported
having a reentry plan.

Self-reported Service Receipt

Exhibit 50. Weighted means for coordination services bundles and items, by group and
wave

Variable
Wave 1
Coordination Score
Rcvd ANY coordination
services*
Rcvd needs assessment
Met with case manager
Rcvd needs assessment
specific for release
Reentry plan developed
Worked with anyone to plan
for release
Wave 2
Coordination Score
Rcvd ANY coordination
services*
Rcvd needs assessment
Met with case manager
Worked with anyone to
reintegrate
Currently on probation or
parole
Wave 3
Coordination Score
Rcvd ANY coordination
services*
Rcvd needs assessment
Met with case manager
Worked with anyone to
reintegrate
Currently on probation or
parole
Wave 4
Coordination Score
Rcvd ANY coordination
services*
Rcvd needs assessment
Met with case manager
Worked with anyone to
reintegrate
Currently on probation or
parole

S
N

NonNonSVORI
S
SVORI
N
Mean Mean

Par.
Est.

SE

Test
Stat.

p Val

OR

863
863

834
834

59.52
0.88

33.00
0.66

26.52
1.27

1.66
0.13

15.96
97.57

0.00
0.00

NA
3.57

860
860
852

830
834
826

0.63
0.66
0.49

0.46
0.41
0.23

0.69
1.06
1.14

0.10
0.10
0.11

47.03
107.24
108.66

0.00
0.00
0.00

2.00
2.89
3.12

843
861

820
834

0.57
0.66

0.25
0.31

1.38
1.42

0.11
0.11

160.58
181.59

0.00
0.00

4.00
4.14

529
529

455
455

57.42
0.92

39.56
0.89

17.87
0.34

1.90
0.22

9.40
2.35

0.00
0.00

NA
1.40

528
527
529

455
455
454

0.44
0.58
0.46

0.18
0.34
0.22

1.31
1.02
1.08

0.15
0.14
0.15

72.11
56.71
54.95

0.00
0.00
0.00

3.69
2.77
2.94

528

454

0.82

0.85

−0.22

0.17

1.54

0.21

0.81

565
565

469
470

43.17
0.81

30.98
0.72

12.19
0.54

1.97
0.15

6.19
12.69

0.00
0.00

NA
1.72

537
539
540

443
441
443

0.29
0.42
0.36

0.18
0.24
0.20

0.59
0.85
0.78

0.16
0.14
0.15

13.65
34.93
26.95

0.00
0.00
0.00

1.80
2.35
2.19

564

468

0.70

0.65

0.25

0.14

3.25

0.07

1.28

580
582

529
531

29.76
0.63

24.41
0.58

5.35
0.20

1.79
0.13

2.99
2.57

0.00
0.11

NA
1.22

476
476
476

443
444
446

0.24
0.31
0.27

0.15
0.22
0.16

0.55
0.43
0.64

0.17
0.15
0.17

9.83
7.79
14.58

0.00
0.01
0.00

1.73
1.54
1.89

579

529

0.53

0.53

0.00

0.12

0.00

1.00

1.00

Note: NA = Not applicable. OR = odds ratio. Par. Est. = parameter estimate. p Val. = probability value. S = SVORI.
SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months postrelease; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations

91

Prisoner reentry experiences of adult males

The Wave 2, 3, and 4 data indicate that most subjects reported
being on probation or parole at the time of the interview and
that SVORI and non-SVORI were equally likely to report being
on parole. The proportion on parole did decline steadily over
time, dropping from more than 80% at 3 months post-release,
to 65% to 70% at 9 months, and to 53% at 15 months
following release.

EMPLOYMENT/EDUCATION/SKILLS
SERVICES
The service area for which the men consistently reported the
greatest need was employment, education, and skills-related
services. On average, as shown in Exhibit 38, the men reported
needing more than 60% of the six items throughout, and there
were no differences between the two groups. Exhibit 51 shows
the weighted means for the service receipt bundle scores and
individual items by group and wave. Greater proportions of
SVORI program participants than non-SVORI comparison
subjects reported receiving most items. Pre-release, 79% of
SVORI and 69% of non-SVORI comparison subjects reported
that they had received at least one of the six services that
included any employment service, any education, money
management, life skills, relationship building, and training to
change their attitudes about criminal behavior. These
percentages dropped to 50% for SVORI and 31% for nonSVORI subjects 3 months following release and declined at each
successive interview until only 26% of SVORI participants and
19% of non-SVORI subjects reported receiving any of these
services at 15 months following release.

The most commonly
reported services
received pre-release were
education and training to
change criminal behavior
attitudes.
Employment-related
services were the most
commonly reported postrelease.

92

The most commonly reported services pre-release were
education (51% SVORI, 44% non-SVORI) and training to
change criminal behavior attitudes (51% SVORI, 36% nonSVORI). Following release, few reported receiving either
education or training to change criminal attitudes—only about
10% reported receiving any education following release,
substantially less than the 85% or so who indicated that they
needed more education. SVORI participants were more likely
than non-SVORI respondents to report receiving training to
change criminal attitudes at 3 and 9 months following release,
but the percentages were small (18% and 16% at 3 and 9
months for SVORI compared to 10% for non-SVORI).

Self-reported Service Receipt

Exhibit 51. Weighted means for employment/education/skills service receipt bundles and
items, by group and wave

Variable
Wave 1
Employ/Ed/Skills Score
Received ANY E/E/S *
Rcvd any employment
Rcvd any education
Rcvd money mgmt training
Rcvd life skills
Rcvd help with personal
relationships
Rcvd training to change
attitudes criminal behavior
Wave 2
Employ/Ed/Skills Score
Received ANY E/E/S*
Rcvd any employment
Rcvd any education
Rcvd money mgmt training
Rcvd life skills
Rcvd help with personal
relationships
Rcvd training to change
attitudes criminal behavior
Wave 3
Employ/Ed/Skills Score
Received ANY E/E/S*
Rcvd any employment
Rcvd any education
Rcvd money mgmt training
Rcvd life skills
Rcvd help with personal
relationships
Rcvd training to change
attitudes criminal behavior
Wave 4
Employ/Ed/Skills Score
Received ANY E/E/S*
Rcvd any employment
Rcvd any education
Rcvd money mgmt training
Rcvd life skills
Rcvd help with personal
relationships
Rcvd training to change
attitudes criminal behavior

S
N

NonSVORI
N

863
863
863
863
862
861
863

834
834
833
834
834
832
834

38.16
0.79
0.37
0.51
0.23
0.42
0.25

863

834

529
529
528
529
529
529
529

S
Mean

NonSVORI
Mean

Par.
Est.

SE

Test
Stat.

p Val.

OR

24.33
0.69
0.20
0.44
0.08
0.21
0.17

13.83
0.51
0.87
0.29
1.28
0.97
0.48

1.39
0.11
0.12
0.10
0.15
0.11
0.12

9.92
20.23
56.72
8.79
68.33
76.28
15.09

0.00
0.00
0.00
0.00
0.00
0.00
0.00

NA
1.67
2.38
1.34
3.59
2.65
1.62

0.51

0.36

0.60

0.10

35.54

0.00

1.83

455
455
455
455
455
455
455

14.40
0.50
0.34
0.11
0.05
0.11
0.07

8.25
0.31
0.20
0.08
0.02
0.05
0.04

6.15
0.79
0.76
0.44
0.75
0.76
0.58

1.09
0.14
0.15
0.23
0.37
0.25
0.28

5.66
34.03
25.11
3.66
4.17
8.96
4.24

0.00
0.00
0.00
0.06
0.04
0.00
0.04

NA
2.21
2.14
1.55
2.12
2.14
1.79

529

455

0.18

0.10

0.64

0.19

10.95

0.00

1.89

540
565
540
540
540
540
540

444
470
444
444
444
443
444

11.48
0.38
0.21
0.11
0.04
0.08
0.09

7.75
0.28
0.14
0.10
0.02
0.05
0.06

3.74
0.45
0.49
0.12
0.80
0.66
0.44

1.05
0.14
0.18
0.21
0.42
0.28
0.26

3.56
10.72
7.57
0.32
3.61
5.59
2.87

0.00
0.00
0.01
0.57
0.06
0.02
0.09

NA
1.56
1.63
1.13
2.23
1.94
1.55

540

444

0.16

0.10

0.48

0.20

5.87

0.02

1.62

476
582
476
476
476
476
476

445
531
445
445
445
445
445

8.47
0.26
0.14
0.10
0.03
0.06
0.07

5.88
0.19
0.10
0.07
0.02
0.03
0.06

2.59
0.40
0.44
0.41
0.60
0.72
0.25

0.93
0.15
0.22
0.25
0.47
0.34
0.27

2.80
7.17
4.19
2.73
1.67
4.40
0.83

0.01
0.01
0.04
0.10
0.20
0.04
0.36

NA
1.49
1.55
1.51
1.83
2.06
1.28

476

445

0.10

0.08

0.24

0.24

1.01

0.31

1.27

Note: NA = Not applicable. E/E/S = employment/education/skills bundle. OR = odds ratio. Par. Est. = parameter
estimate. p Val. = probability value. S = SVORI. SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3
months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations

93

Prisoner reentry experiences of adult males

Employment-related services (e.g., help with resume, job
referrals) were the most commonly reported at each postrelease interview; however, the overall percentage of
individuals reporting receiving employment-related services
was small. Only 34% of SVORI program participants reported
receiving any employment-related services 3 months following
release—significantly higher than the 20% of non-SVORI
respondents who reported employment services, but
substantially below 100%, which one might have expected for
programs with an employment focus.

TRANSITION SERVICES
As was shown in Exhibit 41, the men in the study reported high
levels of need for transition services, which included various
forms of assistance with the reentry process. Nearly all
reported needing at least 1 of the 10 services at each interview,
and a majority reported in the pre-release interview that they
needed help with all of the items except legal assistance.
Among the top needs identified were large services (financial
assistance, public health care insurance) and small (drivers
license, documents for employment). A need for housing was
identified by about half of the respondents pre-release; this
need persisted as about 45% continued to identify housing as a
need in each of the three post-release interviews. Finally, more
than half prior to release and about half following release said
that they needed a mentor, a service that has been linked to
transforming criminal behavior.

Most SVORI program
participants reported
participating in programs
(75%) or classes (65%) to
prepare for release.

94

Exhibit 52 shows the reported receipt of transition services by
group. SVORI program participants were more likely than nonSVORI respondents to report receiving transition services and
to report receiving more transition services from pre-release
through the 9-month post-release interview (Wave 3). The
likelihood and amount declined following release, and the
differences between the two groups diminished over time until
there was no difference at 15 months following release. Most
SVORI participants reported participating in programs (75%) or
classes (65%) to prepare for release, significantly more than
non-SVORI respondents (52% reported participating in
programs, 38% reported participating in classes).

Self-reported Service Receipt

Exhibit 52. Weighted means for transition services receipt bundles and items, by group and
wave

Variable
Wave 1
TransSvcsScore
Rcvd ANY transition services*
Rcvd programs for release
Rcvd classes for release
Rcvd legal assistance
Rcvd financial assist
Rcvd public financial assist
Rcvd public health care
insurance
Rcvd mentor
Rcvd docs for employ
Rcvd place to live
Rcvd transportation
Rcvd drivers license
Rcvd cloth/food bank
Wave 2
TransSvcsScore
Rcvd ANY transition services*
Rcvd legal assistance
Rcvd financial assist
Rcvd public financial assist
Rcvd public health care
insurance
Rcvd mentor
Rcvd docs for employ
Rcvd place to live
Rcvd transportation
Rcvd drivers license
Rcvd cloth/food bank
Wave 3
TransSvcsScore
Rcvd ANY transition services*
Rcvd legal assistance
Rcvd financial assist
Rcvd public financial assist
Rcvd public health care
insurance
Rcvd mentor
Rcvd docs for employ
Rcvd place to live
Rcvd transportation
Rcvd drivers license
Rcvd cloth/food bank

S
N

nonNonSVORI
S
SVORI
N
Mean Mean

Par.
Est.

SE

Test
Stat.

p Val.

OR

863
863
863
862
863
863
862
861

834
834
833
833
834
834
834
834

28.40
0.87
0.75
0.65
0.12
0.13
0.14
0.13

16.77
0.73
0.52
0.38
0.08
0.04
0.11
0.09

11.63
0.89
1.01
1.11
0.39
1.25
0.27
0.42

0.95
0.13
0.11
0.10
0.17
0.21
0.15
0.16

12.20
47.80
90.62
115.94
5.28
35.50
3.20
7.07

0.00
0.00
0.00
0.00
0.02
0.00
0.07
0.01

NA
2.44
2.76
3.03
1.47
3.48
1.31
1.53

863
859
863
862
862
862

834
834
834
834
834
834

0.19
0.41
0.28
0.19
0.22
0.21

0.08
0.26
0.13
0.12
0.08
0.11

0.93
0.66
0.94
0.56
1.15
0.74

0.16
0.11
0.13
0.14
0.16
0.14

36.11
37.59
51.53
16.06
54.76
28.42

0.00
0.00
0.00
0.00
0.00
0.00

2.54
1.93
2.55
1.76
3.17
2.10

529
529
529
529
529
528

455
455
455
455
455
455

13.43
0.59
0.04
0.07
0.20
0.11

8.87
0.43
0.03
0.04
0.16
0.08

4.56
0.66
0.24
0.58
0.29
0.26

0.95
0.13
0.36
0.31
0.17
0.22

4.80
25.31
0.47
3.55
2.90
1.35

0.00
0.00
0.49
0.06
0.09
0.25

NA
1.94
1.28
1.79
1.34
1.30

528
529
529
529
529
528

455
455
455
455
455
455

0.14
0.25
0.12
0.15
0.12
0.16

0.04
0.16
0.11
0.12
0.07
0.08

1.32
0.59
0.09
0.30
0.52
0.74

0.26
0.17
0.21
0.19
0.23
0.21

25.39
12.52
0.18
2.40
5.07
12.62

0.00
0.00
0.68
0.12
0.02
0.00

3.75
1.80
1.09
1.35
1.68
2.10

540
565
539
540
540
540

444
470
444
444
444
444

9.23
0.45
0.06
0.07
0.12
0.09

6.73
0.31
0.07
0.02
0.10
0.07

2.50
0.58
−0.25
1.29
0.20
0.20

0.81
0.13
0.26
0.36
0.21
0.24

3.08
19.12
0.94
12.57
0.88
0.71

0.00
0.00
0.33
0.00
0.35
0.40

NA
1.79
0.78
3.64
1.22
1.22

540
540
540
540
540
540

444
444
444
444
444
444

0.13
0.13
0.07
0.10
0.08
0.07

0.06
0.09
0.06
0.07
0.05
0.07

0.91
0.43
0.08
0.36
0.58
−0.07

0.24
0.21
0.27
0.24
0.29
0.25

14.13
4.09
0.09
2.30
4.14
0.08

0.00
0.04
0.76
0.13
0.04
0.78

2.50
1.54
1.09
1.44
1.79
0.93

(continued)

95

Prisoner reentry experiences of adult males

Exhibit 52. Weighted means for transition services receipt bundles and items, by group and
wave (continued)

Variable
Wave 4
TransSvcsScore
Rcvd ANY transition services*
Rcvd legal assistance
Rcvd financial assist
Rcvd public financial assist
Rcvd public health care
insurance
Rcvd mentor
Rcvd docs for employ
Rcvd place to live
Rcvd transportation
Rcvd drivers license
Rcvd cloth/food bank

S
N

nonNonSVORI
S
SVORI
N
Mean Mean

Par.
Est.

SE

Test
Stat.

p Val.

OR

476
582
476
476
476

445
531
444
445
445

6.00
0.27
0.06
0.05
0.07

6.01
0.26
0.08
0.02
0.11

−0.01
0.07
−0.31
0.69
−0.47

0.76
0.14
0.27
0.38
0.24

−0.02
0.26
1.30
3.32
3.99

0.98
0.61
0.25
0.07
0.05

NA
1.07
0.73
2.00
0.62

476
476
476
476
476
476
476

445
445
445
445
445
445
445

0.06
0.09
0.07
0.04
0.06
0.05
0.05

0.07
0.04
0.06
0.05
0.05
0.05
0.06

−0.16
0.79
0.09
−0.28
0.10
0.12
−0.18

0.26
0.28
0.27
0.32
0.30
0.31
0.29

0.35
8.03
0.11
0.76
0.12
0.15
0.37

0.56
0.00
0.74
0.38
0.73
0.70
0.54

0.86
2.20
1.10
0.76
1.11
1.13
0.84

Note: NA = Not applicable. OR = odds ratio. Par. Est. = parameter estimate. p Val. = probability value. S = SVORI.
SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months postrelease; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations

SVORI participants were much more likely to report receiving
help pre-release obtaining drivers licenses (22%) and
documents for employment (41%) than non-SVORI
respondents (8% and 26%), although at levels substantially
below the more than the 80% who said they needed help
obtaining drivers licenses and the 55% who said they needed
documents for employment. The percentages reporting having
a mentor pre-release were substantially less than the 60% of
both groups who said they needed a mentor. SVORI
participants were 2.5 times more likely to report having a
mentor, but only 19% of SVORI participants (compare 8% nonSVORI) reported having a mentor.

The most commonly
reported service received
across the four interviews
was assistance obtaining
documents for
employment.

96

While most members of both groups reported receiving at least
one of the transition services prior to release, the proportions
declined dramatically following release. Only 59% of SVORI
program participants and 43% of non-SVORI comparison
subjects reported receiving ANY transition services during the
first 3 months following their release from prison. The most
commonly reported service across the four waves was
assistance obtaining documents for employment (e.g., Social
Security card, birth certificate). Perhaps the most surprising

Self-reported Service Receipt

finding in Exhibit 52 is how few of these men received support
as they transitioned back into the community. Although SVORI
participants received significantly more help—at least prerelease and immediately after release—less than 50% reported
receiving any of these services at the 9- and 15-month
interviews.

HEALTH SERVICES
Exhibit 42 showed the bundle scores for health service needs at
each interview wave. Prior to release, subjects reported
needing about one third of the five health services. At the time
of the 15-month post-release interview, men reported needing
about 28% of the services. The most commonly reported need
was for medical treatment, with substantial minorities also
indicating a need for substance use treatment and anger
management programs.
Exhibit 53 provides the means and test statistics for the receipt
of various health services, including medical, dental, mental
health, and substance abuse treatment; included in this bundle
is victims’ group counseling and anger management. Most
subjects indicated that they had received health services while
they were incarcerated, with medical and dental treatment
being the most commonly reported. Not surprisingly, as
medical treatment is mandated by law, there was no difference
in the receipt of medical and dental services while subjects
were incarcerated. There was also no difference in mental
health treatment, although non-SVORI subjects were slightly
more likely than SVORI program participants (19% to 16%) to
report receiving mental health treatment while incarcerated;
this finding is consistent with non-SVORI subjects having been
more likely to report needing mental health treatment (see
Exhibit 42). SVORI program participants were much more likely
to report receiving substance abuse treatment while they were
incarcerated—48% of SVORI and 38% of non-SVORI reported
receiving substance abuse treatment while incarcerated—and
anger management programming (33% SVORI, 26% nonSVORI).
Following release, the respondents were less likely to receive
any service; they also reported receiving many fewer services.
The health services bundle scores declined from about one third
pre-release (35.28 SVORI, 31.21 non-SVORI) to about one

97

Prisoner reentry experiences of adult males

Exhibit 53. Weighted means for health services receipt bundles and items, by group and
wave

Variable
Wave 1
Health Services Score
Rcvd ANY health services*
Rcvd medical tx
Rcvd dental tx
Rcvd MH tx
Rcvd substance use tx
Rcvd group for abuse victims
Rcvd anger mgmt program
Wave 2
Health Services Score
Rcvd ANY health services*
Rcvd medical tx
Rcvd dental tx
Rcvd MH tx
Rcvd substance use tx
Rcvd group for abuse victims
Rcvd anger mgmt program
Wave 3
Health Services Score
Rcvd ANY health services*
Rcvd medical tx
Rcvd dental tx
Rcvd MH tx
Rcvd substance use tx
Rcvd group for abuse victims
Rcvd anger mgmt program
Wave 4
Health Services Score
Rcvd ANY health services*
Rcvd medical tx
Rcvd dental tx
Rcvd MH tx
Rcvd substance use tx
Rcvd group for abuse victims
Rcvd anger mgmt program

S
N

NonNonS
SVORI
SVORI
N
Mean Mean

Par.
Est.

SE

Test
Stat.

p Val.

OR

863
863
861
863
854
862
862
863

834
834
830
833
821
834
834
833

35.28
0.88
0.58
0.50
0.16
0.48
0.07
0.33

31.21
0.83
0.55
0.47
0.19
0.38
0.03
0.26

4.08
0.40
0.14
0.11
−0.22
0.42
0.91
0.33

1.09
0.14
0.10
0.10
0.13
0.10
0.26
0.11

3.75
7.87
1.98
1.30
2.72
17.14
12.41
9.50

0.00
0.01
0.16
0.25
0.10
0.00
0.00
0.00

NA
1.49
1.15
1.12
0.81
1.52
2.48
1.40

529
529
525
526
524
529
529
529

455
455
450
452
449
455
455
455

12.75
0.53
0.25
0.05
0.09
0.28
0.01
0.08

11.06
0.45
0.23
0.07
0.08
0.23
0.01
0.05

1.69
0.32
0.10
−0.26
0.12
0.27
0.35
0.52

0.95
0.13
0.15
0.27
0.24
0.15
0.58
0.26

1.77
5.89
0.41
0.95
0.26
3.30
0.36
3.98

0.08
0.02
0.52
0.33
0.61
0.07
0.55
0.05

NA
1.37
1.10
0.77
1.13
1.31
1.42
1.68

541
565
537
540
535
541
540
540

444
470
440
443
439
443
444
444

12.70
0.50
0.30
0.08
0.08
0.22
0.01
0.08

12.50
0.50
0.29
0.08
0.07
0.23
0.01
0.07

0.20
0.02
0.04
−0.09
0.14
−0.04
−0.55
0.14

0.95
0.13
0.14
0.24
0.25
0.16
0.75
0.25

0.21
0.03
0.09
0.13
0.34
0.07
0.54
0.33

0.83
0.87
0.77
0.72
0.56
0.79
0.46
0.57

NA
1.02
1.04
0.92
1.15
0.96
0.58
1.15

476
582
472
473
470
474
476
476

445
531
442
443
440
443
445
445

10.86
0.37
0.27
0.09
0.07
0.18
0.02
0.03

11.93
0.41
0.27
0.10
0.09
0.21
0.00
0.05

−1.07
−0.14
−0.02
−0.09
−0.31
−0.19
2.08
−0.41

0.96
0.12
0.15
0.22
0.25
0.17
1.06
0.35

−1.12
1.34
0.01
0.17
1.56
1.26
3.81
1.38

0.26
0.25
0.91
0.68
0.21
0.26
0.05
0.24

NA
0.87
0.98
0.91
0.73
0.83
7.99
0.66

Note: NA = Not applicable. MH = mental health. OR = odds ratio. Par. Est. = parameter estimate. p Val. = probability
value. S = SVORI. SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 =
9 months post-release; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations

98

Self-reported Service Receipt

eighth at 3 months post-release (12.75 SVORI, 11.06 nonSVORI). The most commonly reported health service was
medical treatment (23% to 30%), with similar percentages
reporting receipt between the two groups. A slightly smaller
percentage (18% to 28%) reported receiving substance abuse
treatment since the previous interview. Less than 10% reported
receiving dental care. Anger management programming was
also reported to have been received by less than 10% following
release, although more than 25% of respondents said they
needed anger management programming.

DOMESTIC VIOLENCE SERVICES
Subjects were asked about their participation in domestic
violence support groups or batterer intervention programs—
services that were ranked as low-need by most of the men (see
Exhibit 54). About 10% of respondents (9% SVORI, 11% nonSVORI) reported needing domestic violence support groups or
batterer intervention programs at the pre-release interview (see
Exhibit 43), about the same percentage who reported receiving
services. Exhibit 54 shows the proportion of subjects who
reported participating in these two programs. SVORI program
participants were more likely pre-release to have participated in
a domestic violence group (10% versus 6%), but post-release
few subjects reported participating and there were no
differences between the two groups.

CHILD SERVICES
More than 60% of the men reported having children, as noted
earlier. The most commonly reported child-related needs were
help with modifying child support payments, but many also
indicated a need to improve their parenting skills. As shown in
Exhibit 55, 34% of SVORI program participants and 22% of nonSVORI comparison subjects reported receiving any child-related
services while they were in prison. The most commonly reported
services were assistance modifying child support debt and
participation in parenting skills classes. SVORI participants were
significantly more likely to report receiving these two services.

99

Prisoner reentry experiences of adult males

Exhibit 54. Weighted means for domestic violence services receipt bundles and items, by
group and wave

Variable
Wave 1
DomViolScore
Rcvd ANY DV services*
Rcvd batterer intervention
program
Rcvd DV support group
Wave 2
DomViolScore
Rcvd ANY DV services*
Rcvd batterer intervention
program
Rcvd DV support group
Wave 3
DomViolScore
Rcvd ANY DV services*
Rcvd batterer intervention
program
Need DV support group
Wave 4
DomViolScore
Rcvd ANY DV services*
Rcvd batterer intervention
program
Rcvd DV support group

S
N

NonNonS
SVORI
SVORI
N
Mean Mean

Par.
Est.

SE

Test
Stat.

p Val.

OR

863
863
862

834
834
834

7.43
0.11
0.05

4.72
0.07
0.04

2.71
0.53
0.28

1.00
0.18
0.25

2.71
9.02
1.29

0.01
0.00
0.26

NA
1.70
1.33

863

834

0.10

0.06

0.60

0.19

10.02

0.00

1.82

529
529
529

455
455
455

0.97
0.01
0.01

1.17
0.02
0.01

−0.21
−0.19
0.04

0.60
0.54
0.68

−0.35
0.12
0.00

0.73
0.72
0.96

NA
0.83
1.04

529

455

0.01

0.02

−0.35

0.56

0.40

0.53

0.70

540
565
540

444
470
444

1.35
0.02
0.01

1.40
0.02
0.01

−0.05
−0.10
−0.40

0.64
0.44
0.77

−0.07
0.05
0.27

0.94
0.83
0.60

NA
0.91
0.67

540

444

0.02

0.02

0.11

0.47

0.05

0.82

1.11

476
582
476

445
531
445

1.02
0.01
0.01

1.64
0.02
0.01

−0.62
−0.76
−0.31

0.67
0.51
0.74

−0.92
2.24
0.18

0.36
0.13
0.67

NA
0.47
0.73

476

445

0.01

0.02

−0.57

0.53

1.15

0.28

0.57

Note: NA = Not applicable. OR = odds ratio. Par. Est. = parameter estimate. p Val. = probability value. S = SVORI.
SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months postrelease; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations

Receipt of child-related services dropped dramatically following
release. Only 8% of SVORI and 3% of non-SVORI subjects
reported receiving any child-related services at the 3-month
post-release interview, and the percentages declined in
subsequent interviews.

SUMMARY AND CONCLUSIONS
Overall, the examination of service receipt shows that the men
who enrolled in SVORI programming received substantially
higher levels of services than men who received “treatment as
usual.” Although programming was concentrated on the prerelease phase (i.e., levels of service receipt were dramatically

100

Self-reported Service Receipt

Exhibit 55. Weighted means for child services bundles and items, by group and wave

Variable
Wave 1
ChildSvcsScore
Rcvd ANY child services*
Rcvd help getting child
support payments
Rcvd help modifying child
support debt
Rcvd help modifying
custody
Rcvd parenting classes
Rcvd help with child care
Wave 2
ChildSvcsScore
Rcvd ANY child services*
Rcvd help getting child
support payments
Rcvd help modifying child
support debt
Rcvd help modifying
custody
Rcvd parenting classes
Rcvd help with child care
Wave 3
ChildSvcsScore
Rcvd ANY child services*
Rcvd help getting child
support payments
Rcvd help modifying child
support debt
Rcvd help modifying
custody
Rcvd parenting classes
Rcvd help with child care
Wave 4
ChildSvcsScore
Rcvd ANY child services*
Rcvd help getting child
support payments
Rcvd help modifying child
support debt
Rcvd help modifying
custody
Rcvd parenting classes
Rcvd help with child care

S
N

NonNonSVORI
S
SVORI
N
Mean Mean

Par.
Est.

SE

Test
Stat.

p Val.

OR

508
508
507

503
503
502

11.26
0.34
0.06

5.74
0.22
0.02

5.53
0.59
1.22

0.99
0.15
0.37

5.61
16.67
10.90

0.00
0.00
0.00

NA
1.81
3.39

152

158

0.22

0.11

0.84

0.33

6.60

0.01

2.32

507

502

0.04

0.02

0.54

0.39

1.89

0.17

1.72

508
508

503
502

0.25
0.08

0.15
0.03

0.63
1.01

0.16
0.32

14.65
9.87

0.00
0.00

1.88
2.75

314
529
314

273
455
273

4.19
0.08
0.03

1.37
0.03
0.02

2.83
0.95
0.49

0.74
0.31
0.62

3.82
9.38
0.61

0.00
0.00
0.44

NA
2.58
1.62

116

109

0.14

0.04

1.49

0.58

6.59

0.01

4.43

314

273

0.03

0.00

17.68

0.45

0.00

>999

314
314

273
273

0.05
0.04

0.03
0.00

0.54
2.34

0.45
1.04

1.45
5.04

0.23
0.02

1.72
10.42

321
565
321

259
470
258

2.75
0.06
0.02

2.14
0.05
0.01

0.61
0.34
0.47

0.69
0.29
0.87

0.88
1.34
0.30

0.38
0.25
0.59

NA
1.40
1.60

116

105

0.11

0.04

1.07

0.61

3.08

0.08

2.92

321

258

0.01

0.03

−0.86

0.64

1.79

0.18

0.42

321
321

259
257

0.04
0.02

0.03
0.02

0.36
0.36

0.52
0.58

0.49
0.39

0.48
0.53

1.44
1.44

292
582
291

284
531
284

1.71
0.03
0.01

1.68
0.03
0.01

0.03
−0.04
−0.30

0.58
0.34
1.01

0.06
0.02
0.09

0.95
0.90
0.77

NA
0.96
0.74

115

127

0.05

0.05

−0.05

0.60

0.01

0.93

0.95

290

284

0.02

0.01

0.35

0.74

0.22

0.64

1.42

292
290

284
284

0.01
0.02

0.02
0.01

−0.70
0.72

0.64
0.72

1.20
0.99

0.27
0.32

0.50
2.05

1511

Note: NA = Not applicable. OR = odds ratio. Par. Est. = parameter estimate. p Val. = probability value. S = SVORI.
SE = standard error. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months postrelease; Wave 4 = 15 months post-release.
*Included for information; not included in bundle calculations

101

Prisoner reentry experiences of adult males

higher at that time period than at any of the post-release time
periods), SVORI appeared to increase access to services well
beyond release. Even 15 months after release, the SVORI
group still reported significantly higher rates of service receipt
than the non-SVORI group in many service areas.
The services that men were most likely to receive pertained to
the coordination of services, including post-release supervision,
case management, and working with someone to reintegrate
into society. SVORI participants were more likely to receive
transition services pre-release and through 9 months postrelease. Prior to release, most SVORI participants participated
in programs or classes to prepare for release and were also
significantly more likely than non-SVORI respondents to report
receiving the 10 services included in the transition bundle, with
help obtaining documents for employment being the most likely
service reported.
Although most adult SVORI programs listed employment as one
of their priorities and most men reported needing help finding
employment, most men enrolled in SVORI programs did not
report receiving employment-related services. Prior to release,
only 37% of SVORI program participants (compared to 20% of
non-SVORI respondents) reported that they had received any
employment-related services while incarcerated—about the
same as the 34% of SVORI participants who reported having
received employment-related services during the 3 months
following release.

SVORI program
participation greatly
increased the likelihood
of receiving a wide range
of services, but levels of
participation were less
than reported needs.

SVORI program participants were also much more likely than
non-SVORI respondents to report having received programs
such as cognitive behavior therapy that were directed at
changing their attitudes about criminal behavior—change that
some have suggested is a prerequisite to affecting recidivism.
About half (51%) of SVORI participants reported participating
in such programs prior to release compared to 36% of nonSVORI respondents.
Despite the increase in access to services afforded to those
participating in SVORI programs, the proportions of SVORI
program participants who received the service items were
usually substantially less than 100%. Although need clearly
should drive the receipt of services in the sense that individuals
should not receive unneeded services, need exceeded service
receipt for most service items. Overall, aggregate levels of

102

Self-reported Service Receipt

service receipt were substantially lower than aggregate levels
of service need (across all bundles and time periods, and
among both groups). For example, most men reported needing
employment, education, and skills services, and only about half
(51%) of SVORI participants reported receiving any educationrelated services and only 37% reported receiving any
employment-related services while they were incarcerated.
These levels were much higher, however, than what was
observed post-release, as only 34% of SVORI program
participants reported receiving any employment services during
the 3 months following release.
The chapter that follows reports the impact of SVORI
programming on several key domains. Detailed findings for
housing, employment, family/peer/community outcomes,
substance abuse and physical and mental health, and criminal
behavior/recidivism are presented.

103

Outcomes
The purpose of the multi-site evaluation was to identify the
impact of SVORI program participation on a range of postrelease outcomes, with a primary goal of examining the impact
of SVORI program participation on recidivism. The SVORI
programs were intended to affect a range of outcomes that
have been linked to recidivism, including housing, employment,
and substance use, using a case managed or wraparound
approach to provide integrated services based on individual
needs assessments.

…the approach to the
outcome analyses was to
test first-order effects of
SVORI program
participation on each of
the identified outcomes.

The implicit SVORI program model linking inputs, services, and
outcomes is shown in Exhibit 56. This presentation of the
program model makes clear that most services are intended to
affect intermediate outcomes which, in turn, are then assumed
to be correlated with improvements in criminal behavior. (Only
programs directed at changing attitudes towards criminal
behavior were envisioned to directly affect post-release criminal
behavior). Under this model, the outputs (services) directly
affect the intermediate outcomes that, in turn, impact the
criminal behavior outcomes. For example, suppose employment
services increase the likelihood of employment by 20% and
being employed reduces the likelihood of recidivism by 20%,
then receiving employment services would be associated with a
4% reduction in recidivism. Of course, the SVORI program
participants (and, to a lesser extent, the non-SVORI
respondents) received a variety of different services, each of
which could impact one or more intermediate outcomes that
could impact recidivism. There is little theoretical or empirical
guidance for the correct specification of such a complex
recidivism model and, thus, the approach to the outcome
analyses was to test first-order effects of SVORI program
participation on each of the identified outcomes.

105

Prisoner reentry experiences of adult males

Exhibit 56. SVORI Program Model

The previous chapter showed that SVORI program participants
were more likely than non-participants to report receiving a
wide range of reentry services, although the levels of service
provision were less than reported needs. Further, most service
provision was concentrated in the incarceration phase of the
program and diminished following release. In this chapter, a
wide range of outcomes are examined across the domains of
interest to the SVORI—housing, family and community,
employment, physical and mental health (including substance
use), and criminal behavior.
The effect of SVORI program participation on individual
outcomes was assessed by estimating a series of models in
which the outcome variables were regressed on the
dichotomous SVORI indicator variable (see the Introduction to
this report and Lattimore & Steffey, 2009). Weights to control
for observed differences between the SVORI and non-SVORI
groups were applied in the estimation of these outcome
models. The weights were based on propensity scores that
provided individual-level estimates of the probability of

106

Outcomes

assignment to a SVORI program. 34 The weights were
incorporated into the regression models using the survey
procedures in SAS® 9.1. Proc Survey Means was used to
generate weighted group means by group. This procedure
produces weighted means but not weighted standard errors.
So, to provide tests of significance of the effect of SVORI
program participation on outcomes, a second set of models was
estimated in which the outcome variables were regressed on
the dichotomous SVORI indicator variable using Proc Survey
Logistic or Regression for dichotomous or continuous dependent
variables, respectively. The odds ratios from the logistic
regressions provide one measure of treatment effects. The
percentage difference between SVORI and non-SVORI weighted
means provides another.
The following sections address the impact of SVORI program
participation on outcomes in the following domains: housing;
employment; family, peers, and community involvement;
substance use and physical and mental health; and criminal
behavior and recidivism.

HOUSING
As reported in the Introduction, the SVORI PDs said that 54%
of SVORI participants would receive housing placements or
referrals pre-release and 66% would be assisted post-release
(see Exhibit 3). The SVORI evaluation participants were much
less likely to report receiving help with housing. Pre-release,
28% of SVORI participants (compared to 13% of non-SVORI
respondents) reported help with housing, a percentage that
declined to 12% 3 months post-release (compared to 11% of
non-SVORI; see Exhibit 52). Thus, most of SVORI respondents
reported receiving no help with housing. Three “core” housing
outcomes were housing independence, housing stability, and
the extent of challenges faced in locating housing after release.
Prior to discussing these outcomes, data are presented on the
housing experiences of these men prior to and following
incarceration.
34

The propensity score model used observed characteristics to model
the likelihood that an individual with specified characteristics would
be selected or assigned to treatment, in this case to a SVORI
program. A logit model was estimated in SAS® 9.1.3 using PROC
MI and PROC MIANALYZE to accommodate item missingness, which
was relatively rare. Thirty-one independent variables were included
in this model. See Lattimore and Steffey (2009) for details.

107

Prisoner reentry experiences of adult males

In the 6 months prior to their current incarceration, 13% of the
men reported that they were homeless, staying in a shelter, or
had no set place to stay. Following release, most were not
homeless—less than 5% of either group reported homeless
status during the periods since the previous interview.
Exhibit 57 shows the percentage of each group reporting being
homeless, living in a shelter, or without a set place to live in
the 6 months prior to incarceration (Wave 1 interview
conducted 30 days prior to release) or since the previous
interview 35 (Waves 2 through 4 interviews conducted 3, 9, and
15 months post-release). Thus, self-reported housing
difficulties were lower following release than prior to
incarceration, although it is possible that there was higher
nonresponse for unhoused individuals in the post-release
interviews.
Exhibit 57. Self-reported
homeless, living in a
shelter, or without a set
place to live, by group
and wave

100%

SVORI
Non-SVORI

80%
60%
40%
20%

13%

13%
1%

1%

1%

3%

3%

4%

0%
Wave 1

Wave 2

Wave 3

Wave 4

Note: Differences between SVORI and non-SVORI were not significant at the
0.05 level. (Wave 3 difference was significant at the 0.06 level.)

The respondents also reported with whom they were living.
Depending on the interview wave, 7% to 11% of the men
reported that they were living alone. The most common
categories of people with whom the men reported living were

35

108

Individuals who missed the previous follow-up interview were asked
about the 6 months prior to the current interview. Individuals who
were reincarcerated were asked about the period immediately prior
to their current incarceration. Individuals who had been
incarcerated for the entire period since the last interview were
coded as missing on housing questions.

Outcomes

their mother/stepmother and boy/girlfriend/fiancé. 36 Exhibit 58
shows the percentages of each group that reported living with
their mother/stepmother or father/stepfather. More than one
third of the men were living with their mothers during the
period immediately following their release from prison, and
about half that percentage reported living with their fathers.
These percentages declined over time such that by 15 months
post-release only about one quarter reported living with their
mothers.
Exhibit 58. Percentages
living with mothers or
fathers post-release, by
group (SVORI and nonSVORI) and post-release
follow-up wave (2, 3,
and 4)

100%

80%

SVORI-2

Non-SVORI-2

SVORI-3

Non-SVORI-3

SVORI-4

Non-SVORI-4

60%

40%

35

33

30

29

26

27
18

20%

15

17

14

13

14

0%
Mother/stepmother

Father/stepfather

Note: Differences between SVORI and non-SVORI were not significant at the
0.05 level.

Exhibit 59 shows the percentages of each group that reported
living with partners, spouses, or children over the 15 months
following their release. As can be seen, the number reporting
living with spouses increased over time, as did the number
reporting living with children. The complete responses to the
questions related to with whom the respondent was living are
included as Exhibit A-5 in Appendix A.
The co-residence patterns were similar for the SVORI and nonSVORI groups, with the only significant differences being that
the non-SVORI group was significantly more likely to report
living with a boyfriend/girlfriend/fiancé at the 9-month

36

Subjects were allowed to select as many as applied; for example,
someone could indicate that they were living with their mother,
father, and brother(s). On average, the respondents indicated they
were living with individuals in less than two categories.

109

Prisoner reentry experiences of adult males

Exhibit 59. Percentages
living with partners,
spouses, or children
post-release, by group
(SVORI and non-SVORI)
and post-release followup wave (2, 3, and 4)

100%
80%

SVORI-2

Non-SVORI-2

SVORI-3

Non-SVORI-3

SVORI-4

Non-SVORI-4

60%
40%
20%

18 19

23

29 27
26
7

9

12
9 10 10

13 13

17 17 18 19

0%
Boy/girlfriend,fiance

Husband/wife

Child/step

Note: Non-SVORI comparisons were more likely to report living with a
boy/girlfriend/fiancé at the 9 month interview; other differences between
SVORI and non-SVORI were not significant at the 0.05 level

interview and significantly more likely to report living with a
grandparent at the 15-month interview.
Three dimensions of housing—housing independence, housing
stability, and the extent of challenges faced in locating
housing—were examined as reentry outcomes. The SVORI and
non-SVORI groups were compared on these outcomes at 3, 9,
and 15 months post-release. Exhibit 60 shows the weighted
proportion of men in each group (with estimates, standard
errors, and odds ratios from the logistic regression models) who

110

ƒ

were classified as “housing independent” (defined as
living in their own house or apartment, contributing to
the costs of housing, or having their name on the lease
or mortgage of the place where they currently lived);

ƒ

were classified as having stable housing (defined as
having lived in only one place during the reference
period or two places if the move was to secure their own
place or a nicer place); and

ƒ

did not experience housing challenges (respondents
were classified as not having housing challenges if they
were not homeless, reported that they did not have
trouble finding a place to live, and reported that their
current living situation was better or about the same as
their last one).

Outcomes

Exhibit 60. Weighted
means and regression
parameter estimates for
housing outcomes

Wave 2
Housing independence
Housing stability
No housing challenges
Wave 3
Housing independence
Housing stability
No housing challenges
Wave 4
Housing independence
Housing stability
No housing challenges

SVORI
Mean
N = 529
0.723
0.784
0.837
N = 565
0.818
0.695
0.847
N = 582
0.861
0.672
0.815

NonSVORI
Mean
N = 455
0.692
0.781
0.815
N = 470
0.829
0.709
0.820
N = 531
0.798
0.728
0.833

Par.
Est.

SE

OR

0.151
0.017
0.154

0.143
0.157
0.170

1.163
1.017
1.167

−0.074
−0.070
0.201

0.172
0.142
0.175

0.929
0.932
1.222

0.450*
−0.267
−0.123

0.179
0.147
0.176

1.569
0.766
0.884

Note: OR = odds ratio. Par. Est. = parameter estimate for weighted regression.
SE = standard error. N’s are the total responses for each wave of interviews
and do not reflect any item missingness. Wave 1 = 30 days pre-release; Wave
2 = 3 months post-release; Wave 3 = 9 months post-release; Wave 4 = 15
months post-release.
* p < 0.05

There were few
differences in housing
outcomes.
Housing independence
improved gradually for
both groups over the
post-release follow-up
period.

As can be seen, the SVORI and non-SVORI groups are similar
along these core housing dimensions, indicating that SVORI
programming did not significantly improve the post-release
housing experiences for returning adult male prisoners. Given
the relative few who reported receiving help with housing, this
result is not surprising. The single statistically significant
difference (at the 0.05 level) was that the SVORI group, on
average, was more likely than the non-SVORI group to report
being housing independent 15 months after release.
Also of interest in Exhibit 60 are the variable temporal patterns
observed among these three core housing measures. For
example, housing independence improved gradually over the
post-release follow-up period (with 72% of the SVORI and 70%
of the non-SVORI men classified as “housing independent” at
the 3-month interview and 86% of SVORI and 80% of nonSVORI classified as “housing independent” at the final,
15-month interview), housing stability declined over time (with
the highest levels of stability being observed at the immediate
post-release time period and the lowest being observed at the
15-month post-release time period). This pattern may be
because the 9- and 15-month post-release interviews had
longer reference periods (6 months) than the 3-month post-

111

Prisoner reentry experiences of adult males

release interview (3 months) so that respondents had more
opportunities to experience instability during the 9- and 15month interviews. There was little difference over time in the
measure of housing challenges, which is perhaps the broadest
measure of difficulty in finding quality housing. Specifically, the
percentages of men in both groups indicating they had
experienced housing challenges ranged from 81% to 84%
across the three follow-up interviews.
In addition to the three core housing measures, several other
dimensions of housing were measured that provide insight into
the men’s overall post-release housing experiences. For
example, one of the individual measures that was included in
the “housing independence” measure was whether the
respondent lived in his own house or apartment, lived in
someone else’s house or apartment, or was homeless, living in
a shelter, or without a set place to live. Exhibit 61 shows the
percentage of men in each group who reported that they were
living in their own homes at each follow-up. As can be seen,
SVORI program participants were more likely than non-SVORI
comparison subjects to report that they were living in their own
house or apartment, although differences (as measured by the
significance of the logistic regression parameter estimate) were
not significant at the 0.05 level (p values were 0.1, 0.15 and
0.08, for Waves 2, 3, and 4). Further, the percentages of both
groups reporting living in their own places increased over time.
Exhibit 61. Percentage
reporting living in own
house or apartment;
weighted means by
group and follow-up
wave

100%
SVORI
Non-SVORI

80%
60%
36

40%
25

31

37

32

21

20%
0%
Wave 2

Wave 3

Wave 4

Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 =
9 months post-release; Wave 4 = 15 months post-release.

112

Outcomes

Another individual component of housing independence also
showed improvement (for both the SVORI and non-SVORI
groups) during the post-release follow-up period. Specifically,
the percentage of men who contributed to housing costs
increased for SVORI program participants from 70% at 3
months to 79% at 9 months to 85% at 15 months. The
percentages also increased for non-SVORI comparison subjects,
rising from 67% at 3 months post-release to 81% at 9 months
before dropping to 78% at 15 months post-release. SVORI
program participants were more likely to report having their
name on the mortgage or lease at 3 months post-release—22%
versus 16%. Smaller, but insignificant differences persisted at 9
and 15 months (29% versus 25% at 9 months and 35% and
31% at 15 months).
Neighborhood quality was a final relevant dimension of housing.
At each post-release interview the men were asked a series of
questions about the quality of the neighborhoods in which they
lived. The following items were combined to create a score
measuring neighborhood quality:
ƒ

“It is hard to stay out of trouble in your neighborhood.”

ƒ

“Drug selling is a major problem in your neighborhood.”

ƒ

“You think your neighborhood is a good place to live.”

ƒ

“You think your neighborhood is a good place to find a
job.”

ƒ

“Living in your neighborhood makes it hard to stay out
of incarceration.”

When the mean neighborhood quality scores at the three postrelease time periods were examined, little variability over time
was found (weighted means on the 15-point scale ranged from
9.3 to 10 across the groups and follow-up periods), indicating
that the men in both groups had similar perceptions of the
neighborhoods in which they were living at each time period at
which they were interviewed.

EMPLOYMENT
Employment was a primary focus of many of the SVORI
programs, and all 12 adult impact site Program Directors felt
that it would be fair to determine program effectiveness by
examining employment (see Exhibit 2). Employment was also
of considerable importance to the subjects, who consistently

113

Prisoner reentry experiences of adult males

indicated high levels of need for services to improve their
employment, education, and other skills (see Exhibit 40). As
was shown in the previous chapter, the SVORI programs were
successful in increasing the likelihood that individuals received
employment-related services during confinement and during
the first few months following release (see Exhibit 51).
However, levels were far less than reported need, which neared
100% for all respondents. Specifically, only 37% of SVORI
program participants (compared with 20% of non-SVORI)
reported receiving, during their incarcerations, any
employment-related services (such as employment readiness
classes, help preparing a resume, advice about interviewing, or
names of potential employers). Further, only 25% of SVORI
program participants (compared with 16% of non-SVORI)
received help obtaining documents related to employment such
as Social Security cards or birth certificates.
Most of the respondents (64% SVORI, 68% non-SVORI; tstatistic = −2.04) reported working in the 6 months prior to
incarceration and about 60% (59% SVORI, 62% non-SVORI; tstatistic = −0.87) reported that they had supported themselves
with jobs during that same period. Of those working, three
quarters (75% SVORI, 73% non-SVORI; t-statistic = 0.65)
indicated the job was permanent. More than half (57%) of
those who had worked expected to be able to return to a
previous job.
Extensive data were collected from respondents to assess their
post-release employment experiences. Of these measures,
several core employment outcomes were identified:
ƒ

current support of oneself with a job;

ƒ

the number of months worked during the reference
period;

ƒ

worked for each month during the reference period;

ƒ

the number of months at which the same job was held;

ƒ

receipt of formal pay from a job; and

ƒ

whether the job provided benefits (a summary measure
indicating whether the job provided health insurance or
paid leave).

The results for these outcomes at 3, 9, and 15 months postrelease are shown in Exhibit 62.

114

Outcomes

Exhibit 62. Weighted
means and regression
parameter estimates for
employment outcomes

Wave 2
Support self with job
Number months worked
Worked each month
Received formal pay
Job benefits
Wave 3
Support self with job
Number months worked
Worked each month
Received formal pay
Job benefits
Wave 4
Support self with job
Number months worked
Worked each month
Received formal pay
Job benefits

SVORI
Mean
N = 529
0.64
2.04
0.38
0.84
0.47
N = 565
0.68
3.83
0.43
0.80
0.53
N = 582
0.71
3.70
0.44
0.78
0.52

NonSVORI
Par.
Mean
Est.
N = 455
0.59
0.230
1.96
0.079
0.39
−0.056
0.74
0.604
0.39
0.337
N = 470
0.68
0.027
3.73
0.102
0.44
−0.033
0.77
0.178
0.42
0.472
N = 531
0.60
0.481
3.50
0.197
0.42
0.045
0.74
0.183
0.44
0.326

p
Value

OR

0.086
0.420
0.716
0.001
0.028

1.258
NA
0.945
1.829
1.400

0.848
0.536
0.823
0.310
0.001

1.027
NA
0.968
1.195
1.602

0.001
0.252
0.772
0.306
0.034

1.617
NA
1.046
1.201
1.386

Note: OR = odds ratio. Par. Est. = parameter estimate for weighted regression.
NA = not applicable, since model was a regression. p value = probability value
of test statistic. N’s are the total responses for each wave of interviews and do
not reflect any item missingness. Wave 1 = 30 days pre-release; Wave 2 = 3
months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months
post-release.

Although SVORI program
participation was not
associated with increased
likelihood of having a
job, SVORI program
participants were more
likely to report
supporting themselves
with a job and having a
job with formal pay and
benefits.

The results in Exhibit 62 suggest that SVORI program
participants had better post-release employment experiences
than the non-SVORI comparison subjects—if only moderately
so. They were more likely to report that they were currently
supporting themselves with a job at 3 and 15 months postrelease. They reported working about the same number of
months on average—about two thirds of the available months
(2 of 3 months immediately following release, and about 4 of 6
months at the 9- and 15-month interviews)—and were equally
likely to have reported working all months in the reference
period.
SVORI participants appear to have secured better jobs—jobs
that provided formal pay and benefits. A breakdown of the
benefits measure, which includes insurance and paid leave, is
shown in Exhibit 63. As can be seen, SVORI program
participants were more likely than non-SVORI respondents to

115

Prisoner reentry experiences of adult males

Exhibit 63. Weighted
means for self-report
that job provides
insurance or paid leave,
by group (SVORI, nonSVORI) and data
collection wave (2, 3, or
4)

100%

SVORI-2
SVORI-3
SVORI-4

80%
60%

47

44
40%

36
29

33

36

35

NON-SVORI-2
NON-SVORI-3
NON-SVORI-4

41
33

40
32

37

20%
0%
Insurance

Paid leave

report having formal pay and benefits across all waves, but the
most substantial differences are in having a job that provides
insurance. Differences between SVORI and non-SVORI in
reported insurance is significant at p < 0.1 at 3 months and
p < 0.05 at 9 and 15 months. The difference in paid leave is
significant at the p < 0.05 level at 9 months.
Another measure of potential interest is hours worked. SVORI
program participants who reported working reported more
hours worked at 3 and 15 months post-release (39.6 for
SVORI, 37.9 for non-SVORI at 3 months; 42.5 for SVORI, 41.4
for non-SVORI at 15 months). There was no difference at 9
months (40.9 for SVORI, 40.4 for non-SVORI). Additionally,
SVORI program participants were more likely to report not
having problems finding a job (30% versus 25% at 3 months;
35% versus 32% at 9 months; and 37% versus 30% at 15
months; the latter difference was significant at the .05 level).
Program effects were identified, and men’s post-release
employment situations were compared to their preincarceration
employment experiences. Exhibit 64 shows the percentage of
each group who reported supporting themselves with a job in
the preceding period (6 months prior to incarceration or since
the last interview or equivalent time period). As can be seen,
supporting oneself with a job was reported at a higher rate than
in the pre-release interview at all post-release interviews
(Waves 2 through 4) for SVORI program participants, whereas
non-SVORI respondents were about equally as likely as their

116

Outcomes

Exhibit 64. Weighted
means for self-report of
supporting self with a
job, by group and data
collection wave

SVORI

100%

Non-SVORI

80%
60%

59

62

64

68

68

71

59

60

40%
20%
0%
Wave 1

Wave 2

Wave 3

Wave 4

Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 =
9 months post-release; Wave 4 = 15 months post-release.

preincarceration experience to report supporting themselves
with jobs at 3 and 15 months.
Finally, very few men reported receiving money from illegal
activities across all three post-release time periods, in contrast
to the 6 months prior to incarceration. About 5% of both
groups reported receiving money from illegal activity during the
3 months post-release, a percentage that increased to about
10% for both groups at 9 and 15 months post-release. These
percentages are in comparison to the more than half (56%
SVORI, 54% non-SVORI) who reported receiving money from
illegal activity during the 6 months prior to incarceration.
In summary, SVORI program participation led to modest
increases in employment-related services. Although SVORI
program participation appeared to have had little impact on
whether someone was working, program participation was
associated with modest improvements in several employmentrelated outcomes.

FAMILY, PEERS, AND COMMUNITY
INVOLVEMENT
Family Relationships

Family relationships have been shown to be extremely
influential for returning prisoners. However, because none of
the adult SVORI programs focused on family services, no

117

Prisoner reentry experiences of adult males

family-related measures were identified as key outcomes.
However, several aspects of family relationships, including
family emotional support, family instrumental support, quality
of intimate-partner relationships, and quality of relationship
with children were examined. Not surprisingly, given the lack of
emphasis on family-related services among the SVORI
programs, few differences between the SVORI and non-SVORI
groups were observed in these measures at any of the postrelease time periods. In addition, for both groups, the levels of
family emotional support, family instrumental support and the
quality of intimate-partner relationships declined over the three
post-release time periods; little variation was observed (for
both groups) over the three post-release time periods in the
scale measuring the quality of relationship with their children
(for those with children).
Exhibit 65 shows the percentage of men at each interview who
reported (1) being married and (2) being married or in a steady
relationship. Although the proportions of men who reported
being currently married (9–13%) were similar across all time
periods, the men were more likely to report intimate
partnerships at all post-release time periods than during the
preincarceration time period. Interestingly, the percentages
reporting being in a steady relationship or married declined from
the 9-month to the 15-month period, perhaps reflecting the fact
that substantial numbers of these subjects were reincarcerated
by 15 months, as reported later in this chapter.
Somewhat fewer than two thirds of the men reported having
children (62% of SVORI, 64% of non-SVORI) at the pre-release
interview—percentages that remained fairly constant over the
three post-release interviews. Only about one third of those who
reported having children said that they had been required to pay
child support prior to their incarceration. Following release,
about two thirds of those required to make payments said that
they were making the required child support payments.

118

Outcomes

Exhibit 65. Marital status and intimate partnerships by group and wave

SVORI-1
Non-SVORI-2

100%

Non-SVORI-1
SVORI-3

SVORI-2
Non-SVORI-3

80%
64
58

60%
39

40%

20%

9

10

9

11

10

11

11

57

64
56

58

40

13

0%
Married

Married/Steady Relationship

Peer Relationships

Peer relationships may be influential for returning prisoners.
Importantly, as with family relationships, peer relationships
may be both positive and negative. For example, receiving
critical instrumental support from peers, such as help with rides
to appointments, may be beneficial. On the other hand, if one’s
peers are criminally involved, such relationships may be
detrimental to reentry success.
Because many SVORI programs included cognitive-behavioral
components intended to teach inmates to change criminal
behavior by modifying which individuals they associated with
(among other topics), it was appropriate to measure the
“negative exposure” reported by the men both from friends
with whom they spent time and from individuals with whom
they lived. A summary measure of negative exposure was
developed that included indicators of the extent to which the
respondent reported the following:
ƒ

living with people who had ever been incarcerated, used
illegal drugs, engaged in any other illegal activity, or
used alcohol in their presence

119

Prisoner reentry experiences of adult males

ƒ

spending time with friends who were not employed, got
them “in trouble,” had been incarcerated, had assaulted
someone, had committed theft, or had sold drugs

The composite measure has a range of from 0 to 14 and is
coded such that higher values indicate less negative exposure.
The results for this measure suggested that there was little
difference between groups or across the post-release time
periods, as weighted averages ranged between 10.8 and 11.3.
Most men (81%) reported that prior to incarceration they had
friends who had been incarcerated. They were asked in the
follow-up interviews whether they had close friends who had
been incarcerated. Most said they did, although the percentage
declined over time—from about 74% (both groups) at 3 months
to less than 70% at 15 months.
Positive peer support was also measured. Specifically, the men
were asked about the following types of instrumental support
from their friends:
ƒ

help or advice on finding a place to live;

ƒ

help or advice on finding a job;

ƒ

support for dealing with a substance abuse problem;

ƒ

transportation to work or other appointments, if needed;
and

ƒ

financial support.

There was little variation across post-release time periods or
between groups in this measure, which could have values of
from 0 to 15, with higher scores indicating more instrumental
support. Across the three waves, weighted average scores
ranged from 9.3 to 9.9, with no difference between the groups.
Community Involvement

To assess the extent to which men became involved in their
communities after release, they were asked whether they had
(1) done volunteer work in any programs in the community
(e.g., youth groups, programs for the elderly); (2) done
mentoring with peers, youth, or other community members;
(3) participated in any local organizations like clubs, sports
teams, ethnic or racial pride groups, political organizations, or
other community groups; (4) voted in any political election
(including general elections, primary elections, and special
referendums); or (5) served in a Neighborhood Watch or tenant

120

Outcomes

patrol program. 37 A dichotomous measure was developed that
was equal to 1 if the respondent reported doing any of the
activities in the period since the previous interview.
As shown in Exhibit 66, civic action was low for both groups,
although it increased over time—particularly for SVORI program
participants such that by 15 months post-release SVORI
participants were significantly more likely to report participating
in one or more of the civic measures than the non-SVORI
comparison subjects.
Exhibit 66. Percentage
reporting civic action
since release/last
interview

100%

SVORI
Non-SVORI

80%
60%
40%
25

24

29

27

34

28

20%
0%
Wave 2

Wave 3

Wave 4

Note: Difference at Wave 4 is significant at p < 0.05. Wave 1 = 30 days prerelease; Wave 2 = 3 months post-release; Wave 3 = 9 months post-release;
Wave 4 = 15 months post-release.

SUBSTANCE USE AND PHYSICAL AND
MENTAL HEALTH
Substance Use

At the pre-release interview, about 95% of the men in this
study reported having used drugs during their lifetimes and
about two thirds reported having used drugs during the 30 days
prior to their current incarceration. Drug use has been identified
as a correlate of recidivism, and reducing drug use has been
the focus of many interventions. For the 12 SVORI impact
programs, PDs reported that less than half of SVORI program
participants were participating in either comprehensive drug
37

The question initially included “participated in the activities of a
church, mosque, temple, or other religious group” in this measure,
but later excluded the option to better capture the other measures
(most participants participated only in church services).

121

Prisoner reentry experiences of adult males

treatment programs or programs like Alcoholics Anonymous or
Narcotics Anonymous (see Exhibit 3). These estimates were
loosely consistent with the self-reports of the evaluation
respondents: 48% of SVORI program participants reported
receiving alcohol or drug treatment while incarcerated,
significantly more than the 38% of non-SVORI respondents
(see Exhibit 53). SVORI program participants were also
somewhat more likely to report treatment in the immediate
months following release (28% compared with 23%).
Substance use outcomes were measured both by self-report
during all three follow-up interviews and by oral fluids drug
tests administered to nonincarcerated respondents at the
3- and 15-month interviews. The results for the core substance
use outcomes are shown in Exhibit 67. This exhibit shows
abstinence measures, for example “no self-reported drug use.”
Thus, higher percentages are better because they suggest less
use.

Self-reported abstinence
was generally higher for
the SVORI group than for
the non-SVORI group

Only about half of both
groups reported no drug
use or had a negative
drug test at 30 days postrelease.

Self-reported abstinence was generally higher for the SVORI
group than for the non-SVORI group across all follow-up
waves; in several cases these differences were statistically
significant. Similar results obtained for the measure that
combined either self-reported drug use over the past 30 days
or confirmed (by drug tests) drug use, although these
differences are not statistically significant.
The patterns for substance use, based on the combined selfreport and drug test measures, are shown graphically in
Exhibits 68 and 69. Exhibit 68 shows the percentages of both
groups who reported no drug use since release (Wave 2) or the
last interview (Wave 4) and who did not refuse or test positive
on the drug test. Only about half of both groups reported no
drug use or had a negative drug test at 30 days post-release.
This proportion dropped to 40% or less at 15 months postrelease. 38 Exhibit 69 shows the percentages of each group who
reported no drug use in the past 30 days and who did not
refuse or test positive on the drug test. The results are similar
to those in Exhibit 68 and suggest high levels of drug use by
both groups.

38

122

Individuals who were incarcerated for the 6 months prior to the 15month interview were not asked about drug use, nor did they take
a drug test.

Outcomes

Exhibit 67. Weighted means and regression parameter estimates for substance use
outcomes

SVORI
Mean

NonSVORI
Mean

N = 529

N = 455

No self-reported drug use

0.74

No self-reported drug use other than marijuana
or steroids

Par.
Est.

p Value

OR

0.70

0.170

0.243

1.185

0.85

0.85

−0.030

0.868

0.970

No self-reported drug use past 30 days

0.79

0.77

0.160

0.313

1.174

No self-reported drug use other than marijuana
or steroids past 30 days

0.88

0.89

−0.066

0.747

0.936

No self-reported drug use past 30 days or
positive drug test

0.54

0.52

0.093

0.475

1.098

N = 565

N = 470

No self-reported drug use

0.57

0.52

0.201

0.125

1.223

No self-reported drug use other than marijuana
or steroids

0.74

0.71

0.177

0.227

1.194

No self-reported drug use past 30 days

0.69

0.62

0.301

0.028

1.351

No self-reported drug use other than marijuana
or steroids past 30 days

0.81

0.79

0.088

0.589

1.092

—

—

—

—

—

N = 582

N = 531

No self-reported drug use

0.58

0.50

0.337

0.012

1.401

No self-reported drug use other than marijuana
or steroids

0.78

0.72

0.311

0.045

1.365

No self-reported drug use past 30 days

0.66

0.62

0.1512

0.2784

1.163

No self-reported drug use other than marijuana
or steroids past 30 days

0.82

0.80

0.168

0.326

1.182

No self-reported drug use past 30 days or
positive drug test

0.46

0.43

0.118

0.381

1.126

Wave 2

Wave 3

No self-reported drug use past 30 days or
positive drug test
Wave 4

Note: Par. Est. = parameter estimate for weighted regression. p value = probability value of test statistic. OR = odds
ratio. N’s are the total responses for each wave of interviews and do not reflect any item missingness. Wave 1 = 30
days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months postrelease.

123

Prisoner reentry experiences of adult males

Exhibit 68. Percentages
of each group who
reported no drug use
since release/last
interview and had no
positive drug test

100%

SVORI
Non-SVORI

80%
60%

51

49
41

40%

35

20%
0%
Wave 2

Wave 4

Note: Difference at Wave 4 is significant at p < 0.1. Wave 1 = 30 days prerelease; Wave 2 = 3 months post-release; Wave 3 = 9 months post-release;
Wave 4 = 15 months post-release.

Exhibit 69. Percentages
of each group who
reported no drug use in
the past 30 days and
had no positive drug
test

100%

SVORI
Non-SVORI

80%
60%

54

52

46

43

40%
20%
0%
Wave 2

Wave 4

Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 =
9 months post-release; Wave 4 = 15 months post-release.

Marijuana was by far the most common drug reported used in
the self-reports of specific drug use. Additionally, over time,
non-SVORI comparison subjects were less likely to report
having not used marijuana such that the difference was
significant at 15 months. At the 3-month post-release
interview, 80% of SVORI and 76% of non-SVORI subjects
reported no marijuana use. At 9 months post-release, 68% of
SVORI and 63% of non-SVORI subjects reported no marijuana
use. At the final follow-up interview, 66% of SVORI and 59% of
non-SVORI subjects reported no marijuana use. Cocaine use
was also reported—particularly during the later interviews. At 3

124

Outcomes

months, 93% of both groups reported no cocaine use. At 9
months, 84% of SVORI and 87% of non-SVORI subjects
reported no cocaine use. At 15 months, 88% of SVORI and
85% of non-SVORI subjects reported no cocaine use.
The results suggest that SVORI program participants were
doing somewhat better with respect to drug use but that all
men continued to be at high risk for continuing drug use.
Physical Health

Physical health services were not a major programmatic focus
among the SVORI programs; therefore, no core physical health
outcomes were identified as relevant for analysis of program
effects in the evaluation. Of interest, however, is the physical
health status of the men during the post-release time period.
Several dimensions were measured in the post-release
interviews, including specific physical health conditions
experienced by the respondents (including asthma, chronic
back pain, high blood pressure, arthritis, hepatitis B or C, heart
trouble, diabetes, tuberculosis, and HIV/AIDS) and the SF-12
physical health scale, which measures five dimensions of
physical health functioning (including moderate activities such
as moving a table, climbing several flights of stairs,
accomplishing less than one would have like to accomplished
because of physical health, being limited in the kind of work or
activities done as a result of physical health, and pain that
interferes with normal work).
Based on a composite measure that summed the number of
physical health conditions experienced by the men, the results
suggest that members of both groups on average had less than
one of the nine conditions, with no differences evident between
the SVORI and non-SVORI groups and no temporal trend
apparent during the post-release follow-up period. Likewise, the
SVORI and non-SVORI groups scored similarly on the SF-12
physical health scale, with no evidence of temporal trends. The
SF-12 is normed to an average of 50 (range of 0 to 100), with
higher scores indicating better results. Across the three followup interviews, both groups averaged about 52. The postrelease scores were slightly lower than the 53 that the groups
scored on the Wave 1 interview. These findings suggest that no
major differences in physical health status occurred throughout
the entire observation period.

125

Prisoner reentry experiences of adult males

Mental Health

Two core mental health outcomes were identified: the SF-12
mental health scale (a measure of mental health functioning)
and the GSI (an index of mental health status that ranges from
45 to 225, with higher scores indicating better status). There
was no difference between the two groups at Wave 1 on either
of these two scales. At Wave 1, the SVORI group averaged 48.9
on the SF-12 and 66.6 on the GSI; the non-SVORI group
averaged 48.5 on the SF-12 and 68.1 on the GSI. These results
suggest that pre-release, the non-SVORI group had slightly
worse scores, but the differences were not significantly
different. It is important to remember, however, that the nonSVORI group was more likely at Wave 1 to report needing
mental health services a little or a lot (29% versus 22%) and
also more likely to report needing mental health services a lot
(10% versus 6%). The non-SVORI group was also more likely
pre-release to report having received mental health treatment
while they were incarcerated (20% versus 16%).
Exhibit 70 shows the results for the post-release interviews. As
can be seen, SVORI group members had slightly better mental
functioning as indicated by the SF-12 at 3 and 15 months postrelease. Results were also better for SVORI group members at
15 months on the GSI-45.
Exhibit 70. Weighted
means and regression
parameter estimates for
mental health outcomes

Wave 2
SF-12 mental health
GSI-45
Wave 3
SF-12 mental health
GSI-45
Wave 4
SF-12 mental health
GSI-45

SVORI
Mean
N = 529
51.01
164.13
N = 565
48.82
160.56
N = 582
48.82
161.06

Non-SVORI
Par. Est.
Mean
N = 455
49.80
1.211
162.76
1.363
N = 470
48.13
0.692
160.82
−0.257
N = 531
47.15
1.666
158.25
2.810

p Value
0.078
0.285
0.332
0.859
0.018
0.046

Note: Par. Est. = parameter estimate for weighted regression. p value =
probability value of test statistic. N’s are the total responses for each wave of
interviews and do not reflect any item missingness. Wave 1 = 30 days prerelease; Wave 2 = 3 months post-release; Wave 3 = 9 months post-release;
Wave 4 = 15 months post-release.

126

Outcomes

Focusing on these key outcomes, there is some evidence that
SVORI program participation had an impact on the mental
health status of men. In addition, as in the pattern observed for
physical health, the mental health status of men remained fairly
stable over time although weighted average scores diminished
slightly over time.

CRIMINAL BEHAVIOR AND RECIDIVISM
The SVORI logic model suggests that services that are
responsive to needs will result in improvements in intermediate
outcomes including housing, employment, and substance
abuse. Improvements in these outcomes, in turn, are
hypothesized to result in improvements in criminal behavior.
Results presented here suggest that participants in SVORI
programs were more likely to report receipt of a variety of
services, although service receipt was far less than universal.
Further, the SVORI program participants registered modest
improvements in many measures of the intermediate
outcomes. At issue is whether the observed differences in
intermediate outcomes are sufficient to generate measurable
differences in recidivism, even if the research hypothesis is
correct.
Because of the importance of recidivism, multiple measures
were included in the evaluation to determine program effects
on desistance from criminal activity. These measures include
self-reported and official measures of criminal behavior. Core
criminal behavior/recidivism outcomes based on “unofficial”
(i.e., self-reported) data sources are shown in Exhibit 71.
The first measure listed in the exhibit does not directly measure
criminal behavior, but rather perpetration of violence.
Respondents were asked about several specific types of
violence: threatening to hit, throwing, pushing/grabbing/
shoving, slapping/kicking/biting/hitting, and threatening to use
or using a weapon. The measure was scored 1 if the respondent
answered yes to any of these queries and 0 otherwise. The men
in the SVORI group were slightly less likely to report having
perpetrated violence than the men in the comparison group at
each post-release time period, but none of the differences was
statistically significant. Interestingly, a measure parallel to the
perpetration measure (but which assessed victimization) also

127

Prisoner reentry experiences of adult males

Exhibit 71. Weighted means and regression parameter estimates for core self-report
recidivism outcomes

SVORI
Mean

Non-SVORI
Mean

N = 529

N = 455

No perpetration of violence

0.73

Complied with conditions of supervision

Par. Est.

p Value

OR

0.71

0.117

0.422

1.124

0.78

0.78

0.005

0.977

1.005

No criminal behavior

0.79

0.73

0.327

0.034

1.386

No violent or weapons crimes

0.90

0.91

−0.076

0.732

0.927

Not reincarcerated at follow-up

0.93

0.92

0.235

0.352

1.265

N = 565

N = 470

No perpetration of violence

0.64

0.60

0.178

0.175

1.195

Complied with conditions of supervision

0.69

0.70

−0.025

0.881

0.975

No criminal behavior

0.64

0.59

0.207

0.112

1.230

No violent or weapons crimes

0.85

0.82

0.222

0.199

1.249

Not reincarcerated at follow-up

0.73

0.74

−0.067

0.641

0.935

N = 582

N = 531

No perpetration of violence

0.69

0.67

0.069

0.595

1.072

Complied with conditions of supervision

0.66

0.57

0.398

0.023

1.489

No criminal behavior

0.66

0.61

0.189

0.136

1.208

No violent or weapons crimes

0.84

0.83

0.073

0.654

1.076

Not reincarcerated at follow-up

0.64

0.66

−0.065

0.612

0.937

Wave 2

Wave 3

Wave 4

Note: Par. Est. = parameter estimate for weighted regression. p value = probability value of test statistic. OR = odds
ratio. N’s are the total responses for each wave of interviews and do not reflect any item missingness. Wave 1 = 30
days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months postrelease.

showed no evident differences between the two groups (data
not shown).
The second core measure of criminal behavior/recidivism was
compliance with conditions of supervision. This outcome is
important because the majority of men reported being under
post-release supervision throughout the follow-up period. As
shown in Exhibit 71, among the men who were under postrelease supervision, the results were mixed. There was no
difference in reports of compliance at 3 months—on average
78% of both groups reported complying with supervision
conditions. At 9 months, slightly fewer members of the SVORI
group reported complying (69% versus 70%). However, at 15
months, a significantly higher percentage of the SVORI group
reported that they had complied with the conditions of their
supervision (66% versus 57%).

128

Outcomes

Self-reported criminal behavior is another important dimension
of recidivism to capture because it includes criminal behavior
that may not have been detected (and that, therefore, is not
reflected in official measures of criminal activity). Two
outcomes reflecting self-reported criminal behavior are shown
in Exhibits and 72 and 73: any self-reported criminal behavior
(which includes violent crimes, carrying a weapon, other crimes
against people, drug possession crimes, drug sales crimes,
DWI/DUI, property crimes, and lesser types of crimes, such as
prostitution, soliciting, shoplifting, or disorderly conduct) and
self-reported involvement in violent or weapons offenses.
As shown in Exhibit 72, SVORI program participants were more
likely than non-SVORI comparison subjects to report
committing no crimes since release/last interview. This
difference is statistically significant at the 3-month interview
(p < 0.05), but not for subsequent follow-up periods.
Exhibit 72. Weighted
average reports of
committing no crimes
since release/last
interview

100%
80%

SVORI

79

Non-SVORI

73
64

60%

66
59

61

40%
20%
0%
Wave 2

Wave 3

Wave 4

Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 =
9 months post-release; Wave 4 = 15 months post-release.

Exhibit 73 shows the weighted average means of self-reports of
committing any violent crime or carrying a weapon. As can be
seen, about 9–10% of both groups reported committing a violent
crime and/or carrying a weapon in the 3 months following their
release from prison. In subsequent periods, greater numbers
reported either committing a violent crime or carrying a weapon
in the average of 6 months since the previous interview. None of
the differences was statistically significant (the p value for the
Wave 3 SVORI to non-SVORI difference was 0.2).

129

Prisoner reentry experiences of adult males

Exhibit 73. Weighted
average reports of
committing no violent
crimes/no weapons
since release/last
interview

100%

SVORI
Non-SVORI

90

91

85

82

84

83

80%
60%
40%
20%
0%
Wave 2

Wave 3

Wave 4

Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 =
9 months post-release; Wave 4 = 15 months post-release.

The final core criminal behavior/recidivism outcome based on
self-reported data is whether the respondent was not
reincarcerated at the time of his follow-up interview. Exhibit 74
shows the weighted average means and suggests a high
reincarceration rate for these serious and violent offenders
(most of whom were on supervision at release). Somewhat
more than 90% had not been reincarcerated within 3 months of
release—a percentage that declined to less than two-thirds by
the 15 month interview. These percentages imply that by the
time of the 9-month interview more than a quarter had been
reincarcerated and, by the 15-month interview, more than one
third had been reincarcerated.
Exhibit 75 presents an even more sobering picture. This exhibit
shows a composite measure reflecting whether the respondent
was not reincarcerated at the time of the follow-up interview
and reported that he had not been booked into jail or prison
(for 24 hours or more) during the reference period. This
measure is obviously more inclusive because it reflects any
(self-reported) incarceration during the reference period—not
just the point at which the interview was conducted. This
exhibit shows that there is no significant difference on this
outcome between SVORI and comparison men for any follow-up
period. In other words, SVORI and comparison men appeared
equally likely to have been reincarcerated during the follow-up
periods. What is perhaps most surprising, however, are the

130

Outcomes

Exhibit 74. Weighted
average reports of not
reincarcerated at
interview

100%

93

SVORI

92

80%

Non-SVORI

73

74
64

66

60%
40%
20%
0%
Wave 2

Wave 3

Wave 4

Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 =
9 months post-release; Wave 4 = 15 months post-release.

Exhibit 75. Weighted
average reports of not
reincarcerated at
interview and no
jail/prison stay of more
than 24 hours since
release/interview

100%

SVORI

82

Non-SVORI

81

80%
59

60%

59
51

49

40%
20%
0%
Wave 2

Wave 3

Wave 4

Note: Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 =
9 months post-release; Wave 4 = 15 months post-release.

percentages of individuals indicating that they had not been
reincarcerated. Only about 80% were NOT reincarcerated
between release and the 3-month interview, a percentage that
drops to about 60% at 9 months, and 50% at 15 months.
The remaining set of criminal recidivism measures were based
on official data sources and therefore reflect criminal behavior
detected by authorities. These measures include both rearrest
(obtained from the National Crime Information Center [NCIC],
as described in Lattimore and Steffey (2009) and reincarceration
in state prisons (obtained from the state Departments of

131

Prisoner reentry experiences of adult males

Corrections). The rearrest data reflect all data reported to the
NCIC and, thus, may include arrests in states other than the
states in which the evaluation subjects were identified. In
contrast, the reincarceration data reflect only reincarceration
into the same state system (e.g., if one of the SC subjects was
incarcerated in NC or GA, that incarceration is not captured in
the data). Unlike the self-report measures, data were obtainable
for almost all subjects. As with the self-report measures, the
reported means are weighted using the propensity scores, and
the parameter estimates are for the SVORI indicator variable in
the weighted logistic regression models. These core recidivism
measures based on official records are shown in Exhibit 76.
The findings for rearrest are shown graphically in Exhibit 77.
The findings suggest (1) members of the SVORI group were
less likely to be rearrested across the 24 months following
release than the non-SVORI group, but the differences were not
statistically significant; and (2) rearrest rates for these serious
and violent offenders were quite high, with about 70% having
had at least one new arrest within 24 months of release.
The SVORI program participants were less likely than the nonSVORI comparison subjects to have a new arrest for all crime
types (i.e., the slightly lower rearrest rate did not obscure
higher arrest rates for certain types of crimes). The difference
between SVORI and non-SVORI was significant at the 0.05
level for other crimes at 21 and 24 months and at the 0.10
level for public order crimes at 21 and 24 months, including
arrest for a parole or probation violation.
The findings for reincarceration (shown in Exhibit 78) indicate
that the SVORI and non-SVORI groups were equally likely to be
reincarcerated throughout the 24 month follow-up period—and
rates are actually higher for SVORI participants (albeit not
significantly so) after 3 months. These results are somewhat at
odds with both the self-report data and the arrest data, which
consistently, if weakly, suggest less criminal activity by the
SVORI participants.

132

Outcomes

Exhibit 76. Official measures of recidivism

SVORI
Mean
Rearrest
1st rearrest within 3 months of release
1st rearrest within 6 months of release
1st rearrest within 9 months of release
1st rearrest within 12 months of release
1st rearrest within 15 months of release
1st rearrest within 21 months of release
1st rearrest within 24 months of release
Rearrest within 21 months for violent crime
Rearrest within 21 months for property crime
Rearrest within 21 months for drug crime
Rearrest within 21 months for public order crime
Rearrest within 21 months for other crime
Rearrest within 24 months for violent crime
Rearrest within 24 months for property crime
Rearrest within 24 months for drug crime
Rearrest within 24 months for public order crime
Rearrest within 24 months for other crime
Reincarceration
1st reincarceration within 3 months of release
1st reincarceration within 6 months of release
1st reincarceration within 9 months of release
1st reincarceration within 12 months of release
1st reincarceration within 15 months of release
1st reincarceration within 21 months of release
1st reincarceration within 24 months of release

Non-SVORI
Par. Est.
Mean

SE

OR

16%
28%
41%
49%
55%
64%
68%
19%
23%
28%
41%
3%
20%
26%
30%
44%
3%

18%
32%
44%
51%
56%
66%
71%
21%
24%
30%
45%
6%
23%
27%
32%
49%
6%

−0.163
−0.191
−0.129
−0.089
−0.036
−0.089
−0.131
−0.112
−0.080
−0.118
−0.175
−0.560
−0.142
−0.054
−0.117
−0.189
−0.585

0.136
0.111
0.104
0.102
0.103
0.107
0.112
0.129
0.120
0.114
0.104
0.250
0.127
0.117
0.112
0.104
0.249

0.849
0.826
0.879
0.915
0.964
0.914
0.877
0.894
0.923
0.889
0.839
0.571
0.867
0.948
0.890
0.828
0.557

3%
11%
19%
25%
30%
39%
42%

4%
10%
17%
25%
29%
36%
39%

−0.163
0.062
0.113
−0.023
0.033
0.135
0.128

0.280
0.160
0.128
0.114
0.108
0.102
0.102

0.849
1.064
1.120
0.977
1.033
1.145
1.137

Note: Par. Est. = parameter estimate for weighted regression. OR = odds ratio. SE = standard error. For rearrest,
SVORI N = 806 for all periods except N = 787 for 24 month measures; non-SVORI N = 775 for all periods except N
= 759 for 24 month measures. For reincarceration, SVORI N = 863 for all periods; non-SVORI N = 834 except for N
= 817 for 24 month measure.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

133

Prisoner reentry experiences of adult males

Exhibit 77. Cumulative
rearrest rates by group

100%
SVORI

80%

Non-SVORI

64 66
60%
44

41
40%
20%

49 51

68 71

55 56

32

28
16 18

0%
3

6

9

12

15

21

24

Months since release

Exhibit 78. Cumulative
reincarceration rates by
group

100%

SVORI
Non-SVORI

80%
60%

39

40%
20%

11 10

19 17

25 25

30 29

36

42

39

4

3
0%
3

6

9
12
15
Months since release

21

24

TAKING A BROAD VIEW: OUTCOMES OVER
DOMAINS AND TIME
In this chapter, the findings suggest that for a variety of
specific outcomes in each domain men who participated in
SVORI programs had better outcomes, if only moderately so.
This section presents a broader view, looking across about 100
outcomes across the domains for each wave of data collection.
Odds ratios for 98 outcomes for the Wave 2 and Wave 4 data
and 91 outcomes for the Wave 3 data (the difference is due to
outcomes from drug tests conducted at Waves 2 and 4). In all

134

Outcomes

cases, the outcomes were coded such that a positive difference
meant the SVORI group performed better on that outcome
(e.g., arrested was reverse coded to no arrest).
Exhibit 79 shows the odds ratios for 98 dichotomous outcomes
measured 3 months following release. Most of these ratios are
greater than 1, signifying that SVORI participants had even or
higher odds of having these positive outcomes. Specifically, 75
of 98 odds ratios are 1 or more, and 11 of these are significant
at the 0.10 or better level. If SVORI were ineffective, roughly
the same number of positive as negative outcomes would be
expected for SVORI participants in comparison to non-SVORI
subjects. Instead, roughly three out of four outcomes are
neutral or better.
Exhibit 79. Odds ratios
from propensity score
weighted logistic
regressions of 98 Wave
2 (3-month) outcomes
as a function of SVORI
program participation

2.5
2.0
1.5
1.0
0.5
0.0

Exhibit 80 provides similar if somewhat weaker results for Wave
3 data. There were 93 9-month outcomes because drug tests
weren’t conducted in conjunction with these interviews. Here,
Of the 93 odds ratios, 65 are 1 or larger and 13 are significant
at the 0.10 level or better.
Exhibit 81 presents similar results for the 98 outcomes at Wave
4. For the 15-month data, 81 of the 99 outcomes have odds
ratios of 1 or larger, and 22 are significant at the 0.10 level or
better.

135

Prisoner reentry experiences of adult males

Exhibit 80. Odds ratios
from propensity score
weighted logistic
regressions of 93 Wave
3 (9-month) outcomes
as a function of SVORI
program participation

2.5
2.0
1.5
1.0
0.5
0.0

Exhibit 81. Odds ratios
from propensity score
weighted logistic
regressions of 98 Wave
4 (15 month) outcomes
as a function of SVORI
program participation

2.5
2.0
1.5
1.0
0.5
0.0

Exhibit 82 summarizes the findings presented in Exhibits 79
through 81. As noted earlier, if SVORI program participation
had no impact on outcomes, SVORI outcomes would be better
than non-SVORI outcomes about 50% of the time; in this case,
however, for each follow up, the percentage of odds ratios
greater than 1 exceeds 70% (specifically, 76% at 3 months,
70% at 9 months, and 82% at 15 months). The probability of
these outcomes if SVORI were not effective (i.e., if the odds of
a better outcome for SVORI over non-SVORI was 50:50) is
effectively zero (binomial distribution).

136

Outcomes

Exhibit 82. Summary
results of odds ratios
from propensity score
weighted logistic
regressions of three
waves of outcomes as a
function of SVORI
program participation

Number of outcomes
SVORI better than non-SVORI (OR > 1)
SVORI better than non-SVORI (p < 0.05)
SVORI better than non-SVORI (p < 0.1)
non-SVORI better than SVORI (OR > 1)
non-SVORI better than SVORI (p < 0.05)
non-SVORI better than SVORI (p < 0.1)

Follow-up
3
9
15
Months Months Months
98
93
98
75
65
80
4
10
16
11
13
21
23
3
0

28
1
4

18
1
2

Note: OR = odds ratio.

137

Conclusions, Policy
Implications, and
Future Work
In 2003, the U.S. DOJ, DOL, DOEd, HUD, and HHS provided
more than $100,000,000 in grant funds to states to develop,
enhance, or expand programs to facilitate the reentry of adult
and juvenile offenders returning to communities from prisons or
juvenile detention facilities. The $500,000 to $2,000,000 3-year
grants were used to establish programs that were to span the
periods before release, in the community on supervision, and
post-supervision. In addition to the funding, SVORI encouraged
agencies to coordinate with correctional and community
partners and services; these activities were mentioned by
SVORI program directors as a significant—and, for most,
sustained—change from business as usual. The initiative
responded to emerging research findings that suggested that
providing individuals with comprehensive, coordinated services
based on needs and risk assessments could result in improved
post-release outcomes.
The multi-site evaluation was designed to determine whether
participation in SVORI programs resulted in increased service
receipt and better post-release outcomes. The findings in this
report provided information on the characteristics and
experiences of adult male SVORI program participants and
comparison subjects in 12 states who were released from
prison between July 2004 and November 2005. Many of the
services were intended to improve intermediate outcomes that
have been correlated with recidivism—for example,
employment services to improve employment, substance abuse
treatment to reduce use, and cognitive programs to address

139

Prisoner reentry experiences of adult males

criminal thinking. The underlying model suggests that
improvements in these outcomes will lead to reductions in
criminal behavior. The SVORI program participants (and, to a
lesser extent, the non-SVORI respondents) received a variety
of different services, each of which could impact one or more
intermediate outcomes that could impact recidivism. There is
little theoretical or empirical guidance for the correct
specification of such a complex recidivism model and, thus, the
approach to the outcome analyses was to test first-order effects
of SVORI program participation on each of the identified
outcomes including recidivism.
The findings substantiate previous research that male prisoners
returning to their communities after serving more than 2 years
in prison comprise a population with extremely high needs and
that their expressed needs remained high (if somewhat
diminished from pre-release) up to 15 months following release
from prison. Overall, the men in the study had weak
educational and employment histories, extensive substance
abuse histories, substantial experience with the criminal justice
system, and extensive exposure to drug or criminally involved
family members and peers. In particular, most had used drugs
in the past—including two thirds who reported using in the 30
days prior to their incarceration. Most had been previously
incarcerated. A majority had been treated for mental health or
substance use problems.

SVORI program
participation increased
the likelihood of receiving
a wide range of services,
but levels of receipt were
generally much less than
reported needs.

Results from the impact study suggest that SVORI programs
were successful in significantly increasing access to a variety of
services and programming—particularly services related to
transitioning to the community and employment/education/
skills, as well as to substance abuse treatment. For example,
75% of SVORI program participants, in contrast to 51% of nonSVORI comparison subjects, reported involvement in programs
while in prison to prepare for release (see Exhibit 52). 39 This
approximately 50% increase in the likelihood of reentry
program participation that was observed for the SVORI
program participants was seen more broadly in the super
bundle scores, which summarized service receipt at each of the
four interviews. SVORI program participants reported service
receipt while in prison across all domains, resulting in a super
39

140

Values are propensity score weighted means unless otherwise
noted.

Conclusions, Policy Implications, and Future Work

bundle score that was 34–55% higher than the 22 score
calculated for non-SVORI respondents (see Exhibit 46). These
findings add to emerging research regarding the feasibility of
improving service receipt across broadly conceived reentry
programs.

… on average,
respondents reported
needing more than half of
all services (54%)—much
greater than the receipt of
34%, on average, for
SVORI program
participants and 22% for
non-SVORI respondents.

Although the SVORI programs were successful in increasing the
types and amounts of needs-related services provided prior to
and after release from prison, the proportion of individuals who
reported receiving services was less than reported need
(sometimes much less) and, generally, less than the
expectations of the SVORI program directors. Thus,
respondents in both groups reported needs that generated
super bundle service need scores of 54 at the time of the prerelease interview (see Exhibit 37). The super bundle scores
calculated from program director survey data showed that the
program directors expected that their programs on average
were providing services that resulted in an average pre-release
bundle score of 52 and 36 for SVORI and non-SVORI,
respectively (Winterfield et al., 2006). 40 These findings suggest
that, on average, respondents reported needing more than half
of all services (54%)—much greater than the receipt of 34%,
on average, for SVORI program participants and 22% for nonSVORI respondents. Similarly, in a 2005 survey, program
directors reported that on average about 52% of SVORI
program participants were expected to receive each service
(compared to 36% for non-SVORI). Thus, the expressed needs
of the participants and the expectations for SVORI program
service delivery by the program directors both exceeded the
reported levels of service receipt. This finding is consistent with
the fact that SVORI programs were still developing and
implementing their programs and provides a substantial
reminder that starting up complex programs may require a
sustained effort over several years to reach full
implementation.
Service delivery declined substantially, on average, following
release. Thus, overall, the programs were unable to sustain
40

The components and calculations of bundle scores for the program
director data are analogous to, but differ somewhat from, the
components and calculations for the offender interview data. One
interpretation of the program director scores is that it identified the
average percentage of offenders who received each service included
in the bundle (in this case all services). See Winterfield et al.
(2006) for details.

141

Prisoner reentry experiences of adult males

providing support to individuals during the critical, high risk
period immediately following release. This decline may be
attributable to the difficulty programs experienced early on in
their efforts to identify and coordinate services for individuals
released across wide geographic areas and, again, suggests the
need for sustained effort to reach full implementation.

SVORI programs were
unable to sustain levels
of service provision to
respondents with high
levels of expressed
needs following release.

The level of services received diminished quickly over time
following release, regardless of expressed need. For example,
85% or more of both groups reported at all interviews that they
needed more education. While 51% of SVORI program
participants (44% of non-SVORI) reported receiving educational
services while in prison, this percentage dropped to 11% for
SVORI (8% for non-SVORI subjects) during the 3 months
following release and remained essentially unchanged over the
next 12 months (10% SVORI and 7% non-SVORI at the 15month interview). Similar results were seen overall. The service
need super bundle scores ranged between 42 and 45 for the
two groups over the three follow-up interviews—relatively
stable following about a 10 point drop from the 54 score
estimated from pre-release data. As shown in Exhibit 46,
reported service receipt scores declined for SVORI participants
from 34 at the pre-release interview to 18 at 3 months postrelease, 13 at 9 months, and 9 at 15 months post-release.
SVORI participants’ scores were significantly higher than those
for non-SVORI respondents (22 pre-release and 12, 10, and 8
in the three post-release waves) through 9 months postrelease. However, these scores demonstrated that the
programs were unable to sustain levels of service provision to
respondents with high levels of expressed needs following
release.
As previous research suggests the importance of after care to
successful reintegration, the failure of the programs to provide
sustained support during the critical, high-risk period
immediately following release may have contributed to the
modest impact findings. The failure to provide substantial levels
of services following release, however, may also point to the
difficulty of implementing broadly based reentry programs to
provide services across a wide range of domains—a difficulty
exacerbated for programs that released SVORI program
participants across multiple geographic areas or even
statewide. These considerations suggest that ample time should
be provided for development and implementation and that

142

Conclusions, Policy Implications, and Future Work

there may be a need for a sustained, multi-year effort to reach
full implementation.

SVORI program
participation was
associated with
moderately better
outcomes in housing,
employment, substance
use, and self-reported
criminal behavior.

SVORI program
participants had lower
arrest rates but the
differences were not
statistically significant.

The significant—albeit less-than-universal—increase in service
receipt associated with participation in SVORI programs was
associated with moderately better outcomes with respect to
housing, employment, substance use, and self-reported
criminal behavior. For example, SVORI program participants
were more likely to report living in their own house or
apartment at each interview. Further, although SVORI and nonSVORI respondents were equally likely to report having worked
since the last interview, SVORI program participants were more
likely to report currently supporting themselves with a job at 3
and 15 months post-release and to report that they were
working at a job that offered formal pay and provided benefits
(health insurance or paid leave). Additionally, SVORI program
participants were more likely to have abstained from drug use—
with more reporting no drug use since the last interview or in
the last 30 days. Similar results were found when self-report
measures were combined with drug test results. However,
overall, abstinence from illegal drugs was disappointing. For
example, at 15 months following release, only 46% of SVORI
program participants and 43% of non-SVORI comparison
subjects reported no drug use in the previous 30 days and
tested negative on an oral swab drug test. As many of the
previous evaluations of reentry programs have focused
primarily on recidivism and substance use, this evaluation has
extended knowledge about the potential impacts of reentry
programming on an array of other important indicators of
successful reintegration, including housing and employment.
The recidivism results were mixed—with SVORI program
participants less likely to report criminal activity (significantly
so at 3 months post-release, when 79% of SVORI and 73% of
non-SVORI reported no crimes since release). The differences
remained about 6 percentage points over successive time
points (64% versus 59% at 9 months and 66% versus 61% at
15 months), implying that about 10% more SVORI program
participants reported no crimes during the previous 6 months
than non-SVORI comparisons. SVORI program participants
were also less likely to have an officially recorded arrest at a
variety of points during the 24-month period following release,
although these differences were small and not significant.
Rearrest rates, overall, were high for both groups—by 24

143

Prisoner reentry experiences of adult males

months following release, 68% of the SVORI program
participants and 71% of the non-SVORI comparisons had a new
arrest recorded at NCIC.

There was little
difference in
reincarceration rates,
although by 21 months
post-release 39% of
SVORI program
participants versus 36%
of non-SVORI
comparison subjects
had been
reincarcerated.

Although self-reported criminal behavior and official arrest
records were consistent in supporting somewhat lower criminal
activity among SVORI program participants, this was not
associated with lower reincarceration rates—and, in fact, by
21 months post-release the reincarceration rate for SVORI
program participants was about 10% higher than the nonSVORI rate (39% versus 36%). The reincarceration findings
would be consistent with self-reported compliance with
conditions of supervision, which was similar for SVORI and nonSVORI subjects at 3 and 9 months post-release in that
noncompliance could lead to revocation. However, SVORI
program participants were more likely to report complying with
conditions of supervision at 15 months (66% versus 57%).
Additional investigation is needed to determine whether
supervision is in some way associated with the reincarceration
findings.
From a policy perspective, the multi-site SVORI evaluation adds
to the sparse reentry evaluation literature that addresses the
effect of broad-based (wraparound) programmatic efforts on
high risk individuals. Specifically, much of the reentry literature
to date presents findings from single-focus interventions, such
as drug treatment or cognitive behavior therapy, which have
been implemented with low-risk offenders. SVORI was initiated
as consensus began to build that programs needed to address
the multiplicity of needs of offenders and that interventions
were likely to be more successful when focused on high-risk (or
higher risk) offenders. The scale of the evaluation—including
programs in 12 states and enrolling nearly 1,700 men—
provided an opportunity to develop a comprehensive portrait of
high-risk individuals as they attempted to reintegrate into
communities following prison release, providing insight into
their post-release circumstances.
The gaps between reported service needs and receipt for SVORI
program participants may be attributable to the early stage of
implementation when these subjects were enrolled in the
evaluation. Further, most of the SVORI impact programs were
deployed in multiple prisons and enrolled participants who
returned to multiple communities. Developing and

144

Conclusions, Policy Implications, and Future Work

implementing the panoply of services for a comprehensive
reentry program within multiple prisons and identifying and
enlisting community programs and resources are complex tasks
that could easily take several years to fully realize. Thus, for
example, although “only” 57% of SVORI participants reported
having a reentry plan 30 days prior to release, this is a 138%
increase over the percentage of non-SVORI respondents who
reported having a plan. The 57% finding suggests an
opportunity for continued program improvement and more
complete implementation. Indeed, many states appear to have
viewed their program development and implementation with
SVORI funds as a foundation upon which to build better
programs—by enhancing services and expanding the reach of
the services. As reported in Winterfield, Lindquist, and
Brumbaugh (2007), most SVORI program directors said in
response to a 2006 survey that their states were continuing to
build on the programs that they established with SVORI grant
funds.
Importantly, service delivery was not sustained during the
critical, high risk period immediately following release. The
treatment literature suggests that 90 to 270 days of continued
care is optimal (Friedmann, Taxman, & Henderson, 2007;
Taxman, Perdoni, & Harrison, 2007). Larger program effects
would be expected with continuous service delivery after
release—which was one of the primary features of the SVORI.
The modest improvements in intermediate outcomes observed
in the evaluation of SVORI are consistent with findings from
several meta-analyses of single-program efforts. These
analyses suggest treatment effects from 10% to 20% across a
wide range of types of programming for offenders (e.g., Aos et
al., 2006). Whether a multi-focus reentry program can lead to
significantly greater treatment effects of 30% to 50% is
unknown. Results from the SVORI evaluation suggest that
programs will need to be given sufficient time to implement
multi-component, multi-phase programs before this hypothesis
can be tested.
The evaluation was designed to address the question of
whether SVORI programs—enhanced reentry programs—could
impact the post-release outcomes of high-risk offenders. In
other words, the goal was to answer the question “Did SVORI
work?” SVORI programs were “black boxes” that under the

145

Prisoner reentry experiences of adult males

SVORI model were assumed to contain the need-based services
appropriate for each individual. Thus, although the programs
differed across sites, all programs conformed to this higherorder definition of program. Deficiencies in service delivery are,
thus, ascribed to development or implementation shortcomings.
Of course, this is not wholly satisfactory. Indeed, the extensive
data collection on service receipt was intended to allow
examination of other evaluation questions that were beyond the
scope of this evaluation.
Additional evaluation questions will be addressed in future
work, answers to which may help guide policy. This research
will address “what worked?” (Were some SVORI programs more
successful than others? Can the effects of specific program
components be disentangled?); “for whom?” (Are there
identifiable characteristics that are associated with better
outcomes?); “for how long?” (How are study participants faring
5 years after release from prison?); and “at what cost?” (Are
there long-term cost savings associated with the SVORI
programs?). For example, one related hypothesis to be tested
involves the question of whether services directed at proximal
outcomes (e.g., substance abuse treatment) are sufficient to
effect changes in criminal behavior or whether programs
targeted directly at changing criminal thinking may be needed
as well. Additional examination of site-level differences is also
needed to determine whether the characteristics of the sites
(e.g., parole revocation policies, economic climate) have an
impact on “what works” independent of local program and
participant characteristics. The extensive SVORI dataset
provides an opportunity for future research to explore these
questions, as well as related questions such as which services
were helpful, what factors led to reincarceration, and what
factors were associated with remaining out of prison.

146

References
Adams, K., Bennett, K. J., Flanagan, T. J., Marquart, J. W.,
Cuvelier, S. J., Fritsch, E., et al. (1994). A large-scale
multidimensional test of the effect of prison education
programs on offenders' behavior. Prison Journal, 74(4),
443-449.
Andrews, D. A. (2006). Enhancing adherence to risk-needresponsivity: Making quality a matter of policy.
Criminology and Public Policy, 5, 595-602.
Andrews, D. A., & Bonta, J. (2006). The psychology of criminal
conduct (4th ed.). Newark, NJ: Lexis/Nexis/Matthew
Bender.
Andrews, D. A., Ziner, I., Hoge, R. D., Bonta, J., Gendreau, P.,
& Cullen, F. T. (1990). Does correctional treatment
work: A clinically-relevant and psychologically-informed
meta-analysis. Criminology, 28, 369-404.
Aos, S. (2006). Options to stabilize prison populations in
Washington--Interim report. Steve Aos. #06-01-1202.
Aos, S., Miller, M., & Drake, E. (2006). Evidence-based adult
corrections programs: What works and what does not.
Olympia, WA: Washington State Institute for Public
Policy.
Beck, A., Gilliard, D., Greenfeld, L., Harlow, C., Hester, T.,
Jankowski, L., et al. (1993). Survey of state prison
inmates, 1991 (NCJ 136949). Washington, DC: U.S.
Bureau of Justice Statistics.
Bonczar, T. P., & Beck, A. J. (1997). Lifetime likelihood of going
to state or federal prison. Bureau of Justice Statistics
Special Report No. NCJ 1600092. Washington, DC: U.S.
Bureau of Justice Statistics.
Boudin, K. (1993). Participatory literacy education behind bars.
Harvard Educational Review, 63(2), 207-232.

147

Prisoner reentry experiences of adult males

Buck, M. L. (2000). Getting back to work: Employment
programs for ex-offenders. Field report series.
Philadelphia: Public/Private Ventures.
Bushway, S., & Reuter, P. (2002). Labor markets and crime. In
J. Q. Wilson & J. Petersilia (Eds.), Crime: Public policies
for crime control (Rev. ed., pp. 191-224). San Francisco,
CA: ICS Press.
Byrnes, M., Macallair, D., & Shorter, A. (2008). Aftercare as
afterthought: Reentry and the California Youth
Authority. In R. G. Shelden & D. Macallair (Eds.),
Juvenile justice in America: Problems and prospects (pp.
83-114). Long Grove, IL: Waveland Press.
Chalfin, A., Tereshchenko, B., Roman, J., Roman, C., & Arriola,
C. (2007). Cost-benefit analysis of building bridges.
Washington, DC: Urban Institute.
Clear, T. R., Rose, D. R., & Ryder, J. A. (2001). Incarceration
and the community: The problem of removing and
returning offenders. Crime & Delinquency, 47, 335-351.
Cowell, A., Roman, J., & Lattimore, P. K. (2009). An economic
evaluation of the Serious and Violent Offender Reentry
Initiative. Research Triangle Park, NC: RTI International.
Cullen, F. T., & Gendreau, P. (2000). Assessing correctional
rehabilitation: Policy, practice, and prospects. In J.
Horney (Ed.), Criminal justice 2000: Policies, processes,
and decisions of the criminal justice system (Vol. 3, pp.
109-175). Washington, DC: National Institute of Justice.
D'Agostino, R. B., Jr. (1998). Tutorial in biostatistics:
Propensity score methods for bias reduction in the
comparison of a treatment to a non-randomized control
group. Statistics in Medicine, 17, 2265-2281.
Finn, P. (1998). Successful job placement for ex-offenders: The
center for employment opportunities. Washington, DC:
National Institute of Justice.
Franz, J. (2003). Wraparound: A primer. Retrieved December
23, 2009, from http://www.paperboat.com/.
Friedmann, P. D., Taxman, F. S., & Henderson, C. (2007).
Evidence-based treatment practices for drug-involved
adults in the criminal justice system. Journal of
Substance Abuse Treatment, 32(3), 267-277.

148

References

Gaes, G. G., Flanagan, T. J., Motiuk, L. L., & Stewart, L.
(1999). Adult correctional treatment. In M. Tonry & J.
Petersilia (Eds.), Prisons (pp. 361-426). Chicago:
University of Chicago Press.
Gerber, J., & Fritsch, E. (1994). The effects of academic and
vocational program participation on inmate misconduct
and reincarceration: Prison education research project:
Final report. Huntsville, TX: Sam Houston State
University.
Glueck, S., & Glueck, E. (1974). Of delinquency and crime.
Springfield, IL: Thomas.
Glynn, R. J., Schneeweiss, S., & Sturmer, T. (2006). Indications
for propensity scores and review of their use in
pharmacoepidemiology. Basic Clinical Pharmacology and
Toxicology, 98(3), 253-259.
Gottfredson, M. R., & Hirschi, T. (1990). A general theory of
crime. Stanford, CA: Stanford University Press.
Greifinger, R. B., Bick, J. A., & Goldenson, J. (2007). Public
health behind bars: From prisons to communities. Dobbs
Ferry, NY: Springer-Verlag.
Hagan, J., & Coleman, J. P. (2001). Returning captives of the
American war on drugs: Issues of community and family
reentry. Crime & Delinquency, 47(3), 352-267.
Hagan, J., & Dinovitzer, R. (1999). Collateral consequences of
imprisonment for children, communities, and prisoners.
In J. Petersilia & M. Tonry (Eds.), Prisons (pp. 121-162).
Chicago: University of Chicago Press.
Harer, M. D. (1995). Prison education program participation
and recidivism: A test of the normalization hypothesis.
Washington, DC: Federal Bureau of Prisons Office of
Research and Evaluation.
Harrison, P. (2000). Total sentenced prisoners released from
state or federal jurisdiction, National Prisoners Data
Series (NPS-1). Washington, DC: U.S. Bureau of Justice
Statistics. Available at:
http://www.ojp.usdoj.gov/bjs/data/corpop22.csv.
Hawkins, S., Dawes, D., Lattimore, P. K., & Visher, C. A.
(2009). Detainee reentry experiences of juvenile males:
Characteristics, service receipt, and outcomes of
participants in the SVORI Multi-site Evaluation. Research
Triangle Park, NC: RTI International.

149

Prisoner reentry experiences of adult males

Healy, K., Foley, D., & Walsh, K. (1999). Parents in prison and
their families. Queensland, Australia: Catholic Prison
Ministry.
Holsinger, A. M., Lurigio, A. J., & Latessa, E. I. (2001).
Practitioner's guide to understanding the basis of
assessing offender risk. Federal Probation, 65(1), 46-50.
Holzer, H. J., & Martinson, K. (2005). Can we improve job
retention and advancement among low-income working
parents? Washington, DC: Urban Institute.
Holzer, H. J., Raphael, S., & Stoll, M. A. (2006). Perceived
criminality, criminal background checks, and the racial
hiring practices of employers. Journal of Law &
Economics, 49(2), 451-480.
James, D. J., & Glaze, L. E. (2006). Mental health problems of
prison and jail inmates. Washington, DC: U.S. Bureau of
Justice Statistics.
Karberg, J. C., & James, D. J. (2005). Substance dependence,
abuse, and treatment of jail inmates 2002. Washington,
DC: U.S. Department of Justice.
Lattimore, P. K. (2007). The challenges of reentry. Corrections
Today, 69(2), 88-91.
Lattimore, P. K., Brumbaugh, S., Visher, C., Lindquist, C. H.,
Winterfield, L., Salas, M., et al. (2004). National portrait
of the Serious and Violent Offender Reentry Initiative.
Research Triangle Park, NC: RTI International.
Lattimore, P. K., & Steffey, D. M. (2009). The Multi-Site
Evaluation of SVORI: Methodology and analytic
approach. Research Triangle Park, NC: RTI
International.
Lattimore, P. K., Steffey, D. M., & Visher, C. A. (2009). Prisoner
reentry experiences of adult males: Characteristics,
service receipt, and outcomes of participants in the
SVORI Multi-site Evaluation. Research Triangle Park, NC:
RTI International.
Lattimore, P. K., & Visher, C. A. (2009). The Multi-site
Evaluation of SVORI: Summary and synthesis. Research
Triangle Park: RTI International.
Lattimore, P. K., Visher, C. A., & Steffey, D. M. (2008). Prerelease characteristics and service receipt among adult
male participants in the SVORI Multi-site Evaluation.
Research Triangle Park, NC: RTI International.

150

References

Lattimore, P. K., Visher, C. A., Winterfield, L., Lindquist, C., &
Brumbaugh, S. (2005). Implementation of prisoner
reentry programs: Findings from the Serious and Violent
Offender Reentry Initiative Multi-site Evaluation. Justice
Research and Policy, 7(2), 87-109.
Laub, J. H., Nagin, D. S., & Sampson, R. J. (1998). Trajectories
of change in criminal offending: Good marriages and the
desistance process. American Sociological Review,
63(2), 225-238.
Laub, J. H., & Sampson, R. J. (1993). Turning points in the life
course: Why change matters to the study of crime.
Criminology, 31, 301-326.
Laub, J. H., & Sampson, R. J. (2001). Understanding desistance
from crime. In M. H. Tonry (Ed.), Crime and Justice: A
Review of Research. Chicago: University of Chicago
Press.
Laub, J. H., & Sampson, R. J. (2003). Shared beginnings,
divergent lives: Delinquent boys to age 70. Cambridge,
MA: Harvard University Press.
Lawrence, S., Mears, D. P., Dubin, G., & Travis, J. (2002). The
practice and promise of prison programming.
Washington, DC: Urban Institute.
Lindquist, C. (2005). Reentry research in action:
Implementation of SVORI programs. Research Triangle
Park, NC: RTI International.
Lindquist, C., & Winterfield, L. (2005). Reentry research in
action: Characteristics of prisoner reentry programs.
Research Triangle Park, NC: RTI International.
Lindquist, C. H., Barrick, K., Lattimore, P. K., & Visher, C. A.
(2009). Prisoner reentry experiences of adult females:
Characteristics, service receipt, and outcomes of
participants in the SVORI Multi-site Evaluation. Research
Triangle Park, NC: RTI International.
Lipsey, M. W., & Cullen, F. T. (2007). The effectiveness of
correctional rehabilitation: A review of systematic
reviews. Annual Review of Law and Social Science, 3,
297-320.
Lowenkamp, C. T., Latessa, E. J., & Smith, P. (2006). Does
correctional program quality really matter? The impact
of adhering to the principles of effective intervention.
Criminology and Public Policy, 5, 575-608.

151

Prisoner reentry experiences of adult males

Lynch, J., & Sabol, W. (2001). Prisoner reentry in perspective.
Washington, DC: Urban Institute.
MacKenzie, D. L. (2006). What works in corrections: Reducing
the criminal activities of offenders and delinquents. New
York: Cambridge University Press.
Marlowe, D. B. (2006). When "what works" never did: Dodging
the "scarlet M" in correctional rehabilitation. Criminology
and Public Policy, 5(2), 339-346.
Maruna, S. (2001). Making good: How ex-offenders reform and
reclaim their lives. Washington, DC: American
Psychological Association Books.
Mauer, M. (2000). Race to incarcerate. Washington, DC: The
Prison Project.
Moore, G. E., & Mears, D. P. (2003). A meeting of the minds:
Researchers and practitioners discuss key issues in
corrections-based drug treatment. Washington, DC:
Urban Institute.
National Research Council (2007). Parole, desistance from
crime, and community integration. Washington, DC:
National Academy Press.
Petersilia, J. (2003). When prisoners come home: Parole and
prisoner reentry. New York: Oxford University Press.
Petersilia, J. (2004). What works in prisoner reentry? Reviewing
and questioning the evidence. Federal Probation, 68(2),
4-8.
Petersilia, J. (2005). Hard time: Ex-offenders returning home
after prison. Corrections Today, 155, 66-71.
Pew Center on the States (2009). Maximum impact: Targeting
supervision on higher-risk people, places, and times.
Public Safety Policy Brief. Retrieved December 28, 2009,
from http://www.pewcenteronthestates.org/uploadedFiles/
Maximum_Impact_web.pdf

Piehl, A. M. (1998). Economic conditions, work, and crime. In
M. Tonry (Ed.), Handbook on crime and punishment (pp.
302-319). Oxford, England: Oxford University Press.
Re-entry Policy Council (2005). Report of the Re-entry Policy
Council: Charting the safe and successful return of
prisoners to the community. New York: Council of State
Governments.

152

References

Rhine, E., Mawhorr, T., & Parks, E. C. (2006). Implementation:
The bane of effective correctional programs. Criminology
and Public Policy, 5(2), 347-358.
Roman, J., Brooks, L. E., Lagerson, E., Chalfin, A., &
Tereshchenko, B. (2007). Impact and cost-benefit
analysis of the Maryland reentry partnership initiative.
Washington, DC: Urban Institute.
Rosenbaum, P., & Rubin, D. B. (1983). The central role of the
propensity score in observational studies for causal
effects. Biometrika, 70, 41-55.
Rosenbaum, P., & Rubin, D. B. (1985). The bias due to
incomplete matching. Biometrics, 41, 103-116.
Rossman, S. B., & Roman, C. G. (2003). Case-managed reentry
and employment: Lessons from the Opportunity to
Succeed Program. Justice Research and Policy, 5(2), 75100.
Rubin, D. B. (1997). Estimating causal effects from large data
sets using propensity scores. Annals of Internal
Medicine, 127, 757-763.
Rubin, D. B. (2006). Matched sampling for causal effects. New
York: Cambridge University Press.
Sampson, R. J., & Laub, J. H. (1990). Crime and deviance over
the life course: The salience of adult social bonds.
American Sociological Review, 55(5), 609-627.
Sampson, R. J., & Laub, J. H. (1993). Crime in the making:
Pathways and turning points through life. Cambridge,
MA: Harvard University Press.
Seiter, R. P., & Kadela, K. R. (2003). Prisoner reentry: What
works, what does not, and what is promising. Crime &
Delinquency, 49(3), 360-388.
Stillman, J. (1999). Working to learn: Skills development under
"Work First." Philadelphia, PA: Public/Private Ventures.
Strategic Advantages (2000). Symptom Assessment-45
Questionnaire (SA-45). New York: Multi-Health Systems.
Sung, G. (2001). Rehabilitation felony drug offenders through
job development: A look into a prosecutor-led diversion
program. Prison Journal, 81(2), 271-286.
Taxman, F. S., Perdoni, M., & Harrison, L. (2007). Treatment
for adult offenders: A review of the state of the state.
Journal of Substance Abuse Treatment, 32(3), 239-254.

153

Prisoner reentry experiences of adult males

Travis, J. (2002). Invisible punishment: An instrument of social
exclusion. In M. Mauer & M. Chesney-Lind (Eds.),
Invisible punishment: The collateral consequence of
mass imprisonment (pp. 15-36). New York: New Press.
Travis, J. (2005). But they all come back: Facing the challenges
of prisoner reentry. Washington, DC: Urban Institute
Press.
Visher, C. A. (2006). Effective reentry programs [Editorial
introduction]. Criminology and Public Policy, 5(2), 299302.
Visher, C. A. (2007). Returning home: Emerging findings and
policy lessons about prisoner reentry. Federal
Sentencing Reporter, 20(2), 93-102.
Visher, C. A., Debus, S., & Yahner, J. (2008). Employment after
prison: A longitudinal study of releasees in three states.
Washington, DC: Urban Institute.
Visher, C. A., & Mallik-Kane, K. (2007). Reentry experiences of
men with health problems. In R. Greifinger (Ed.), Public
health is public safety: Improving public health through
correctional health care (pp. 432-447). London:
Spinger-Verlag.
Visher, C. A., & Travis, J. (2003). Transitions from prison to
community: Understanding individual pathways. Annual
Review of Sociology, 29, 89-113.
Visher, C. A., Winterfield, L., & Coggeshall, M. B. (2005). Exoffender employment programs and recidivism: A metaanalysis. Journal of Experimental Criminology, 1(3),
295-316.
Ware, J. E., Jr., Kosinski, M., Turner-Bowker, D. M., & Gandek,
B. (2002). How to score version 2 of the SF-12 health
survey (with a supplement documenting version 1).
Lincoln, RI: QualityMetric.
Warr, M. (1998). Life-course transitions and desistance from
crime. Criminology, 36, 183–216.
West, H. C., Sabol, W. J., & Cooper, M. (2009). Prisoners in
2008 (NCJ-228417). U.S. Bureau of Justice Statistics.
Retrieved December 27, 2009, from
http://bjs.ojp.usdoj.gov/index.cfm?ty=pbdetail&iid=1763.

154

References

Western, B. (2007). The penal system and the labor market. In
S. Bushway, M. A. Stoll, & D. F. Weiman (Eds.), Barriers
to re-entry? The labor market for released prisoners in
postindustrial America. New York: Russell Sage
Foundation.
Wiebush, R. G., McNulty, B., & Le, T. (2000). Implementation
of the intensive community-based aftercare program.
Washington, DC: Office of Juvenile Justice and
Delinquency Prevention.
Wilson, D. B., Gallagher, C. A., & MacKenzie, D. L. (2000). A
meta-analysis of corrections-based education, vocation,
and work programs for adult offenders. Journal of
Research on Crime and Delinquency, 37(4), 347-368.
Wilson, J. A., & Davis, R. C. (2006). Good intentions meet hard
realities: An evaluation of the project greenlight reentry
program. Criminology and Public Policy, 5(2), 303-338.
Winterfield, L., & Brumbaugh, S. (2005). Reentry research in
action: Overview of the Serious and Violent Offender
Reentry Initiative. Research Triangle Park, NC: RTI
International.
Winterfield, L., & Castro, J. (2005). Matching drug treatment to
those in need: An analysis of correctional service
delivery in Illinois and Ohio. Justice Research and Policy,
7(2), 30-55.
Winterfield, L., Lattimore, P. K., Steffey, D. M., Brumbaugh, S.,
& Lindquist, C. (2006). The Serious and Violent Offender
Reentry Initiative: Measuring the effects on service
delivery. Western Criminology Review, 7(2), 3-19.
Winterfield, L., & Lindquist, C. (2005). Reentry research in
action: Characteristics of prisoner reentry programs.
Research Triangle Park, NC: RTI International.
Winterfield, L., Lindquist, C., & Brumbaugh, S. (2007).
Sustaining adult reentry programs after SVORI. The
Multi-site Evaluation of the Serious and Violent Offender
Reentry Initiative: Reentry research in action. Retrieved
December 18, 2009, from http://www.svori-evaluation.org/
%5Cdocuments%5Creports%5CRRIA_2007_Adult_Sustainabili
ty.pdf

Young, D. (2004). First count to ten: Innovation and
implementation to juvenile reintegration programs.
Retrieved December 15, 2009, from http://www.uscourts.
gov/fedprob/September_2004/juvenile.html

155

Appendix A. Data Tables
Exhibit A-1. Adult male case disposition—Wave 1 (pre-release)

TOTAL ALL CASES

Case Disposition—Eligible Cases
Completed
Interview completed
Released Early
R released prior to Wave 1 interview
Refused
Final refusal by R, guardian or other
Access Denied
Access to R denied by prison
Other Noninterview
R absconded
Private setting not available
R deceased
Language barrier—Spanish
Language barrier—Other
Physically/mentally incapable
Other noninterview
TOTAL ELIGIBLE CASES

Case Disposition—Ineligible Cases
Ineligible Cases
R transferred to non-study facility
R releasing to non-study area
R not releasing during data collection
period
Date of release unknown
Case fielded incorrectly
R ineligible to participate
Site dropped from study
Other ineligible
TOTAL INELIGIBLE CASES

SVORI
N
%
1,406 43.92%
SVORI
% of
Eligible
N
SVORI

Non-SVORI
N
%
1,795
56.08%
Non-SVORI
% of
Eligible
N
non-SVORI

All Cases
N
%
3,201 100.00%
All Cases

N

% of
Eligible

863

73.70%

834

59.87%

1,697

66.19%

169

14.43%

369

26.49%

538

20.98%

126

10.76%

166

11.92%

295

11.51%

6

0.51%

8

0.57%

14

0.55%

2
2
1
1
0
1
0
1,171

0.17%
0.17%
0.09%
0.09%
0.00%
0.09%
0.00%
100.00%

3
1
0
5
1
2
1
1,393

0.22%
0.07%
0.00%
0.36%
0.07%
0.14%
0.07%
100.00%

5
3
1
6
1
3
1
2,564

0.20%
0.12%
0.04%
0.23%
0.04%
0.12%
0.04%
100.00%

SVORI
% of
Ineligible
N
SVORI

Non-SVORI
% of
Ineligible
N
non-SVORI

All Cases

N

% of
Ineligible

21
7

8.94%
2.98%

56
37

13.93%
9.20%

77
41

12.09%
6.44%

100

42.55%

92

22.89%

192

30.14%

2
5
86
4
10
235

0.85%
2.13%
36.60%
1.70%
4.26%
100.00%

25
158
12
18
4
402

6.22%
39.30%
2.99%
4.48%
1.00%
100.00%

32
163
98
28
6
637

5.02%
25.59%
15.38%
4.40%
0.94%
100.00%

Note: R = respondent. Wave 1 = 30 days pre-release.

A-1

Prisoner reentry experiences of adult males

Exhibit A-2. Respondent characteristics, by group (pre-release)

Characteristic
Demographics and Housing
Age at incarceration
Age at pre-release (Wave 1) interview
White
Black
Hispanic
Multiracial/other
Born in United States
English is primary language
Homeless/shelter/no set place to live prior
to incarceration
Employment History
Ever held a job
Employed during 6 months prior to
incarceration
Source of support 6 months prior to
incarceration: Family
Source of support 6 months prior to
incarceration: Friends
Source of support 6 months prior to
incarceration: Government
Source of support 6 months prior to
incarceration: Illegal income
Source of support 6 months prior to
incarceration: Other
Last job: Hours worked per week
Last job: Hourly salary
Last job: Was permanent
Last job: Received formal pay
Last job: Health insurance provided
Completed 12th grade or GED/other high
school equivalent
Currently in school
Ever served in the military
Family and Peers
Married
Involved in steady relationship 6 months
prior to incarceration
Currently married or in steady relationship
Lived with spouse/partner before
incarceration

N

SVORI
Mean (SD)

Non-SVORI
Mean (SD)

1,697
1,697
1,694
1,694
1,694
1,694
1,697
1,697

26.13 (7.49)
28.89 (7.14)
0.32 (0.46)
0.57 (0.50)
0.04 (0.20)
0.08 (0.27)
1.00 (0.07)
0.98 (0.13)

27.06 (7.41)
29.30 (7.48)
0.37 (0.48)
0.50 (0.50)
0.04 (0.20)
0.09 (0.29)
0.98 (0.13)
0.97 (0.16)

−2.57
−1.17
−2.30
2.74
−0.13
−0.89
2.59
1.59

1,695

0.12 (0.33)

0.12 (0.33)

0.18

1,696

0.89 (0.31)

0.92 (0.27)

−2.21

1,696

0.64 (0.48)

0.68 (0.47)

−2.04

1,693

0.32 (0.47)

0.31 (0.46)

0.15

1,693

0.16 (0.37)

0.14 (0.35)

1.40

1,693

0.11 (0.31)

0.10 (0.30)

0.48

1,693

0.45 (0.50)

0.43 (0.50)

0.99

1,693

0.07 (0.25)

0.10 (0.30)

−2.14

1,107
1,083
1,117
1,120
1,094

41.72 (13.86)
10.91 (8.51)
0.75 (0.43)
0.74 (0.44)
0.37 (0.48)

41.76 (14.07)
10.13 (6.87)
0.73 (0.44)
0.72 (0.45)
0.34 (0.47)

−0.04
1.67
0.65
0.64
0.93

1,695

0.61 (0.49)

0.58 (0.49)

0.88

1,697
1,697

0.15 (0.35)
0.05 (0.22)

0.13 (0.34)
0.05 (0.21)

0.83
0.39

1,697

0.09 (0.28)

0.10 (0.30)

−1.05

1,693

0.68 (0.47)

0.69 (0.46)

−0.28

1,690

0.39 (0.49)

0.40 (0.49)

−0.33

670

0.59 (0.49)

0.67 (0.47)

−2.15

t-statistic

(continued)

A-2

Appendix A – Data Tables

Exhibit A-2. Respondent characteristics, by group (pre-release; continued)

Characteristic
Family and Peers (continued)
Have any living children
Number of children (only respondents with
children)
Number of children (respondents with and
without children)
Have child(ren) under 18
Primary care responsibilities for any
children under 18 6 months prior to
incarceration
Number of children under 18 supported 6
months prior to incarceration
Required to pay child support 6 months
prior to incarceration
Made court-ordered child support payments
6 months prior to incarceration
Court order for support changed while
incarcerated
Owe back child support
Dollar amount of back child support owed
State has forgiven/decreased back child
support
Have people in life that are considered
family
Have a family member who has been
convicted of a crime
Have a family member who has been in a
correctional facility
Have a family member who has had
problems with drugs/alcohol
Family emotional support scale (0–30:
>more support)
Had a friend (before incarceration) who has
been convicted of a crime
Had a friend (before incarceration) who has
been in a correctional facility
Had a friend (before incarceration) who has
had problems with drugs or alcohol
Physical and Mental Health
Physical health scale (>better)
Mental health scale (>better)
Received treatment for mental health
problem prior to this incarceration
Global Severity Index (45–225: >worse)
Positive Symptom Total (0–45: >worse)

N

SVORI
Mean (SD)

Non-SVORI
Mean (SD)

1,684

0.62 (0.49)

0.64 (0.48)

−0.88

1,056

2.22 (1.63)

2.29 (1.60)

−0.65

1,684

1.37 (1.67)

1.46 (1.69)

−1.07

1,684

0.59 (0.49)

0.61 (0.49)

−0.59

1,009

0.47 (0.50)

0.49 (0.50)

−0.59

527

1.17 (1.18)

1.19 (1.18)

−0.23

1,007

0.30 (0.46)

0.32 (0.47)

−0.56

312

0.59 (0.49)

0.56 (0.50)

0.51

283

0.26 (0.44)

0.27 (0.44)

−0.01

301

0.93 (0.25)
9127.02
(11281.27)

0.91 (0.29)
10728.93
(12558.94)

0.73

253

0.05 (0.21)

0.09 (0.28)

−1.21

1,697

0.97 (0.16)

0.97 (0.17)

0.27

1,574

0.75 (0.43)

0.76 (0.43)

−0.22

1,602

0.75 (0.44)

0.74 (0.44)

0.21

1,591

0.72 (0.45)

0.74 (0.44)

−0.99

1,615

21.63 (4.87)

21.35 (4.71)

1.18

1,540

0.83 (0.37)

0.83 (0.37)

−0.07

1,556

0.81 (0.39)

0.81 (0.39)

0.03

1,572

0.82 (0.39)

0.83 (0.38)

−0.42

1,673
1,673

53.63 (9.23)
48.93 (10.54)

53.34 (9.19)
48.51 (10.65)

0.64
0.80

1,693

0.24 (0.43)

0.25 (0.44)

−0.52

1,697
1,697

66.64 (21.43)
12.62 (9.77)

68.09 (23.07)
13.33 (10.07)

−1.34
−1.47

234

t-statistic

−1.03

(continued)

A-3

Prisoner reentry experiences of adult males

Exhibit A-2. Respondent characteristics, by group (pre-release; continued)

Characteristic
Physical and Mental Health (continued)
Anxiety Scale (5–25: >worse)
Depression Scale (5–25: >worse)
Hostility Scale (5–25: >worse)
Interpersonal Sensitivity Scale (5–25:
>worse)
Obsessive-Compulsive Scale (5–25:
>worse)
Paranoid Ideation Scale (5–25: >worse)
Phobic Anxiety Scale (5–25: >worse)
Psychoticism Scale (5–25: >worse)
Somatization Scale (5–25: >worse)
No physical health-related limitations
Ever had asthma
Currently have asthma
Receiving treatment for asthma
Taking prescription for asthma
Ever had diabetes
Currently have diabetes
Receiving treatment for diabetes
Taking prescription for diabetes
Ever had heart trouble
Currently have heart trouble
Receiving treatment for heart trouble
Taking prescription for heart trouble
Ever had high blood pressure
Currently have high blood pressure
Receiving treatment for high blood pressure
Taking prescription for high blood pressure
Ever had arthritis
Currently have arthritis
Receiving treatment for arthritis
Taking prescription for arthritis
Ever had chronic back pain
Currently have chronic back pain
Receiving treatment for chronic back pain
Taking prescription for chronic back pain
Ever had tuberculosis
Tuberculosis is currently active
Ever diagnosed as being HIV positive or
having AIDS

N

SVORI
Mean (SD)

Non-SVORI
Mean (SD)

1,696
1,696
1,697

7.42 (2.90)
8.31 (3.94)
6.41 (2.52)

7.67 (3.18)
8.45 (3.84)
6.69 (2.88)

−1.75
−0.76
−2.11

1,691

7.50 (3.30)

7.60 (3.55)

−0.62

1,697

8.12 (3.67)

8.17 (3.66)

−0.25

1,697
1,697
1,695
1,697
1,697
1,697
1,690
175
175
1,696
1,693
24
24
1,695
1,687
53
53
1,695
1,664
143
144
1697
1696
85
85
1,697
1,697
205
205
1,695
1,692

8.84 (3.66)
6.42 (2.32)
6.58 (2.38)
7.05 (2.78)
0.59 (0.49)
0.20 (0.40)
0.11 (0.31)
0.48 (0.50)
0.48 (0.50)
0.02 (0.15)
0.01 (0.11)
0.91 (0.30)
0.91 (0.30)
0.05 (0.23)
0.03 (0.17)
0.36 (0.49)
0.36 (0.49)
0.17 (0.38)
0.09 (0.29)
0.73 (0.45)
0.71 (0.46)
0.05 (0.23)
0.05 (0.21)
0.13 (0.33)
0.13 (0.33)
0.15 (0.35)
0.11 (0.32)
0.14 (0.35)
0.18 (0.39)
0.06 (0.23)
0.00 (0.00)

8.85 (3.74)
6.56 (2.74)
6.89 (2.59)
7.16 (3.04)
0.56 (0.50)
0.19 (0.39)
0.10 (0.30)
0.58 (0.50)
0.61 (0.49)
0.02 (0.13)
0.02 (0.12)
0.77 (0.44)
0.69 (0.48)
0.05 (0.22)
0.03 (0.18)
0.36 (0.49)
0.39 (0.50)
0.16 (0.37)
0.08 (0.27)
0.65 (0.48)
0.61 (0.49)
0.06 (0.23)
0.05 (0.23)
0.22 (0.42)
0.24 (0.43)
0.16 (0.37)
0.13 (0.33)
0.14 (0.35)
0.12 (0.33)
0.07 (0.25)
0.00 (0.03)

−0.04
−1.12
−2.61
−0.82
1.20
0.40
0.33
−1.32
−1.64
0.75
−0.49
0.89
1.29
0.49
−0.54
0.02
−0.24
0.70
0.80
1.09
1.25
−0.28
−0.71
−1.17
−1.41
−0.84
−0.93
0.05
1.24
−0.97
−1.00

1,697

0.01 (0.08)

0.01 (0.10)

−0.60

t-statistic

(continued)

A-4

Appendix A – Data Tables

Exhibit A-2. Respondent characteristics, by group (pre-release; continued)

Characteristic
Physical and Mental Health (continued)
Receiving treatment for HIV/AIDS
Taking prescription for HIV/AIDS
Ever had hepatitis B or C
Currently have hepatitis B or C
Receiving treatment for hepatitis B or C
Taking prescription for hepatitis B or C
Wear glasses or corrective lenses
Need eye glasses
Currently use a hearing aid
Need a hearing aid
Ever received care for mental health or
alcohol/drug problems
Ever received care for: Alcohol
abuse/dependence
Ever received care for: Anxiety
Ever received care for: Attention deficit
hyperactivity disorder
Ever received care for: Bipolar disorder
Ever received care for: Conduct disorder
Ever received care for:
Depression/dysthymia
Ever received care for: Drug
abuse/dependence
Ever received care for: Obsessivecompulsive disorder
Ever received care for: Oppositional defiant
disorder
Ever received care for: Posttraumatic stress
disorder
Ever received care for: Phobia (social or
specific)
Ever received care for: Schizophrenia
Ever received care for: Other
problem/diagnosis
Did not receive care for problem/no
diagnosis
Currently receiving treatment: Alcohol
abuse/dependence
Currently receiving treatment: Anxiety
disorder
Currently receiving treatment: Attention
deficit hyperactivity disorder

N

SVORI
Mean (SD)

Non-SVORI
Mean (SD)

14
14
1,691
1,689
60
60
1,697
1,238
1,697
1,690

0.83 (0.41)
0.67 (0.52)
0.03 (0.18)
0.03 (0.16)
0.23 (0.43)
0.14 (0.35)
0.27 (0.45)
0.22 (0.41)
0.00 (0.05)
0.02 (0.15)

0.88 (0.35)
0.88 (0.35)
0.05 (0.22)
0.05 (0.21)
0.11 (0.31)
0.05 (0.23)
0.26 (0.44)
0.22 (0.42)
0.01 (0.08)
0.05 (0.21)

−0.20
−0.90
−1.61
−2.25
1.27
1.00
0.34
−0.10
−1.17
−2.54

1,696

0.56 (0.50)

0.55 (0.50)

0.32

925

0.25 (0.44)

0.28 (0.45)

−0.87

925

0.06 (0.23)

0.07 (0.26)

−0.88

925

0.12 (0.33)

0.13 (0.33)

−0.31

925
925

0.10 (0.30)
0.03 (0.18)

0.12 (0.33)
0.04 (0.19)

−1.23
−0.34

925

0.19 (0.39)

0.20 (0.40)

−0.32

925

0.42 (0.49)

0.34 (0.48)

2.33

925

0.01 (0.12)

0.02 (0.12)

−0.10

925

0.01 (0.11)

0.00 (0.07)

1.36

925

0.03 (0.18)

0.02 (0.15)

0.66

925

0.01 (0.08)

0.01 (0.11)

−1.08

925

0.04 (0.21)

0.05 (0.21)

−0.18

925

0.18 (0.39)

0.18 (0.38)

0.12

925

0.17 (0.38)

0.15 (0.36)

0.97

783

0.07 (0.25)

0.10 (0.29)

−1.38

783

0.02 (0.14)

0.03 (0.17)

−0.94

783

0.01 (0.09)

0.03 (0.16)

−1.98

t-statistic

(continued)

A-5

Prisoner reentry experiences of adult males

Exhibit A-2. Respondent characteristics, by group (pre-release; continued)

Characteristic
Physical and Mental Health (continued)
Currently receiving treatment: Bipolar
disorder
Currently receiving treatment: Conduct
disorder
Currently receiving treatment:
Depression/dysthymia
Currently receiving treatment: Drug
abuse/dependence
Currently receiving treatment: Obsessivecompulsive disorder
Currently receiving treatment: Oppositional
defiant disorder
Currently receiving treatment: Posttraumatic
stress disorder
Currently receiving treatment: Phobia
(social or specific)
Currently receiving treatment:
Schizophrenia
Currently receiving treatment: Other
problem/diagnosis
Currently not receiving treatment for any
condition
Doctor prescribed medication for
emotional/psychological problem during
this incarceration
Received the prescribed medication
Any victimization (6 months prior to
incarceration)
Victimization severity prior to incarceration
(0–30: >worse)
Any victimization (during incarceration)
Victimization severity during incarceration
(0–36: >worse)
Substance Use
Ever drank any type of alcoholic beverage
Age at first drink
Used alcohol 30 days prior to this
incarceration
Age at last drink if no alcohol 30 days prior
Ever used drugs
Number of drugs used in lifetime
Used drugs 30 days prior to this
incarceration
Number of drugs used 30 days prior to this
incarceration

N

SVORI
Mean (SD)

Non-SVORI
Mean (SD)

783

0.05 (0.21)

0.06 (0.24)

−0.69

783

0.01 (0.07)

0.01 (0.09)

−0.47

783

0.06 (0.23)

0.10 (0.29)

−2.10

783

0.10 (0.31)

0.09 (0.28)

0.90

783

0.00 (0.00)

0.01 (0.09)

−1.74

783

0.00 (0.00)

0.00 (0.00)

783

0.01 (0.11)

0.01 (0.10)

0.31

783

0.00 (0.05)

0.00 (0.05)

−0.01

783

0.04 (0.19)

0.03 (0.18)

0.15

783

0.05 (0.22)

0.06 (0.24)

−0.68

783

0.72 (0.45)

0.67 (0.47)

1.48

1,697

0.13 (0.34)

0.19 (0.39)

−3.23

268

0.95 (0.23)

0.96 (0.21)

−0.33

1,696

0.59 (0.49)

0.58 (0.49)

0.61

1,696

3.87 (5.61)

3.75 (5.49)

0.47

1,696

0.55 (0.50)

0.54 (0.50)

0.47

1,696

2.71 (3.64)

2.88 (4.05)

−0.93

1,696
1,616

0.96 (0.19)
13.71 (3.85)

0.97 (0.17)
13.64 (3.76)

−0.80
0.34

1,693

0.68 (0.47)

0.67 (0.47)

0.43

479
1,697
1,697

24.18 (7.41)
0.94 (0.24)
3.39 (2.78)

25.66 (7.86)
0.96 (0.21)
3.84 (2.93)

−2.11
−1.67
−3.26

1,696

0.66 (0.48)

0.69 (0.46)

−1.56

1,697

1.37 (1.56)

1.58 (1.75)

−2.63

t-statistic

(continued)

A-6

Appendix A – Data Tables

Exhibit A-2. Respondent characteristics, by group (pre-release; continued)

Characteristic
Substance Use (continued)
Used drugs other than marijuana and
steroids 30 days prior to this incarceration
Ever used sedatives
Age first used sedatives
Used sedatives 30 days prior to this
incarceration
Age last used sedatives
Ever used tranquilizers
Age first used tranquilizers
Used tranquilizers 30 days prior to this
incarceration
Age last used tranquilizers
Ever used stimulants
Age first used stimulants
Used stimulants 30 days prior to this
incarceration
Age last used stimulants
Ever used pain relievers
Age first used pain relievers
Used pain relievers 30 days prior to this
incarceration
Age last used pain relievers
Ever used methadone
Age first used methadone
Used methadone 30 days prior to this
incarceration
Age last used methadone
Ever used anabolic steroids
Age first used anabolic steroids
Used anabolic steroids 30 days prior to this
incarceration
Age last used anabolic steroids
Ever used marijuana
Age first used marijuana
Used marijuana 30 days prior to this
incarceration
Age last used marijuana
Ever used hallucinogens
Age first used hallucinogens
Used hallucinogens 30 days prior to this
incarceration

N

SVORI
Mean (SD)

Non-SVORI
Mean (SD)

1,696

0.42 (0.49)

0.47 (0.50)

−1.92

1,695
328

0.18 (0.39)
17.62 (4.24)

0.21 (0.41)
17.13 (4.45)

−1.63
1.02

1,693

0.06 (0.23)

0.09 (0.29)

−2.88

205
1,695
461

22.48 (5.32)
0.25 (0.43)
17.93 (4.34)

24.12 (7.09)
0.31 (0.46)
18.47 (5.04)

−1.86
−2.86
−1.22

1,691

0.08 (0.28)

0.13 (0.33)

−2.86

285
1,696
298

22.79 (5.62)
0.16 (0.36)
16.66 (4.09)

23.04 (6.47)
0.20 (0.40)
17.05 (4.77)

−0.35
−2.31
−0.75

1,696

0.07 (0.25)

0.09 (0.29)

−1.84

165
1,695
454

21.05 (5.30)
0.24 (0.43)
18.21 (4.96)

22.84 (6.91)
0.30 (0.46)
18.53 (5.59)

−1.88
−2.78
−0.64

1,693

0.11 (0.31)

0.14 (0.34)

−1.97

251
1,695
132

23.38 (5.46)
0.06 (0.24)
23.71 (8.24)

24.67 (7.15)
0.09 (0.29)
23.10 (6.62)

−1.61
−2.28
0.47

1,695

0.02 (0.13)

0.02 (0.13)

−0.28

103
1,696
30

26.95 (8.63)
0.02 (0.13)
17.94 (4.54)

26.27 (7.36)
0.02 (0.13)
19.50 (3.20)

0.43
0.27
−1.07

1,696

0.00 (0.00)

0.00 (0.00)

30
1,695
1,568

18.94 (5.32)
0.92 (0.27)
13.94 (3.15)

21.79 (4.04)
0.94 (0.24)
14.14 (3.33)

−1.63
−1.25
−1.24

1,694

0.52 (0.50)

0.53 (0.50)

−0.76

675
1,695
784

23.33 (7.20)
0.43 (0.50)
17.16 (3.45)

23.61 (6.72)
0.49 (0.50)
17.58 (3.95)

−0.53
−2.51
−1.59

1,694

0.09 (0.28)

0.09 (0.29)

−0.30

t-statistic

(continued)

A-7

Prisoner reentry experiences of adult males

Exhibit A-2. Respondent characteristics, by group (pre-release; continued)

Characteristic
Substance Use (continued)
Age last used hallucinogens
Ever used cocaine
Age first used cocaine
Used cocaine 30 days prior to this
incarceration
Age last used cocaine
Ever used heroin
Age first used heroin
Used heroin 30 days prior to this
incarceration
Age last used heroin
Ever used amphetamines
Age first used amphetamines
Used amphetamines 30 days prior to this
incarceration
Age last used amphetamines
Ever used inhalants
Age first used inhalants
Used inhalants 30 days prior to this
incarceration
Age last used inhalants
Received alcohol/drug treatment before this
incarceration
Current Incarceration and Criminal Historya
Duration of incarceration at Wave 1
interview (years)
Wave 1 conviction offense(s) category:
Person/violent crime
Robbery
Assault
Lethal crime
Sex offense
Other person/violent crime
Wave 1 conviction offense(s) category:
Property crime
Burglary
Theft
Car theft
Fraud/forgery
Other property crime

N

SVORI
Mean (SD)

Non-SVORI
Mean (SD)

626
1,694
935

20.92 (4.07)
0.53 (0.50)
19.39 (5.32)

21.56 (5.35)
0.58 (0.49)
19.52 (4.90)

−1.69
−2.09
−0.39

1,694

0.22 (0.42)

0.26 (0.44)

−1.77

528
1,695
343

24.65 (7.31)
0.18 (0.38)
20.90 (6.13)

24.62 (6.95)
0.23 (0.42)
21.34 (5.62)

0.04
−2.59
−0.68

1,695

0.08 (0.27)

0.09 (0.28)

−0.83

206
1,692
473

26.19 (8.70)
0.26 (0.44)
17.10 (3.76)

24.75 (6.65)
0.30 (0.46)
18.47 (4.81)

1.30
−1.86
−3.47

1,690

0.13 (0.33)

0.14 (0.34)

−0.55

251
1,694
267

22.58 (5.97)
0.15 (0.36)
15.83 (3.91)

23.74 (6.49)
0.16 (0.37)
15.76 (3.34)

−1.45
−0.63
0.16

1,693

0.01 (0.10)

0.01 (0.08)

0.71

252

18.06 (4.87)

17.34 (4.17)

1.26

1,696

0.42 (0.49)

0.41 (0.49)

0.42

1,697

2.76 (2.46)

2.26 (2.63)

4.10

1,688

0.42 (0.49)

0.40 (0.49)

0.92

1,688
1,688
1,688
1,688
1,688

0.15 (0.36)
0.19 (0.39)
0.04 (0.21)
0.05 (0.22)
0.03 (0.18)

0.13 (0.33)
0.16 (0.36)
0.03 (0.17)
0.07 (0.25)
0.06 (0.24)

1.28
1.88
1.68
−1.51
−2.46

1,688

0.24 (0.43)

0.27 (0.44)

−1.35

1,688
1,688
1,688
1,688
1,688

0.11 (0.31)
0.08 (0.28)
0.03 (0.16)
0.02 (0.15)
0.04 (0.20)

0.12 (0.32)
0.08 (0.27)
0.03 (0.18)
0.05 (0.21)
0.05 (0.21)

−0.71
0.43
−0.55
−2.52
−0.50

t-statistic

(continued)

A-8

Appendix A – Data Tables

Exhibit A-2. Respondent characteristics, by group (pre-release; continued)

SVORI
Characteristic
N
Mean (SD)
Current Incarceration and Criminal Historya (continued)
Wave 1 conviction offense(s) category:
1,688
0.36 (0.48)
Drug crime
Drug dealing/manufacturing
1,688
0.21 (0.41)
Drug possession
1,688
0.22 (0.41)
Other drug offense
1,688
0.01 (0.11)
Wave 1 conviction offense(s) category:
1,688
0.17 (0.37)
Public order crime
Wave 1 conviction offense(s) category:
1,688
0.02 (0.13)
Other crime
Current incarceration for probation or parole
1,695
0.27 (0.44)
violation
Current incarceration for probation violation
1,695
0.05 (0.22)
Current incarceration for parole violation
1,695
0.22 (0.41)
Parole violation: Technical violation
459
0.59 (0.49)
Parole violation: New crime
459
0.42 (0.49)
Age at first arrest
1,685
15.92 (4.78)
Number of lifetime arrests
1,586
12.42 (11.45)
Number of lifetime convictions
1,658
5.48 (6.05)
Number of lifetime convictions/age at
1,658
0.21 (0.24)
incarceration
Ever locked up in a juvenile correctional
1,696
0.51 (0.50)
facility for committing a crime
Number of times in juvenile lockup (only
833
3.58 (3.89)
those who reported ever being locked up)
Number of times in juvenile lockup (all
1,680
1.82 (3.30)
respondents)
Ever been in jail/prison more than 24 hours
1,694
0.83 (0.38)
at one time
Number of times sent to prison (only those
1,434
1.45 (1.82)
who reported ever having been in prison)
Number of times sent to prison (all
1,688
1.20 (1.74)
respondents)
Any disciplinary infractions during this
1,694
0.65 (0.48)
incarceration
One disciplinary infraction during this
1,694
0.17 (0.38)
incarceration
Two or more disciplinary infractions during
1,694
0.47 (0.50)
this incarceration
Placed in administrative segregation during
1,692
0.45 (0.50)
this incarceration

Non-SVORI
Mean (SD)

t-statistic

0.31 (0.46)

2.36

0.15 (0.36)
0.21 (0.41)
0.01 (0.10)

3.34
0.65
0.16

0.22 (0.42)

−2.92

0.01 (0.10)

1.21

0.35 (0.48)

−3.71

0.06 (0.25)
0.29 (0.45)
0.64 (0.48)
0.37 (0.48)
16.03 (5.09)
13.14 (11.39)
5.70 (6.26)

−1.42
−3.18
−1.05
1.03
−0.47
−1.25
−0.73

0.22 (0.25)

−0.25

0.49 (0.50)

1.07

3.49 (3.64)

0.35

1.69 (3.07)

0.86

0.87 (0.33)

−2.42

1.69 (2.05)

−2.35

1.47 (1.99)

−2.97

0.56 (0.50)

3.50

0.17 (0.37)

0.36

0.40 (0.49)

3.17

0.40 (0.49)

2.41
(continued)

A-9

Prisoner reentry experiences of adult males

Exhibit A-2. Respondent characteristics, by group (pre-release; continued)

SVORI
Characteristic
N
Mean (SD)
Current Incarceration and Criminal Historya (continued)
Current gang member
1,688
0.05 (0.21)
Considers gang to be family
92
0.53 (0.51)
Relatives are members of the gang
92
0.55 (0.50)
Any perpetration of violence (6 months
1,697
0.69 (0.46)
prior to incarceration)
Current Gang Member
Have institution job
1,697
0.63 (0.48)
Have prison service job
1,692
0.60 (0.49)
Have prison industry job
1,692
0.04 (0.19)
Hours per week at institution job
1,035
23.78 (16.95)
Have work release job
1,697
0.03 (0.17)
Hours per week at work release job
59
39.92 (14.28)

Non-SVORI
Mean (SD)

t-statistic

0.06 (0.24)
0.52 (0.50)
0.58 (0.50)

−1.45
0.05
−0.26

0.67 (0.47)

0.80

0.61 (0.49)
0.57 (0.50)
0.04 (0.21)
22.29 (16.12)
0.04 (0.20)
30.97 (11.91)

0.95
1.40
−0.77
1.45
−1.31
2.63

Note: GED = general educational development. SD = standard deviation. Wave 1 = 30 days pre-release.
a

Results for W1 Conviction Offenses may not sum to 100% because some respondents reported multiple conviction
offenses.

A-10

Appendix A – Data Tables

Exhibit A-3. Proportion of respondents who reported needing specific services, by group
(pre-release)

Service
Transition Services
Legal assistance
Financial assistance
Public financial assistance
Public health care insurance
Mentor
Documents for employment
Place to live
Transportation
Driver’s license
Access to clothing/food banks
Health Services
Medical treatment
Mental health treatment
Substance use treatment
Victims’ group for abuse
Anger management program
Employment/Education/Skills Services
Job
Job training
More education
Money management skills
Life skills
Work on personal relationships
Change attitudes on criminal behavior
Domestic Violence Services
Batterer intervention program
Domestic violence support group
Child Services
Child support payments
Modification of child support debt
Modification of child custody
Parenting skills
Child care

N

SVORI
Mean (SD)

Non-SVORI
Mean (SD)

1,690
1,696
1,695
1,693
1,695
1,697
1,695
1,696
1,697
1,696

0.45 (0.50)
0.86 (0.35)
0.52 (0.50)
0.75 (0.43)
0.60 (0.49)
0.55 (0.50)
0.49 (0.50)
0.72 (0.45)
0.83 (0.38)
0.60 (0.49)

0.48 (0.50)
0.82 (0.39)
0.54 (0.50)
0.73 (0.45)
0.61 (0.49)
0.56 (0.50)
0.46 (0.50)
0.71 (0.46)
0.81 (0.39)
0.55 (0.50)

−1.38
2.61
−0.94
1.19
−0.37
−0.15
1.32
0.59
1.02
2.30

1,696
1,693
1,696
1,697
1,694

0.56 (0.50)
0.22 (0.42)
0.37 (0.48)
0.04 (0.20)
0.36 (0.48)

0.57 (0.50)
0.29 (0.45)
0.43 (0.50)
0.04 (0.20)
0.38 (0.48)

−0.19
−3.09
−2.64
0.22
−0.82

1,696
1,696
1,697
1,696
1,690
1,694
1,693

0.80 (0.40)
0.82 (0.39)
0.94 (0.24)
0.71 (0.45)
0.75 (0.43)
0.64 (0.48)
0.64 (0.48)

0.76 (0.43)
0.76 (0.43)
0.92 (0.27)
0.68 (0.47)
0.73 (0.44)
0.64 (0.48)
0.69 (0.46)

1.94
2.62
1.23
1.38
0.96
0.15
−2.12

1,694
1,695

0.08 (0.27)
0.06 (0.24)

0.08 (0.27)
0.09 (0.28)

−0.02
−2.23

995
276
1,002
1,009
1,007

0.45 (0.50)
0.88 (0.33)
0.35 (0.48)
0.60 (0.49)
0.39 (0.49)

0.48 (0.50)
0.86 (0.35)
0.38 (0.49)
0.63 (0.48)
0.39 (0.49)

−1.04
0.48
−0.97
−1.11
0.08

t-statistic

SD = standard deviation.

A-11

Prisoner reentry experiences of adult males

Exhibit A-4. Proportion of respondents who reported receiving specific services, by group
(pre-release)

Variable Label
Coordination Services
Received needs assessment
Received release-specific needs assessment
Met with case manager
Developed reentry plan
Worked with anyone to plan for release
Transition Services
Participated in programs to prepare for release
Took class specifically for release
Received legal assistance
Received assistance accessing financial assistance
Received assistance accessing public financial assistance
Received assistance accessing public health care
assistance
Received mentoring services
Received assistance obtaining documents
Received assistance finding transportation
Received assistance finding place to live
Received assistance getting driver’s license
Received assistance accessing clothing/food banks
Health Services
Received any medical treatment
Received dental services
Received any mental health treatment
Received any substance use treatment
Participated in groups for victims of abuse
Participated in anger management program

N

SVORI

NonSVORI

t-statistic

1,690
1,678
1,694
1,663
1,695

0.63 (0.48)
0.49 (0.50)
0.66 (0.47)
0.57 (0.50)
0.66 (0.48)

0.45 (0.50)
0.23 (0.42)
0.40 (0.49)
0.24 (0.43)
0.31 (0.46)

7.43
11.61
11.05
14.69
15.22

1,696
1,695
1,697
1,697
1,696

0.75 (0.43)
0.65 (0.48)
0.12 (0.32)
0.13 (0.34)
0.14 (0.35)

0.51 (0.50)
0.37 (0.48)
0.08 (0.27)
0.04 (0.19)
0.11 (0.31)

10.64
11.89
2.38
7.11
1.81

1,695 0.13 (0.34)

0.09 (0.29)

2.46

1,697
1,693
1,696
1,697
1,696
1,696

0.20 (0.40)
0.41 (0.49)
0.19 (0.39)
0.28 (0.45)
0.22 (0.41)
0.21 (0.41)

0.08 (0.27)
0.26 (0.44)
0.12 (0.32)
0.13 (0.33)
0.08 (0.27)
0.11 (0.32)

6.92
6.66
4.30
7.82
8.46
5.54

1,691
1,696
1,675
1,696
1,696
1,696

0.58 (0.49) 0.55 (0.50)
0.50 (0.50) 0.47 (0.50)
0.16 (0.36) 0.20 (0.40)
0.48 (0.50) 0.38 (0.48)
0.07 (0.25) 0.03 (0.16)
0.34 (0.48) 0.26 (0.44)

1.55
1.38
−2.17
4.44
4.02
3.88
(continued)

A-12

Appendix A – Data Tables

Exhibit A-4. Proportion of respondents who reported receiving specific services, by group
(pre-release; continued)

Variable Label
Employment/Education/Skills Services
Received any employment services
Participated in employment readiness program
Participated in job training program
Talked to potential employer
Given advice about job interviewing
Given advice about answering questions about criminal
history
Given advice about how to behave on the job
Given names of people to contact in community to find
job
Put together a resume
Received any educational services
Received money management services
Received other life skills training
Received assistance with personal relationships
Received training to change criminal behavior attitudes
Domestic Violence Services
Participated in batterer intervention programs
Participated in domestic violence support groups
Child Services
Received assistance making child support payments
Received assistance modifying child support debt
Received assistance modifying child custody
Participated in parenting classes
Received assistance finding child care

N
1,696
1,693
1,696
1,696
1,696

SVORI
0.37 (0.48)
0.23 (0.42)
0.17 (0.38)
0.15 (0.35)
0.32 (0.47)

NonSVORI

tstatistic

0.19 (0.39)
0.09 (0.28)
0.04 (0.20)
0.06 (0.23)
0.14 (0.35)

8.71
8.06
9.16
6.37
9.01

1,695 0.30 (0.46) 0.13 (0.34)

8.53

1,696 0.31 (0.46) 0.13 (0.34)

9.12

1,695 0.27 (0.44) 0.13 (0.33)

7.37

1,696
1,697
1,696
1,693
1,697
1,697

8.01
4.06
9.28
9.84
4.32
6.76

0.24 (0.43)
0.53 (0.50)
0.24 (0.43)
0.42 (0.49)
0.25 (0.43)
0.52 (0.50)

1,696 0.05 (0.22)
1,697 0.11 (0.31)
1,009
310
1,009
1,011
1,010

0.10 (0.30)
0.43 (0.50)
0.08 (0.27)
0.21 (0.41)
0.17 (0.37)
0.36 (0.48)
0.03 (0.18)
0.06 (0.23)

1.44
3.91

0.07 (0.25) 0.02 (0.14)
0.22 (0.42) 0.11 (0.31)
0.04 (0.19) 0.02 (0.15)
0.25 (0.43) 0.15 (0.36)
0.08 (0.27) 0.03 (0.16)

3.70
2.77
1.29
4.04
3.73

A-13

Prisoner reentry experiences of adult males

Exhibit A-5. Proportion of respondents who reported living with the indicated individuals in
the period since the last interview

Variable Label
Mother/stepmother
Boy/girlfriend/fiancé
Father/stepfather
Sister/stepsister
Child/stepchild
Nobody
Brother/stepbrother
Someone else
Niece/nephew
Husband/wife
Friend
Aunt/uncle
Grandparent
Facility/shelter
residents
Cousin
In-laws
Foster parent
Ex-husband/wife

Wave 2
NonSVORI SVORI tstat
0.35
0.33
0.6984
0.18
0.19 −0.4843
0.18
0.15
1.1948
0.16
0.12
1.8106
0.13
0.13 −0.1810
0.11
0.07
1.6992
0.11
0.09
0.5784
0.09
0.10 −0.6759
0.08
0.07
0.4181
0.07
0.09 −0.9268
0.05
0.07 −1.3948
0.05
0.05 −0.2640
0.05
0.08 −1.9078
0.04
0.04 −0.0205
0.03
0.02
0.01
0.00

0.03
0.03
0.00
0.00

Note: NA = not applicable. S = SVORI.

A-14

0.6960
−0.7852
1.7353
NA

Wave 3
NonSVORI SVORI
0.30
0.29
0.23
0.29
0.17
0.14
0.13
0.09
0.17
0.17
0.10
0.08
0.10
0.10
0.10
0.10
0.06
0.06
0.09
0.10
0.08
0.09
0.06
0.05
0.07
0.06
0.02
0.03
0.04
0.02
0.00
0.00

0.03
0.03
0.01
0.00

tstat
0.4019
−2.0785
1.0535
1.9282
−0.2178
0.9764
0.3645
0.0836
0.4632
−0.7288
−0.5302
0.7254
0.5972
−1.0485
1.4518
−1.2500
−1.0064
−0.1271

Wave 4
NonSVORI SVORI
0.26
0.27
0.27
0.26
0.13
0.14
0.11
0.12
0.18
0.19
0.10
0.07
0.10
0.08
0.13
0.11
0.05
0.04
0.10
0.12
0.09
0.09
0.05
0.05
0.05
0.08
0.04
0.04
0.04
0.02
0.00
0.01

0.03
0.02
0.00
0.00

tstat
−0.4099
0.2058
−0.2771
−0.1980
−0.2829
1.5907
1.2370
1.1021
0.6577
−1.2257
−0.1815
−0.0921
−2.0237
0.2925
0.4846
−0.2261
0.5025
0.9270

 

 

Prison Profiteers - Side
Advertise Here 2nd Ad
The Habeas Citebook Ineffective Counsel Side