Prisoner Reentry Experiences
of Adult Females:
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

Christine H. Lindquist

Kelle Barrick

RTI International
3040 East Cornwallis Road
Research Triangle Park, NC 27709
Phone: (919) 485-5706
Fax: (919) 541-5985
Lindquist@rti.org

RTI International
3040 East Cornwallis Road
Research Triangle Park, NC 27709
Phone: (919) 541-6435
Fax: (919) 541-5985
Kbarrick@rti.org

Pamela K. Lattimore

Christy A. Visher

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

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
Christine H. Lindquist, RTI International
Kelle Barrick, RTI International
Pamela K. Lattimore, RTI International
Christy A. Visher, University of Delaware and Urban Institute
Staff Contributors
Danielle M. Steffey, RTI International
Laurin Parker, University of Delaware
Debbie Dawes, RTI International
Alexander Cowell, RTI International
Susan Brumbaugh, 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 program
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 agencies in 2003 to develop programs to improve
criminal justice, employment, education, health, and housing
outcomes for released prisoners. Sixty-nine agencies received
federal funds to develop 89 programs. The SVORI Multi-site
Evaluation was funded by the National Institute of Justice to
examine the extent to which the SVORI (1) improved access to
appropriate, comprehensive, integrated services; (2) improved
employment, health, and personal functioning; and (3) reduced
criminal recidivism. Sixteen programs—12 adult and 4
juvenile—were included in an impact evaluation to determine
the effectiveness of the programming provided under SVORI.
Research Subjects
This report presents findings from the pre-release and postrelease interviews conducted with women in 11 impact sites.
The sample includes 153 females enrolled in SVORI programs
and 204 comparison females who did not receive SVORI
programming. The respondent profile revealed a high-risk,
high-need study group. The women reported many physical and
mental health problems, with half reporting receiving treatment
for mental health problems before the current period of
incarceration. Whereas more than half of the women reported
working during the six months before prison, nearly as many
reported receiving income from illegal activities. The women
reported an average of 11 arrests, with the first occurring at 19
years of age, and nearly all reported at least one previous
incarceration. The women reported very high levels of current
service need; among the most commonly reported were
education, public health insurance, financial assistance,
employment, and mentoring.

v

Study Methods
The focus of the evaluation was to assess whether SVORI
respondents received more services than non-SVORI
respondents and to examine differences between the groups on
a variety of post-release outcomes. Propensity score weights
were developed, tested, and applied to improve the
comparability of the SVORI and non-SVORI groups. Weighted
analyses were used to examine the treatment effect of SVORI.
Major Findings
In terms of service receipt, SVORI and non-SVORI respondents
reported the highest levels of service receipt during
confinement. Whereas both groups reported low levels of postrelease service receipt, SVORI respondents generally reported
higher levels of service receipt than non-SVORI respondents.
However, the levels of post-release service receipt reported by
both groups were considerably lower than their reported levels
of service need.
SVORI programming appeared to have a positive impact on
both employment outcomes and abstinence from drug use. The
findings for criminal behavior were mixed; the women enrolled
in SVORI had positive outcomes for self-reported criminal
behavior and official measures for rearrest but had negative
outcomes for self-reported compliance with conditions of
supervision and official measures of reincarceration. SVORI
programming did not appear to affect core housing outcomes,
familial or peer relationships, or physical or mental health
outcomes.
Conclusions
Study findings clearly demonstrate that female prisoners
returning to society are a population with high needs. While the
SVORI programs were successful in increasing services
provided to female participants, the levels of services that
female SVORI participants received failed to match their high
levels of need. However, the findings support the notion that
enhanced access to a variety of reentry services results in
modest improvements among several key reentry domains for
women. The current evaluation’s detailed documentation of
service areas for which women reported high needs can be
used for effective planning and service delivery. Because of the
variety of challenges that returning women prisoners face,
particularly with respect to mental and physical health

vi

problems, extensive family responsibilities, and lack of
employment experience (compared with reentering male
prisoners), effective coordination of services is necessary.

vii

Contents
Section

Page

Executive Summary

ES-1 

Introduction

1 

Previous Research on Women and Reentry

5 

Women Prisoners: Prevalence and Needs ........................ 5 
Reentry Experiences Among Women ............................... 6 
Reentry Programming for Women .................................. 7 
An Overview of SVORI Programming for Female
Offenders

11 

The SVORI Multi-site Evaluation—Design and
Methods

15 

Study Design .............................................................15 
Data Collection Procedures ...........................................17 
Approach for Addressing Nonresponse and Attrition .........18 
Approach for Addressing Selection Bias ..........................19 
Pre-release Experiences of Returning Female
Prisoners

21 

Background Characteristics ..........................................22 
In-Prison Experiences ..................................................42 
Service Needs ............................................................44 
Service Receipt...........................................................57 
Post-release Experiences of Returning Female
Prisoners

69 

Service Needs ............................................................70 
Service Receipt...........................................................80 
Housing .....................................................................92 
Employment ............................................................ 100 
Family, Peers, and Community Involvement ................. 107 
Substance Use and Physical and Mental Health.............. 112 
Criminal Behavior and Recidivism ................................ 119 

ix

x

Conclusions

131 

References

139 

Appendix A. Data Tables

A-1 

Exhibits
Exhibit Number

Page

1.

Adult female sample sizes, by state and group...............16 

2.

Demographic characteristics of respondents at time
of interview, by group ................................................23 

3.

Percentages of mothers reporting on child care or
child support responsibilities, by group .........................24 

4.

Criminal history and substance use of family and
peers, by group.........................................................26

5.

Lifetime substance use, by group .................................28 

6.

Use of specific substances during the 30 days
preceding incarceration, by group ................................29 

7.

Lifetime health problems, by group ..............................31 

8.

Current health problems, by group...............................32 

9.

Average scores on SA-45 subscales, by group ...............34 

10. Employment before incarceration, by group...................35 
11. Characteristics of respondents’ jobs before
incarceration, by group ..............................................36 
12. Sources of income during the 6 months before
incarceration, by employment status and group .............38 
13. Criminal history of respondents, by group .....................40 
14. Conviction offenses for current incarceration, by
group ......................................................................41 
15. Disciplinary infractions and administrative
segregations during current incarceration, by group .......43 
16. Frequency of in-prison contact with family members
and friends, by group .................................................45 
17. Service need bundle scores across service bundles,
by group ..................................................................47 
18. Self-reported need for specific transition services,
by group ..................................................................49 
19. Self-reported need for specific health services, by
group ......................................................................51 

xi

20. Self-reported need for employment, education, and
skills services, by group .............................................52 
21. Self-reported need for specific child services, by
group ......................................................................54 
22. Most commonly reported service needs, by group ..........56 
23. Service receipt bundle scores across service
bundles, by group .....................................................59 
24. Self-reported receipt of specific coordination
services, by group .....................................................60 
25. Self-reported receipt of specific transition services,
by group ..................................................................61 
26. Self-reported receipt of specific health services, by
group ......................................................................63 
27. Self-reported receipt of specific employment,
education, and skills services, by group ........................64 
28. Most commonly reported services received, by
group ......................................................................66 
29. Weighted means and parameter estimates of the
effect of SVORI on service need...................................71 
30. Transition service needs bundle score...........................74 
31. Health service needs bundle score ...............................76 
32. Employment/education/life skills service needs
bundle score .............................................................77 
33. Domestic violence service needs bundle score ...............78 
34. Child service needs bundle score .................................79 
35. Weighted means and parameter estimates of the
effect of SVORI on service receipt ................................81 
36. Coordination services receipt bundle score ....................84 
37. Transition services receipt bundle score ........................86 
38. Health services receipt bundle score .............................87 
39. Employment/education/life skills services receipt
bundle score .............................................................89 
40. Domestic violence services receipt bundle score .............90 
41. Child services receipt bundle score ...............................91 
42. Weighted means and parameter estimates of the
effect of SVORI on housing outcomes ...........................93 
43. Self-reported housing independence since
release/last interview .................................................94 
44. Self-reported housing stability .....................................95 
45. Self-reported lack of housing challenges since
release/last interview .................................................96 
46. Self-reported being homeless, living in a shelter, or
being without a set place to live ..................................98 

xii

47. Weighted means and parameter estimates of the
effect of SVORI on employment outcomes................... 101 
48. Self-reported currently supporting self with job ............ 102 
49. Self-reported number of months worked since
release/last interview ............................................... 103 
50. Self-reported number of months worked at same
job since release/last interview .................................. 104 
51. Self-reported receipt of formal pay for current job ........ 105 
52. Self-reported having a job with benefits...................... 106 
53. Negative peer exposure (0–14, <better) ..................... 110 
54. Self-reported high civic action since release/last
interview ................................................................ 112 
55. Weighted means and parameter estimates of the
effect of SVORI on self-reported substance use
outcomes ............................................................... 114 
56. No drug use reported or detected since release/last
interview ................................................................ 115 
57. No drug use reported or detected in past 30 days ........ 116 
58. Physical health scale (12-Item Short-Form Health
Survey) ................................................................. 118 
59. Self-reported overall physical health is very good or
excellent ................................................................ 119 
60. Global Severity Index (45–225, >better) .................... 120 
61. Self-reported overall mental health “very good” or
“excellent” .............................................................. 121 
62. Weighted means and parameter estimates of the
effect of SVORI on self-reported recidivism
outcomes ............................................................... 122 
63. Self-reported no perpetration of violence since
release/last interview ............................................... 123 
64. Self-reported complying with conditions of
supervision since release/last interview ...................... 124 
65. Self-reported not committing any crimes since
release/last interview ............................................... 125 
66. Self-reported not committing any violent or
weapons crimes since release/last interview ................ 126 
67. Not reincarcerated at follow-up interview .................... 127 
68. Not booked or reincarcerated since release/last
interview ................................................................ 128 
69. Official measures of recidivism .................................. 129 
70. Months to first rearrest............................................. 130 
71. Months to first reincarceration ................................... 130 
A-1. Adult female case disposition—Wave 1 (pre-release) .... A-1 
A-2. Respondent characteristics, by group ......................... A-2 

xiii

A-3. Proportion of respondents who reported needing
specific services, by group ...................................... A-11 
A-4. Proportion of respondents who reported receiving
specific services, by group ...................................... A-12 

xiv

Executive Summary
The Serious and Violent Offender Reentry Initiative (SVORI)
funded agencies in 2003 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. Across the grantees, programming was
provided to adult males, adult females, and juveniles.
The SVORI Multi-site Evaluation was funded by the National
Institute of Justice in the spring of 2003; it 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 programs only unless otherwise stated):
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 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.
This report presents findings for the female participants in the
impact evaluation, including 153 SVORI participants and 204
comparison women from 11 of the 12 adult impact sites. 1 After
1

Because the Maryland SVORI program served only men, no women
were recruited from that site.

ES-1

Prisoner Reentry Experiences of Adult Females

a brief review of the literature on women and reentry, a
description of the programming delivered to the women
through SVORI funding, and a summary of the methods with
which the impact evaluation was conducted, data are presented
on the pre- and post-release characteristics and experiences
among the female participants.
The data presented in the pre-release section of this report,
which are based on the interview conducted 30 days (on
average) before release, are primarily descriptive; they convey
characteristics of the respondents, as well as their preprison
and incarceration experiences. In addition to providing
descriptive information, the pre-release section assesses the
comparability of the SVORI and non-SVORI groups, examining
whether the SVORI participants received more services than
non-SVORI participants during their incarcerations. These prerelease findings on service receipt therefore constitute an initial
assessment of whether SVORI funding, compared with
“treatment as usual,” increased women prisoners’ access to
pre-release services.
The post-release section of the report, which is based on the
interviews conducted 3, 9, and 15 months after release,
describes the post-release experiences among the women,
assesses whether SVORI participants continued to receive more
services than comparable women not enrolled in SVORI during
the post-release follow-up period, and examines differences
between the groups on a variety of outcomes. In the postrelease section, weighted outcome analyses (which adjust for
selection into the SVORI programs) examine the treatment
effect of SVORI. Both the pre- and the post-release sections
highlight gender differences based on comparisons of the entire
female and male subsamples (using unweighted t-tests).
The implications of the findings for policy and practice
pertaining to female returning prisoners conclude the report.
Pre-release Characteristics of the SVORI and Non-SVORI
Female Respondents
Demographics
ƒ

ES-2

Female respondents, on average, were 31 years old at
the pre-release interview, with nearly equal numbers
self-identifying as white (44%) and black (41%).

Executive Summary

ƒ

Approximately 62% of the women reported having a
high school diploma or General Educational Development
(GED) credential.

ƒ

While 41% of the women reported living in their own
homes during the 6 months before incarceration, more
than one fifth were homeless, living in a shelter, or
without a set place to live.

ƒ

The vast majority of women were mothers; more than
half of those with minor children reported that they had
primary care responsibilities before incarceration.

Substance Use and Physical and Mental Health
ƒ

Nearly all women reported having used alcohol and
marijuana during their lifetimes, and three-quarters
reported cocaine use. More than two thirds of the
women reported having used one or more illicit drugs
during the 30 days before incarceration.

ƒ

Women reported many physical and mental health
problems; at the time of the pre-release interview,
fewer than half rated their physical health and fewer
than one third rated their mental health as excellent or
very good.

ƒ

Half of the women reported receiving treatment for
mental health problems before their current
incarceration.

Employment History and Financial Support
ƒ

Most women reported having worked at some point
during their lifetimes; more than half reported working
during the 6 months before prison.

ƒ

Of those who worked during the 6 months before prison,
about three quarters reported that their most recent job
was permanent and that they received formal pay.

ƒ

Nearly half of the women reported receiving income
from illegal activities, with those lacking a job before
prison being more likely to report illegal income.

Criminal History
ƒ

The women reported an average of 11 arrests, with the
first arrest occurring, on average, at 19 years of age.

ƒ

Nearly all women reported at least one previous
incarceration; one third had been detained in a juvenile
facility.

ƒ

At the time of the interview, women reported an
average length of incarceration of less than 2 years.

ES-3

Prisoner Reentry Experiences of Adult Females

Levels of Service Needs
ƒ

The women reported very high levels of current (30 days
before release, on average) service need across the 29
services addressed in the interview; on average, women
reported needing nearly two thirds of the services.

ƒ

Service need was substantially higher for women than
for men.

ƒ

The needs the women most commonly reported were
education (95%), public health insurance (91%),
financial assistance (87%), employment (83%), and a
mentor (83%).

ƒ

The SVORI and non-SVORI respondents reported similar
levels of need for most services, indicating that the two
groups were similar on service need at the time of the
pre-release interview.

Levels of Service Receipt
ƒ

Participants in SVORI programs had greater access to a
wide range of pre-release services and were more likely
to receive most of the documented service areas
(“bundles”).

ƒ

The most common services SVORI respondents reported
receiving during their incarceration were participating in
programs to prepare for release, taking a class
specifically for release, working with someone to plan for
release, receiving a needs assessment, and developing a
reentry plan.

ƒ

The women enrolled in SVORI reported substantially
higher levels of service receipt than the non-SVORI
respondents across 22 of the 36 services.

ƒ

Overall, SVORI respondents reported receiving about
half of the service items—in contrast to the one quarter
that non-SVORI respondents reported receiving.

The results from the pre-release interviews show that the
SVORI and non-SVORI groups were similar on most background
characteristics and largely similar on self-reported service need,
which was extremely high among the women. Women who
participated in SVORI programming were more likely to receive
pre-release programming and services, which indicates that the
SVORI funding did increase access to services for female
prisoners returning to society.

ES-4

Executive Summary

Post-release Experiences Among the SVORI and NonSVORI Female Respondents
Levels of Service Needs
ƒ

Compared with the extremely high self-reported service
needs at the pre-release interview, women reported
needing substantially fewer services at the 3-month
post-release interview. Need for services continued to
decline at the 9- and 15-month interviews, as well.
Nonetheless, absolute levels of service need remained
quite high. Women continued to report high levels of
service need (in the 40–50% range) for many services
even 15 months after release.

ƒ

The needs most commonly reported by the women
across all 3 follow-up waves were more education (87–
93%), public health care insurance (66–77%), and
financial assistance (64–73%).

ƒ

The SVORI and non-SVORI groups reported similar
levels of need for most services, indicating that the
needs of the two groups were comparable at each
follow-up time point.

ƒ

As with the findings at the pre-release interview, service
need was significantly higher for women than it was for
men across several service areas (primarily health
services and family services), at each time period.

Levels of Service Receipt
ƒ

As with the decline of self-reported need for services
over time, the likelihood of receiving services declined
over time.

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 very small proportions of women received
the services they needed.

ƒ

The women enrolled in SVORI programming reported
substantially higher rates of services receipt than the
non-SVORI respondents at each of the post-release
interviews. Similar to the overall trend in declining
service receipt over time, the number of differences
between these groups decreased at each follow-up
wave.

ƒ

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

ES-5

Prisoner Reentry Experiences of Adult Females

ƒ

Women reported much higher levels of service receipt
than men during the post-release period, particularly at
the 3- and 15-month time periods.

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.

ƒ

For both the SVORI and non-SVORI groups, housing
situations improved gradually over the 3-, 9-, and 15month post-release time periods.

ƒ

Several gender differences in housing were identified,
including women’s being significantly more likely than
men to report

ƒ

–

living in their own house or apartment (at the 3- and
9-month post-release time periods),

–

being homeless (at the 3- and 9-month post-release
time periods), and

–

living with their children (at all time periods).

Men were significantly more likely than women to report
contributing to housing costs and to report living with
their mothers or stepmothers, sisters, and brothers (at
all time periods).

Employment
ƒ

ES-6

SVORI programming appeared to positively affect many
dimensions of employment, with effects being strongest
for the time period reflecting 15 months post-release.
Compared with the non-SVORI respondents, women
who enrolled in SVORI programming
–

were more likely to report supporting themselves
with a job (at 15 months post-release),

–

worked significantly more months (at 15 months
post-release),

–

worked significantly more months at the same job
(at 15 months post-release),

–

were more likely to receive formal pay for their jobs
(at all time periods), and

–

were less likely to report receiving money from
illegal activities (at 3 months post-release).

Executive Summary

ƒ

Overall, women fared worse than men on most
dimensions of employment, including likelihood of
working, number of months worked, and likelihood of
working at jobs that offered benefits.

Family, Peers, and Community Involvement
ƒ

ƒ

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

When several dimensions of familial relationships
were examined, including emotional support,
instrumental support, quality of intimate-partner
relationships, and quality of relationships with
children, no differences between the SVORI and nonSVORI respondents were observed.

–

The SVORI and non-SVORI groups reported similar
levels of negative peer exposure. The levels of
instrumental support from peers were significantly
higher for the SVORI group than for the non-SVORI
group at both the 9- and the 15-month time periods.

–

SVORI participants reported marginally higher levels
of civic action than the non-SVORI group (p < 0.10)
at the 3- and 15-month time periods.

Overall, levels of familial and peer support (based on
scales measuring both emotional and instrumental
support) were similar for men and women. Women
reported higher-quality intimate partnerships at the 3month post-release time period—a time period at which
men were significantly more likely to live with their
spouse or romantic partner. At all time periods, women
were significantly more likely than men to report having
primary care responsibilities for their children and had
higher scores on the scale assessing the respondent’s
relationship with children.

Substance Use
ƒ

SVORI appeared to have a positive impact on abstinence
from drug use. Results for a composite self-report and
oral fluids drug tests outcome measure indicated that
the SVORI participants were significantly less likely than
the non-SVORI participants to have used drugs during
the reference period and during the 30 days before the
interview, at both the 3- and 15-month post-release
time periods.

ƒ

Overall, women’s substance use increased across the
post-release follow-up periods; levels of use did not
differ between women and men.

ES-7

Prisoner Reentry Experiences of Adult Females

Physical and Mental Health
ƒ

SVORI did not appear to influence physical or mental
health outcomes. No differences were observed, at any
time period, between the SVORI and non-SVORI groups
on the 12-Item Short Form Health Survey (SF-12; Ware,
Kosinski, Turner-Bowker, & Gandek, 2002) physical
health scale, the number of physical conditions
experienced, overall perceptions of physical health, the
SF-12 mental health scale, the Global Severity Index, or
overall perceptions of mental health.

ƒ

Significant gender differences were observed for all of
the measures of mental and physical health at all followup time periods, with women consistently faring worse
than men.

Criminal Behavior and Recidivism
ƒ

ƒ

The findings for criminal behavior and recidivism were
mixed. The women enrolled in SVORI had positive
outcomes (i.e., lower criminality/recidivism) for
–

self-reported perpetration of violence (15 months
post-release),

–

self-reported criminal behavior (excluding violent
and weapons crimes; 15 months post-release), and

–

official measures of rearrest (within 9, 12, 15, and
21 months of release).

The women enrolled in SVORI had negative outcomes
(i.e., higher criminality or recidivism) for
–

self-reported compliance with conditions of
supervision (at 9 months post-release) and

–

official measures of reincarceration in state prisons
(within 12, 15, 21, and 24 months of release).

Conclusions
The findings reported here clearly demonstrate that female
prisoners returning to society are a population with high needs.
SVORI funding offered correctional agencies an opportunity to
intervene by providing a range of services designed to facilitate
successful reentry for prisoners. The programs were extremely
successful in increasing the services provided to female
participants. Across almost all types of services, the women
who participated in SVORI reported substantially higher levels
of service receipt than comparable women not enrolled in
SVORI. In addition, although the SVORI programs focused
resources on increasing women’s access to services during

ES-8

Executive Summary

incarceration, the higher levels of service receipt found for
SVORI participants persisted even after their release (with
significantly higher levels of service receipt documented at 3, 9,
and 15 months post-release). Importantly, however, the levels
of services that female SVORI participants received, although a
significant improvement over “treatment as usual,” failed to
match their high levels of need.
Even though service receipt was insufficient to meet the
women’s high needs, the enhanced service delivery that SVORI
programs provided appears to have improved participants’
reentry outcomes in several dimensions. Employment and
substance abuse were the domains for which the most
consistent program effects were observed; the outcomes were
much less clear for criminal behavior and recidivism. Women
who participated in SVORI had positive outcomes for several
dimensions of criminal behavior and recidivism, including lower
likelihood of rearrest. According to official corrections data,
however, SVORI participants had significantly higher
reincarceration rates. Two possible explanations for this pattern
are (1) site-specific effects (because the women were not
evenly distributed by group across sites and site-level practices
or policies may influence reincarceration rates) and (2) that
SVORI program participants were more likely than comparison
subjects to have been at risk for post-release supervision
revocation—because either they were more likely to be on
supervision, or they were subject to more conditions of
supervision.
The findings support the notion that enhanced access to a
variety of reentry services results in modest improvements
among several key reentry domains for women. With these
outcomes, the reentry efforts initiated through SVORI funding
provide a promising foundation for future reentry efforts.

ES-9

Introduction
SVORI was a collaborative federal effort, established in 2003, 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. Funded by the U.S.
Departments of Justice, Labor, Education, Housing and Urban
Development, and Health and Human Services, 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.
SVORI funding was intended 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
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,
Lattimore, Steffey, Brumbaugh, & Lindquist, 2006; Winterfield
& Lindquist, 2005).

1

Prisoner Reentry Experiences of Adult Females

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 a comprehensive 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 (see Lattimore et al., 2004). 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 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
with those of a comparable group of individuals who received
“treatment as usual” in the 16 sites participating in the impact
evaluation. The impact evaluation included a longitudinal study
of 2,391 returning prisoners (adult males, adult females, and
juvenile males) who were interviewed approximately one month
before release and then again at 3, 9, and 15 months after
release. 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.
This report presents findings for the female participants in the
impact evaluation, which included 153 SVORI participants and
204 comparison women returning from prison in 11 states. The
data presented in the pre-release section of this report, which
are based on the interview conducted 30 days (on average)
before release, are primarily descriptive; they convey
characteristics of the respondents, as well as their preprison
and incarceration experiences. In addition to providing
descriptive information, the pre-release section assesses the
comparability of the SVORI and non-SVORI respondents,
examining whether the SVORI participants received more
services than non-SVORI participants during their

2

Introduction

incarcerations. These pre-release findings on service receipt
therefore constitute an initial assessment of whether SVORI
funding, compared with “treatment as usual,” increased women
prisoners’ access to pre-release services.
The post-release section of the report, which is based on the
interviews conducted 3, 9, and 15 months after release,
describes the post-release experiences among the women,
assesses whether SVORI participants continued to receive more
services than comparable women not enrolled in SVORI during
the post-release follow-up period, and examines differences
between the groups on a variety of outcomes. In the postrelease section, weighted outcome analyses (which adjust for
selection into the SVORI programs) examine the treatment
effect of SVORI. Both the pre- and the post-release sections
highlight gender differences based on comparisons of the entire
female and male subsamples (using unweighted t-tests).
In the sections that follow, the literature on women and reentry
is briefly reviewed (“Previous Research on Women and
Reentry”) and the programming delivered to the women
through SVORI funding is described (“An Overview of SVORI
Programming for Female Offenders”). “The SVORI Multi-Site
Evaluation—Design and Methods” section summarizes the
methods for the impact evaluation, including the selection of
respondents, the interview process, the technique applied to
ensure comparability between the SVORI and non-SVORI
samples, and the manner in which selection and attrition bias
were addressed. In the “Pre-release Experiences of Returning
Female Prisoners,” detailed findings on the pre-release
characteristics of the women are presented, including the
demographic characteristics of the women, their preprison
experiences, self-reported service needs, and in-prison service
receipt. In the “Post-release Experiences of Returning Female
Prisoners” section, the findings for service need and receipt
during the post-release time periods are presented, in addition
to the impact results for a variety of reentry outcomes. A
discussion of policy implications concludes the report.

3

Previous Research
on Women and
Reentry
WOMEN PRISONERS: PREVALENCE AND
NEEDS
The U.S. Bureau of Justice Statistics reports that 208,300
women were incarcerated in state and federal prisons or local
jails in June 2007 (Sabol & Couture, 2008). Women represent
approximately 10% of the total incarcerated population and
have notably different circumstances and needs than
incarcerated men. Female offenders are more likely to
participate in property and drug offenses than males, who are
more likely to participate in violent offenses (West & Sabol,
2008). Morash et al. (1998) note that female offenders are
more likely to have been victims of sexual and physical abuse.
Moreover, research indicates that female offenders are more
likely than male offenders to battle drug addiction and suffer
from mental illness (Covington, 2003; Morash et al., 1998).
Indeed, in a study of reentering prisoners, it was estimated that
77% of women and 54% of men had chronic physical and
mental health conditions at the time of release from prison;
women were more likely to have comorbid mental health and
substance abuse conditions (Mallik-Kane & Visher, 2008). The
same study nonetheless found that men with substance abuse
disorders were more likely to receive tangible assistance from
family than other men, whereas substance-abusing women
were less likely to receive such help than other women.
Female inmates are also more likely to report being a parent
than men (62% and 51%, respectively) and to having more

5

Prisoner Reentry Experiences of Adult Females

than one child (41% and 29%, respectively; Glaze, 2008).
Additionally, women who are released from prison appear to
experience more housing challenges, lower employment, less
family support, and more substance abuse than men (MallikKane & Visher, 2008).

REENTRY EXPERIENCES AMONG WOMEN
Because of their high level of needs, female offenders
encounter unique obstacles on reentry into the community. In
addition, women may have different concerns and priorities
with regard to reentry than men. Research has documented
that the primary reentry concerns for female offenders are to
successfully reunite with their children, maintain a suitable
lifestyle, and sustain relationships with family and intimate
partners (La Vigne, Brooks, & Shollenberger, 2008; O’Brien,
2001; Richie, 2001). In an analysis of female offenders
returning to Houston, Texas, communities (La Vigne et al.,
2008), a majority of the women were found to be highly
optimistic, especially about reuniting with their families and
children. Many women expected to have emotional and financial
support from their families and relationships upon release.
LaVigne et al. (2008) report that the expectations women had
in reuniting with their families were met in the first 8 to 10
months after their release, which contributed to their successful
reintegration into the community.
Research has also found that female offenders prioritize finding
suitable housing and employment upon their release, which can
assist in their personal recovery (O’Brien, 2001). O’Brien also
reports that former female offenders want to feel a sense of self
confidence or resiliency upon their release.
Richie (2001) has found that having access to sufficient health
care to cover medical needs is also a concern for female
offenders. Whether women have more successful reentry
experiences than men is unclear. In addition, the factors that
contribute to successful reintegration for men and women have
not been conclusively identified. In a review of 32 studies
evaluating prisoner reentry programs, Seiter and Kadela (2003)
found that vocational training or work release programs, drug
rehabilitation, halfway house programs, and pre-release
programs were associated with reductions in recidivism.
Additionally, drug rehabilitation programs were found to be

6

Previous Research

successful at reducing drug use, and education programs were
shown to increase achievement scores. However, this review
did not distinguish whether these studies included both men
and women or include a discussion of any differences in
program effects by gender. There is limited research examining
correlates of successful reentry specifically among female
offenders.
However, one factor that has been found to be responsible for
the successful reintegration of former female offenders is the
existence of community programs that foster a positive
transition (O’Brien & Harm, 2002). Similarly, LaVigne et al.
(2008) have reported that former female offenders often
describe the importance of having community programs and
supervision that emphasize substance abuse treatment,
employment-based skills, counseling, and job training (La Vigne
et al., 2008). Bloom, Owen, and Covington (2003) stress the
importance of comprehensive, collaborative, and well
coordinated services.
The limited research on gender differences and desistance has
found female offenders more likely than male offenders to
credit their children and religion for changes in their behavior;
male offenders reported the prison environment, treatment,
and their family as the primary reasons for their desistance
(Giordano, Cernkovich, & Rudolph, 2002).

REENTRY PROGRAMMING FOR WOMEN
Although many programs are developed to serve the needs of
both men and women, female offenders arguably have unique
needs and circumstances that require some gender-specific
programming (Koons, Burrow, Morash, & Bynum, 1997).
According to Bloom et al. (2003, p. vii), “gender-responsive
policy and practice target women’s pathways to criminality by
providing effective interventions that address the intersecting
issues of substance abuse, trauma, mental health, and
economic marginality.” In some correctional settings,
rehabilitative, educational, and reintegration programs that
specifically target women offenders are offered. For example,
the Women’s Prison Association (WPA), which is stationed in
New York, is a core agency dedicated to supporting,
rehabilitating, and educating female offenders. The WPA
provides programming services that include residential and

7

Prisoner Reentry Experiences of Adult Females

family services, as well as reentry services designed for
incarcerated mothers. Rather than incarceration, some
correctional agencies argue that there should be other options
available to women offenders, especially those who are
mothers.
The WPA, for example, provides the Hopper Home Alternative
to Incarceration (ATI) program. This program is specifically
designed to allow female offenders to complete court-mandated
programs, which can include drug treatment and anger
management training, in the community as opposed to in a
prison setting. The WPA also provides services to female
offenders who wish to regain custody of their children through
the Sarah Powell Huntington House Family Reunification
Residence (SPHH), Incarcerated Mothers Law Project (IMLP),
and the Family Preservation Program. Moreover, the WPA also
establishes services that assist these women in finding suitable
housing through their Sunflower House program.
The SPHH, for instance, targets homeless women who have a
criminal record. More specifically, SPHH helps women regain
custody of their children by assisting in their quest to find
appropriate housing. The Sunflower House program allows
former offenders to assist each other in their recovery process
from substance addictions.
The IMLP and Family Preservation Program provide services to
women offenders who are concerned with the welfare of their
children. More specifically, the purpose of the IMLP is to inform
incarcerated mothers of their parental rights pertaining to the
institutional custody of their children, whereas the Family
Preservation Program employs an intensive case management
model that is designed to treat drug-addicted mothers through
substance abuse treatment and mental illness rehabilitation.
The Family Preservation Program also assists these mothers in
locating housing, obtaining an education, and gaining
employment.
Other programs, such as the Children’s Center at Bedford Hills
Correctional Facility, provide nurseries for babies of
incarcerated women and help teach them parenting (Morash et
al., 1998). A national survey on promising correctional
programs for incarcerated women found that the most common
programmatic focus areas were substance abuse education and
treatment, parenting, life skills, relationships, and education.

8

Previous Research

Both program administrators and participants attributed the
success of programs to characteristics of the staff members
(e.g., personal experience with addiction and caring attitudes)
and to the women’s acquisition of a comprehensive skill set
related to both work and family responsibilities (Koons et al.,
1997).
On the basis of the limited research of former female prisoners
and even more limited evaluations of reentry programs, the
factors that appear most important for successful reentry for
women include establishing suitable housing, finding gainful
employment, and reuniting with children and family. Moreover,
women’s reentry programs need to emphasize post-release
treatment and counseling for infectious diseases and substance
abuse as well as protection from abusive relationships.

9

An Overview of
SVORI
Programming for
Female Offenders
The SVORI funding stream provided an opportunity for state
and local agencies to develop reentry programming for women
(in addition to men and juveniles). The federal guidance
accompanying SVORI funding placed few restrictions on the
state agencies with respect to the design of the individual
SVORI programs. The primary restrictions placed on local
SVORI programs were an age limit—the programs were
required to target prisoners aged 35 years or younger—and a
requirement for post-release community supervision. Other
broad requirements were that the program include three
phases (in-prison, supervised post-release, and postsupervision); provide holistic case management and service
delivery; improve participants’ quality of life and self-sufficiency
through employment, housing, family, and community
involvement; improve participants’ health by addressing
substance use and physical and mental health; and reduce
participants’ criminality through supervision and monitoring of
noncompliance. The programs also were encouraged to include
needs and risk assessments, reentry plans, transition teams,
community resources, and graduated sanctions (see Winterfield
et al., 2006). Because a SVORI program model was not
specified, each program was locally designed; therefore, the
programs varied considerably in approach, services provided,
and target populations.

11

Prisoner Reentry Experiences of Adult Females

Through the SVORI implementation assessment, program
directors from each of the 52 adult programs were surveyed
about their program goals, as well as the types of services they
provided and the proportion of offenders—both SVORI and nonSVORI— who received these services before and after release.
Because this assessment was not designed to capture gender
differences in programming, little is known about the extent to
which SVORI programming differed by gender across the 89
SVORI programs.
Information on the reentry services and programming
implemented appears elsewhere (Lattimore, Visher, & Steffey,
2008; Lindquist, 2005; Winterfield & Brumbaugh, 2005;
Winterfield et al., 2006; Winterfield & Lindquist, 2005).
Descriptions of SVORI programming at each impact site can be
found in Lattimore and Steffey (2009). Overall, the evaluation
documented that employment and community integration
tended to be the primary focus of SVORI programs and that,
according to information provided by program directors, higher
proportions of SVORI participants than comparable offenders
not participating in SVORI were receiving nearly all of the
services available.
From the programs selected for the impact evaluation, more
detailed documentation on programming was attained during
evaluation staff’s impact site visits. This documentation enabled
a comparison of SVORI programming designed for men with
that designed for women. In sites that assigned SVORI
participants to case managers or parole officers dedicated to
SVORI participants, women were often assigned to a particular
case manager or parole offer. In addition, in sites that
transferred SVORI participants to community corrections
facilities, the men and women were, not surprisingly,
transferred to separate facilities. Two sites noted the
availability of post-release housing specifically for women (in
one site this option was limited to women with children).
Overall, however, the evaluation documentation suggests that
the SVORI programs were not designed to deliver genderspecific programming. For example, among the 11 impact sites
that serve adult females, none of the sites appeared to
differentiate between men and women in their overall approach
to reentry programming.

12

Overview of Programming

In many ways, the lack of gender-specific programming among
the SVORI programs is to be expected because of the small
number of women enrolled in them. Overall, among the SVORI
programs that participated in the impact evaluation, only 15%
of the SVORI participants enrolled in the impact study were
women. This figure is only an estimate of the gender
distribution of the SVORI programs selected for the impact
evaluation (because it reflects cases that were actually fielded
and resulted in a completed pre-release interview), but it is
consistent with the impact sites.
Stakeholders from some sites reported that the upper age limit
imposed by the federal funding requirements (35 years of age)
severely restricted the number of women they were able to
enroll. One site received permission to establish a higher upper
age limit for women than the one used for men. It is clear,
however, that across impact sites customized programming for
women was not emphasized.
Interestingly, although no differences were evident in the
design of SVORI programming for men and women, information
from the site visits does suggest gender differences in overall
pre-release programming availability (offered as “treatment as
usual” in the male and female facilities). Specifically,
documentation of the services available to the general prison
population at the pre-release institutions served by the SVORI
programs suggests that, in general, more programs and
services were available at the women’s prisons than at the
men’s prisons. Services such as substance abuse treatment,
mental health treatment, mentoring programming, life skills
programming, and domestic violence services appeared to be
more commonly available at the women’s prisons.2

2

In at least one site, however, a service available at several men’s
prisons (video-conferencing) was not available at the women’s
prison served by the SVORI program.

13

The SVORI Multisite Evaluation—
Design and Methods
Here the methods employed in the SVORI Multi-site Evaluation
are summarized. A detailed description of the design, data
collection procedures, instrumentation, and analytic strategy
can be found in Lattimore and Steffey (2009).

STUDY DESIGN
The impact evaluation included a longitudinal study of adult
male, adult female, and juvenile male prisoners returning to
their communities. 3 On the basis of an extensive site selection
process, 16 programs were chosen (from among the 89 SVORI
programs) for the impact study, with the objective of achieving
diversity in programmatic approach and geographical
representation. The 16 programs included 12 adult programs
and 4 juvenile programs located in 14 states: 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 adult female sample reflected
women enrolled in Indiana, Iowa, Kansas, Maine, Missouri,
Nevada, Ohio, Oklahoma, Pennsylvania, South Carolina, and
Washington. As shown in Exhibit 1, the women were unevenly
distributed across group and site, with half of the non-SVORI

3

Juvenile females were excluded from the impact evaluation because
of the extremely small number of SVORI participants in this
subgroup.

15

Prisoner Reentry Experiences of Adult Females

Exhibit 1. Adult female
sample sizes, by state
and group

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

SVORI
35
12
17
7
22
9
15
3
6
24
3
153

Non-SVORI
3
101
31
2
0
8
12
7
0
31
9
204

Total
38
113
48
9
22
17
27
10
6
55
12
357

% of Total
10.6
31.7
13.5
2.5
6.2
4.8
7.6
2.8
1.7
15.4
3.4
100.0

respondents residing in Indiana. 4 Other sites provided very few
total cases or no non-SVORI respondents. This uneven
distribution of cases across group by site limits the possibilities
of addressing the impact of site in the outcome analyses and
raises concerns of the potential for undue influence by a group
on site, independent of effect.
A site-specific research design was developed for each impact
site. In two sites (Iowa and Ohio), a random-assignment
evaluation design was implemented by the programs. 5 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.
In some instances, comparison participants were identified as
those who met all eligibility criteria except pre- or post-release
4

5

16

Indiana had a surplus of eligible comparison participants who were
interviewed in the expectation that, if shown to be comparable to
the women in other states, they could compensate for the deficit of
comparison women in other states.
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 very few comparison women
were enrolled in that site).

Design and Methods

geographic parameters. When this exception occurred, either
the comparison sample was selected from pre-release facilities
that were comparable to facilities in which SVORI was available,
or individuals from SVORI facilities who were returning to a
separate but similar geographic area were selected. Eligible
respondents (both SVORI and comparison) were identified
monthly during the 17-month enrollment period for the impact
evaluation.

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 adult
respondents who were interviewed in a community setting. For
examination of recidivism outcomes, the interview and drug
test data were supplemented with arrest data and with
administrative records obtained from state correctional
agencies.
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 were
designed to obtain 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 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
extensive instrumentation process involving substantive domain

17

Prisoner Reentry Experiences of Adult Females

experts and the use of existing, validated measures and scales
used in previous RTI and Urban Institute studies.
In addition to obtaining approval from 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).

APPROACH FOR ADDRESSING
NONRESPONSE AND ATTRITION
A total of 516 women were eligible to be included in the study.
Completed Wave 1 (pre-release) interviews were obtained with
69% of the women. Among eligible sample members
approached for interviews, refusal rates were very low: on
average, 7% across the 11 adult female sites. A breakdown of
the categories of refusals and ineligible cases is available in
Appendix Exhibit A-1. As shown in the exhibit, most of the
noninterviews among eligible women were due to the women’s
being released before their Wave 1 interview could be
completed.
Nearly 90% of the women who were interviewed at Wave 1 also
responded to at least one of the follow-up interviews. The
response rates for the Wave 2, 3, and 4 interviews were 68%,
71%, and 77%, respectively.
Although the response rates for the women were fairly high,
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 nonSVORI groups similarly, at each wave the SVORI and
comparison groups were found to be similar at each wave on a
range of characteristics. Unfortunately, the relatively small
sample size precluded for the women the more rigorous
examination of nonresponse conducted for the men (see
Lattimore & Steffey, 2009). Diagnostic tests for response bias
in the male sample did not indicate any problems. These
results, combined with the higher response rates found, at each
wave, in the female sample as compared with the male sample,
and with the comparability between groups across waves,

18

Design and Methods

suggest that attrition did not introduce any substantial problem
into the data on women prisoners.

APPROACH FOR ADDRESSING SELECTION
BIAS
In addition to limitations posed by attrition, the potential for
selection bias must be examined because, in most sites, women
were not randomly assigned to SVORI or non-SVORI conditions.
On initial examination, the raw data showed that the SVORI
participants were more likely to be incarcerated for a violation
of parole than their non-SVORI counterparts (see Lattimore &
Steffey, 2009). Differences in age, race, and prior mental
health treatment approached significance.
Propensity score models were then used to improve the
comparability between the SVORI and non-SVORI groups. This
technique uses observed characteristics to model the likelihood
that an individual with those characteristics will be selected (or
assigned) to the intervention. One limitation of propensity score
models is the possibility that a variable related to both the
treatment and outcome, and thus potentially responsible for
any observed treatment effects, is omitted. However, this is an
accepted methodology that has been used in hundreds of
research studies in a variety of fields (Rubin, 2006).
In the first step, a logit model to generate the probability of
assignment to SVORI was estimated with 24 variables
measured before SVORI assignment, including characteristics
such as age, race, criminal history, and employment before
incarceration. Propensity score weights were developed to
examine balance (and program effects). Once the propensity
score weights were applied, the SVORI and non-SVORI
respondents exhibited balance on each variable included in the
propensity model, conferring confidence that the groups are
indeed comparable and permitting examination of the effect of
SVORI on outcomes measured in the follow-up interviews.
As an additional check, differences between the SVORI and
comparison groups on these 24 Wave 1 characteristics were
examined at each follow-up wave. At the 3-month interview,
the difference in incarceration rate (measured at the time of
pre-release interview) was statistically significance (p < 0.05);
however, the SVORI and non-SVORI respondents did not differ
significantly on any other variables included in the propensity

19

Prisoner Reentry Experiences of Adult Females

model at any other wave. The results suggest that the
propensity score model provided balance across all four waves
of interview data. For a more detailed discussion of this
approach, see Lattimore and Steffey (2009).

20

Pre-release
Experiences of
Returning Female
Prisoners
Provided here is descriptive information about the 357 adult
female SVORI and non-SVORI respondents interviewed in the
11 adult impact sites in which women were enrolled.
Background characteristics of the women are summarized, in
addition to their preprison and incarceration experiences. These
data allow us to assess the comparability of the SVORI and
non-SVORI respondents.
This section presents detailed information on the service needs
reported by the women and examines the extent to which the
women received a variety of services during their
incarcerations. These data constitute an initial assessment of
program implementation, in that service receipt reported by
women who were enrolled in the SVORI programs can be
compared with that of comparable women who received
“treatment as usual.”
The data are presented for the total sample of women, as well
as the SVORI and non-SVORI groups. Also discussed are key
gender differences observed, although data for men are not
presented (for complete pre-release data on the male sample,
see Lattimore et al., 2008). Although a gender comparison is
not the explicit purpose of this evaluation, these differences are
presented to take advantage of the unique opportunity of
having comparable measures for both the male and female
samples. Because the pre-release data are used entirely for
descriptive purposes in this section, the data here are

21

Prisoner Reentry Experiences of Adult Females

unweighted. As will be discussed in the “Post-release
Experiences of Returning Female Prisoners” section, weighting
for selection bias is necessary to examine actual program
effects among the SVORI and non-SVORI respondents.
Appendix Exhibit A-2 provides the means, standard deviations,
and t-statistics for all the variables discussed in this section.

BACKGROUND CHARACTERISTICS
The subsections on background characteristics provide
descriptive information on demographic characteristics;
housing; family and children; substance use and physical and
mental health; employment history and financial support;
criminal history, violence, victimization, and gang involvement;
and in-prison experiences.
Demographic Characteristics
Almost all of the women in the sample were born in the United
States (99%) and reported that English was their primary
language (96%), with no significant differences appearing
between the SVORI and non-SVORI groups. The racial/ethnic
breakdown of the female sample is shown in Exhibit 2, with
approximately 44% of the sample being white and no
significant differences emerging between the SVORI and nonSVORI groups. The women reported an average age of 31
years. Not surprisingly, given the demographic composition of
the prison population as a whole, the female sample was
significantly (p < 0.05) older than the adult male sample
(which had a mean age of 29 years). Slightly more than 62% of
the women had at least a 12th grade education or GED, with a
significantly higher proportion of the SVORI group falling into
this classification (71%, as opposed to 55%).
Housing
When asked about their housing situation during the 6 months
before incarceration, more than one fifth of the women (22%)
reported as their primary housing situation during that time
period that they were homeless, living in a shelter, or had no
set place to live. This percentage is significantly higher than
that for adult males, among whom only 13% reported being
homeless, living in a shelter, or being without a set place to live
(for gender difference, p < 0.05). Interestingly, women were
more also more likely than men to be living in their own house
or apartment (41%, as opposed to 34%; p < 0.05), whereas

22

Pre-release Experiences

Exhibit 2. Demographic
characteristics of
respondents at time of
interview, by group

Variable
Race
White
Black
Hispanic
Other race
Age
Age at Wave 1 interview
Education
12th grade/GED*

SVORI

Non-SVORI

All

48%
35%
8%
9%

41%
45%
5%
10%

44%
41%
6%
10%

31

32

31

71%

55%

62%

Note: GED = General Educational Development.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

men were more likely than women to report primarily living in
someone else’s house or apartment (47% of men and 34% of
women, p < 0.001). This pattern suggests that, before
incarceration, women were less likely to have friends and family
members who would assist with housing. Consequently, women
who lacked their own housing ended up homeless, living in a
shelter, or being without a set place to live, whereas men
without their own housing ended up living with someone else.
Family and Children
Fourteen percent of the women reported being married at the
time of the pre-release interview—a significantly higher
percentage than among the men (10%; for gender difference, p
< 0.05). Slightly less than half of the female sample (48%)
were either currently married or in a steady relationship, which,
once again, is a significantly higher percentage than that
observed among the male sample (40%; p < 0.05). Among the
women who currently had a romantic partner, about three
quarters (74%) reported living with that person before their
incarceration. No gender differences in the likelihood of living
with the current romantic partner before incarceration were
evident, nor were there any differences in these variables
between the SVORI and non-SVORI female respondents.
The great majority of women (84%) had children, compared
with 63% of men (p < 0.05). Among mothers, the average
number of children was 2.8 (which was significantly higher than
the 2.3 reported by fathers; p < 0.05). As shown in Exhibit 3,
among the mothers who reported having children younger than
18 years of age, 54% reported that they had primary care

23

Prisoner Reentry Experiences of Adult Females

Exhibit 3. Percentages of mothers reporting on child care or child support responsibilities,
by group
100%

95%
91%

SVORI
Non-SVORI

85%

All
80%

60%

55%

54%

54%
45%
39%

40%
31%

20%

17%
11%

13%

0%
Primary care for children
under 18 a

Required to pay child
support a

Made required child support
paymentsb

Owed back child support b

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

responsibilities for any of their children during the 6 months
before incarceration (primary care being defined as the
children’s living with the respondent most of the time and the
respondent’s being responsible for feeding and clothing them).
Significantly more mothers than fathers with minor children
reported that they had primary care responsibilities for their
children before incarceration (54%, as opposed to 48%; p <
0.05) and, among those who had primary care responsibilities
for any of their children, mothers reported having responsibility
for significantly more children than fathers (2.15, as opposed to
1.75; p < 0.05). Interestingly, among respondents who
reported not having primary care responsibilities for any of
their children younger than 18, fathers reported that they
financially supported (in any way) significantly more children
than mothers during the 6 months before incarceration (1.14,
as opposed to 0.71; p < 0.05).

24

Pre-release Experiences

As shown in Exhibit 3, among mothers of children under 18,
only 13% reported that they were required by a court to pay
child support for any of their children during the 6 months
before incarceration. This percentage was much lower than the
comparable percentage of fathers (31%; p < 0.05). Similarly,
while only 39% of mothers who owed child support indicated
that they made court-ordered child support payments (during
the 6 months before incarceration), a significantly higher
proportion of fathers (57%) reported making such payments.
Of mothers and fathers who had child support requirements
during the 6 months before incarceration, the great majority
(91% for women and 92% for men) indicated that they owed
back child support, with an average amount of $6,687 for
mothers and $11,132 for fathers. While one third of mothers
who owed back child support owed more than $5,000, more
than half of fathers who owed back child support (58%) owed
more than $5,000 (for gender difference, p < 0.05). Among the
women, no significant differences between the SVORI and nonSVORI subgroups were evident for any of the child-related
variables.
Several dimensions of family support were captured in the prerelease interview. The adult females reported high levels of
family emotional support at the time of the pre-release
interview. For example, more than half of the women strongly
agreed with statements such as “I feel close to my family,” “I
have someone in my family to talk to about myself or my
problems,” and “I have someone in my family to turn to for
suggestions about how to deal with a personal problem.” More
than two thirds strongly agreed that they wanted their families
to be involved in their lives and that they had someone in their
families to love them. The items were combined to create a
family emotional support scale with possible values ranging
from zero to 30, with higher scores indicating higher levels of
support. The scores on this family emotional support scale
revealed no differences between male (21.4) and female (21.0)
respondents and no differences between the female SVORI
(21.1) and non-SVORI (20.8) respondents.
The role of family members as a positive influence becomes
questionable when one examines the criminal and substance
abuse involvement of respondents’ family members. As shown
in Exhibit 4, more than three quarters of the women (both
SVORI and non-SVORI) reported having family members who

25

Prisoner Reentry Experiences of Adult Females

Exhibit 4. Criminal history and substance use of family and peers, by group
100%
SVORI
Non-SVORI
All
80%

79%

83%

81%
77% 78%

78%

80%

81% 82%
78%

80% 79% 79%

78%

75% 76%

74%
70%

60%

40%

20%

0%
Family who had been Family who had been Family with drug or
convicted
incarcerated
alcohol problems

Friends who had
been convicted

Friends who had
been incarcerated

Friends with drug or
alcohol problems

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

had at some time been convicted of a crime, been incarcerated,
or had problems with drugs or alcohol. Similarly, the prevalence
of criminal behavior and substance abuse problems among
friends were high among the women. A large majority of
women in both groups reported having friends who at some
time had been convicted, had been incarcerated, or had
experienced drug or alcohol problems.
Interestingly, gender differences in criminal or substance abuse
behaviors among family and friends were evident (data not
shown). Although women were significantly more likely than
men to report having family members who had been
incarcerated (79% of women, as opposed to 74% of men; p <
0.05) and family members who had had problems with drugs or
alcohol (82% of women, as opposed to 73% of men; p < 0.05),
men were significantly more likely to report having friends who
had been convicted of a crime (83% of men, as opposed to
76% of women; p < 0.05) and friends who had been

26

Pre-release Experiences

incarcerated (81% of men, as opposed to 74% of women;
p < 0.05).
Substance Use and Physical and Mental Health
The pre-release interview elicited information about several
dimensions of pre-incarceration use of alcohol and drugs, in
addition to substance abuse treatment received. In addition,
respondents were asked about physical and mental health,
including medical diagnoses, health-related limitations
(capturing both physical and mental health–related limitations),
and mental health symptoms. Three widely used scales were
used, including the SF-12 physical health scale (Ware et al.,
2002), the SF-12 mental health scale, and the SA-45 Global
Severity Index (GSI; Strategic Advantages, 2000).
Substance Use and Treatment
Virtually all of the women reported at least some experience
with alcohol (96% had at some time drunk alcohol, and 53%
drank during the 30 days before incarceration). The average
age of first use was 14.5 years, which was significantly older
than that reported by the male sample (13.7). Interestingly,
the SVORI female participants had a significantly younger age
of first use than the non-SVORI women (13.9, as opposed to
14.9; p < 0.05).
Many women had used illicit substances at some point in their
lives. Exhibit 5 shows the prevalence of lifetime use for the
most common drugs. Most (90%) of the women reported
having used marijuana, and 75% had used cocaine. Forty-four
percent reported having at some time used hallucinogens, and
about a third reported that they had used tranquilizers (36%),
amphetamines (35%), and pain relievers (35%; without a
prescription or for reasons other than those for which the drugs
were prescribed, or in larger amounts or more often than the
respondent’s doctor ordered). Use of other substances was very
low. 6
A few differences between the SVORI and non-SVORI groups of
women were evident, with the lifetime prevalence being higher
for the SVORI participants for heroin and amphetamines and
the age of first use being significantly younger for the SVORI
group for cocaine.
6

Less than 10% of women reported use of methadone; no female
respondent reported ever having used anabolic steroids.

27

Prisoner Reentry Experiences of Adult Females

Exhibit 5. Lifetime substance use, by group

90%
89%
90%

Marijuana
75%
75%
74%

Cocaine
44%
40%

Hallucinogens

48%
36%
34%
39%
35%
31%
41%

Tranquilizers
Amphetamines*

35%
33%
37%

Pain relievers
25%
23%
29%
23%
20%
27%

Sedatives
Stimulants

All
Non-SVORI
SVORI

22%
18%

Heroin*

27%
18%
16%
22%

Inhalants
10%
10%
9%

Methadone
Anabolic steroids

0%
0%
0%

0%

20%

40%

60%

80%

100%

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

In addition, several gender differences are of interest. For six
substances, including cocaine, amphetamines, sedatives,
tranquilizers, stimulants, and pain relievers, the prevalence of
lifetime use was significantly higher among women than among
men. The only substance for which the lifetime prevalence was
higher for men was marijuana. Also of interest is that, among
users, women tended to initiate use at an older age than men.
For six substances, including marijuana, cocaine, tranquilizers,
pain relievers, methadone, and inhalants, women reported a
significantly older age at first use than men.
Respondents were also asked about substance use during the
30 days before their incarceration. About two thirds of the
women (66% of SVORI and 70% of non-SVORI; difference not
statistically significant) reported use of any illicit drug during
this time period. Exhibit 6 shows use of various substances

28

Pre-release Experiences

Exhibit 6. Use of specific substances during the 30 days preceding incarceration, by group

43%
45%
41%

Marijuana

43%
Cocaine*

48%
37%
18%
17%
18%

Amphetamines

All

17%
16%
17%

Pain relievers

Non-SVORI
SVORI

17%
18%
15%

Tranquilizers

11%
10%
12%

Sedatives

9%
10%
9%

Stimulants

0%

20%

40%

60%

80%

100%

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

during the 30 days before incarceration. Slightly less than half
of the women reported use of marijuana and cocaine
immediately before incarceration, with significantly higher
cocaine use being reported by the non-SVORI respondents.
Interestingly, drug use before incarceration was higher among
women than among men for several types of drugs.
Specifically, women were more likely than men to report having
used cocaine, amphetamines, sedatives, tranquilizers, and pain
relievers during the 30 days before incarceration. In contrast,
men were significantly more likely than women to report use of
marijuana and hallucinogens during this time period.
A final indicator of substance use was lifetime receipt of
treatment for alcohol or drugs. Respondents were asked if they
had ever received professional treatment for drugs or alcohol
before their incarceration. More than half of female respondents

29

Prisoner Reentry Experiences of Adult Females

(55%) answered affirmatively, with no significant differences
being found between the SVORI and non-SVORI respondents.
Among the women who had ever received substance abuse
treatment, the average number of times they had started a
treatment program was 3.1 times. Women were significantly
more likely to have received substance abuse treatment before
incarceration than men (among whom only 41% indicated
preincarceration treatment) and had, on average, started a
treatment program significantly more times than men (among
whom the average number of times was 2.3).
Physical Health
At the pre-release interview, respondents were asked to report
whether their current physical health was excellent, very good,
good, fair, or poor. Fewer than half the women (43%)
described their health as very good or excellent; 11% rated it
as poor (with no significant differences between SVORI and
non-SVORI respondents). Men had more positive selfassessments of their health than women. Nearly two thirds
(64%) of men reported very good or excellent health; less than
3% indicated poor physical health.
As part of the pre-release interview, respondents were given
the SF-12, a scale that measures physical and mental
functioning. Scores on the SF-12 physical health scale for
women averaged 49.1, which was significantly lower (indicating
worse physical health) than the 53.5 reported for men. Only
37% of the women reported no physical health limitations
(when asked about five dimensions of physical health
functioning including: moderate activities such as moving a
table, climbing several flights of stairs, accomplishing less than
she would have liked because of her physical health, being
limited in kinds of work or activities because of her physical
health, and pain interfering with her normal work), compared
with 57% of men—a statistically significant gender difference.
Among the women, although the SVORI and non-SVORI groups
had similar overall scores on the SF-12 physical health scale, a
significantly higher percentage of SVORI participants (46%)
reported no physical health limitations than comparison group
members (among whom only 31% reported experiencing no
physical health related limitations).

30

Pre-release Experiences

The percentages of women who reported that they currently
had or had ever had specific diagnoses are shown in Exhibits 7
and 8. Among both “lifetime” and “current” health problems,
the most commonly reported were asthma and chronic back
pain. Very few respondents reported having tested positive for
HIV or tuberculosis.
Exhibit 7. Lifetime health problems, by group

28%
Asthma

29%
27%
24%
25%
24%

Chronic back
pain
17%
19%
14%

High blood
pressure

14%
15%
13%

Arthritis

All

12%
13%
12%

Hepatitis B or C

Non-SVORI
SVORI

11%
14%
8%

Heart trouble

10%
11%
9%

Diabetes
3%
3%
4%

Tuberculosis

2%
2%
1%

HIV positive/AIDS

0%

20%

40%

60%

80%

100%

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

No statistically significant differences were found between the
female SVORI and non-SVORI subgroups; however, when the
women’s data were compared with those of the male sample
(data not shown), several gender differences emerged.
Although the most common health problems were similar for
men and women (with the most commonly reported conditions
for men being asthma, high blood pressure, and chronic back
pain), the prevalence of almost all health conditions was higher
for women. Women reported significantly higher lifetime
prevalence rates for asthma, chronic back pain, diabetes, heart

31

Prisoner Reentry Experiences of Adult Females

Exhibit 8. Current health problems, by group

22%
25%
20%
21%
21%
22%

Asthma

Chronic back pain
11%
13%
9%
12%
14%
10%
11%
12%
10%
7%
8%
6%
7%
9%
4%

High blood pressure

Arthritis

Hepatitis B or C

Heart trouble

Diabetes

Tuberculosis

All
Non-SVORI
SVORI

0%
0%
0%
2%
2%
1%

HIV positive/AIDS
0%

20%

40%

60%

80%

100%

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

trouble, arthritis, tuberculosis, and hepatitis B or C. In addition,
for all of these conditions except tuberculosis, women were
significantly more likely than men to report currently having the
condition. Not surprisingly, the average number of physical
health diagnoses (both lifetime and current) was also
significantly higher for women than men: women averaged 1.2
lifetime diagnoses (compared with 0.8 for men) and 0.9 current
diagnoses (compared with 0.4 for men).
Mental Health
Women also perceived their pre-release mental health status to
be low. Asked to rate their current emotional and mental health
as “excellent,” “very good,” “good,” “fair,” or “poor,” less than
one third of the women (31%) rated their heath as excellent or
very good; 9% rated it as poor. SVORI participants and
comparison group members reported similar mental health
ratings. When gender differences were examined, the men
reported significantly better emotional and mental health than

32

Pre-release Experiences

the women. More than half of the men (51%) reported
excellent or very good mental health, while less than 3%
indicated it was poor.
In terms of mental health functioning, the average score among
the women for the SF-12 mental health scale was 44.8
(significantly lower than the 48.7 average among men). Among
the female sample, the SVORI group had significantly better
mental health functioning than the non-SVORI group, according
to the SF-12 mental health scale scores (46.0 for SVORI and
42.2 for non-SVORI; p < 0.01).
On the SA-45 GSI, an index of mental health status with a
range of 45 to 225 (with higher scores indicating greater
psychopathology), women averaged 79.0. Once again, the
women in the SVORI group had better mental health than the
non-SVORI group. The average GSI score was 74.9 for SVORI
and 82.1 for non-SVORI (p < 0.05). Not surprisingly, among
women as a whole, GSI scores were significantly higher than
those of men (who averaged 67.3). On the Positive Symptom
Total (PST) of the SA-45, the same gender differences emerge.
Among women, the average PST score of 17 indicates that, on
average, women experienced 17 of the 45 symptoms included
in the SA-45 during the 7 days before the interview. In
contrast, men experienced an average of 13 symptoms. No
differences between the female SVORI and non-SVORI groups
were evident.
In addition to the GSI and PST, the SA-45 includes subscales
measuring symptoms of specific psychopathologies. Average
scores for the nine subscales are shown in Exhibit 9. Although
the SVORI and non-SVORI women scored similarly for many
subscales, the rates of depression, psychoticism, and
somatization were significantly higher among the non-SVORI
respondents. For all nine subscales, women’s scores were
significantly higher than those reported for men (data not
shown), reflecting the overall pattern of poorer physical and
mental health among women.
The final indicators of mental health status pertain to mental
health treatment (including prescriptions for medications).
Women were asked if they had received treatment for a mental
health problem before incarceration. Overall, half of the women
reported having received mental health treatment, with a
significantly higher percentage of comparison group members

33

Prisoner Reentry Experiences of Adult Females

Exhibit 9. Average
scores on SA-45
subscales, by group

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

SVORI
9.28
9.39
7.04
8.46
9.28
9.11
7.28
7.15
7.91

Non-SVORI
9.78
10.91
7.62
9.30
9.62
9.95
8.05
7.97
8.92

All
9.57
10.26
7.37
8.94
9.48
9.59
7.72
7.62
8.48

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

(55%) than SVORI participants (44%) having received
treatment. Twice as many women reported prior mental health
treatment as men, among whom only 25% indicated that they
had received treatment for a mental health problem before
incarceration. Specific mental health conditions for which
women reported having ever gone to providers were examined,
and the conditions that were most commonly treated were
depression (33% of the women who had ever received
treatment for a mental health or substance abuse problem),
bipolar disorder (26% of women who had received treatment),
and anxiety (14% of the women who had received treatment).
Although women were more likely to have received such
treatment than men, no differences in prior treatment for these
conditions were evident between the SVORI and non-SVORI
female respondents. During their current period of
incarceration, 33% of women (twice the percentage of men)
reported having been prescribed medication for emotional
problems; 55% (more than twice the percentage of men) felt
that they needed treatment for mental health problems, with
no differences being evident between the SVORI and nonSVORI groups.
Employment History and Financial Support
Employment History
Almost all women (95%) reported having had a job at some
point in their lifetimes, but only about half (53%) were working
during the 6 months before incarceration (Exhibit 10). Prior
employment rates were similar for the SVORI and non-SVORI
respondents, but for both indicators women were significantly
less likely to have worked than men.

34

Pre-release Experiences

Exhibit 10. Employment before incarceration, by group
100%

96%
93%

95%
SVORI
Non-SVORI
All

80%

60%
54%
51%

53%

40%

20%

0%
Ever held a job

Held job 6 months before prison

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

As shown in Exhibit 11, among respondents who worked during
the 6 months before incarceration, three quarters (75%) of the
women described their most recent job as a permanent job and
one for which they received formal pay. Interestingly, women in
the SVORI group appeared to have higher-quality jobs, because
they were more likely to report that their jobs were permanent
and provided formal pay. The great majority of working women
in both groups reported that they worked for more than 20
hours per week at their most recent job during the 6 months
before incarceration. The average number of hours worked
during this time period was 39.5, with an average hourly salary
of $10.15. No significant differences between the SVORI and
non-SVORI groups were evident in hours worked or salary, nor
were there any significant differences between working women
and working men.

35

Prisoner Reentry Experiences of Adult Females

Exhibit 11. Characteristics of respondents’ jobs before incarceration, by group
100%

SVORI
Non-SVORI

92%

91%

91%

91%

All

82%
78%

80%
74%
69%

68%

Held a permanent job*

Worked for formal pay*

60%

40%

20%

0%
Worked more than 20 hrs/week

Note: Among respondents who worked during the 6 months before incarceration.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

Although the majority of working women described their most
recent job as a permanent one, one third (34%) of the
respondents reported having had more than one job during the
6 months before incarceration. Furthermore, 45% reported that
they had 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, nearly half (48%) of the respondents reported jobs
lasting more than 2 years. There were no significant differences
between the SVORI and non-SVORI women in the number of
jobs worked in the 6 months before incarceration, the length of
most recent job, or the longest job ever worked.
The jobs that the women held before incarceration typically
were in the service industry and were similar for SVORI and
non-SVORI respondents. Nearly half of the respondents who
had been employed during the 6 months before incarceration
(49%) reported that their jobs had been in the service industry,

36

Pre-release Experiences

which includes jobs as cooks, waiters, janitors, cashiers, and
dishwashers. Many respondents also reported having worked in
sales, holding positions such as sales representative, realtor,
sales clerk, and telemarketer (12%), or as laborers (12%),
holding positions such as construction worker, day laborer,
landscaper, and roofer. Very few respondents reported having
held professional or technical occupations (4%) or jobs as
managers or administrators (3%). Differences between men
and women in occupation type were found. Significantly more
men reported working as skilled craftspersons or operators,
whereas more women reported sales and service jobs.
Financial Support
All respondents were asked how they had supported
themselves during the 6 months before incarceration, whether
by legal employment or illegal activity and including financial
support from family, financial support from friends, and support
from government programs. Slightly less than half of the
women reported that they supported themselves with a job
(49%) and with support from their family (49%). Nearly as
many (45%) reported that they supported themselves with
illegal activities. More than a quarter reported having received
support from a government program (30%) and from friends
(27%). The only difference between the women enrolled in
SVORI and the comparison group was that comparison group
was significantly more likely to have received support from
their friends than the SVORI participants (32%, compared with
21%).
A deeper understanding of the role of formal employment in
incarcerated women’s overall sources of financial support can
be achieved by examining the breakdown of sources of income
shown in Exhibit 12. In this exhibit, information on sources of
financial support during the 6 months before incarceration is
shown separately for women who had a job during this time
period and for women who did not. Interestingly, it appears
that the main way that employment status affected sources of
financial support (other than legal income) was with respect to
illegal activities. Women who reported having held a job during
the 6 months before incarceration were much less likely to
support themselves with illegal activities than women who did
not hold a job during this time period. Employment status did
not, however, substantially affect the women’s likelihood of

37

Prisoner Reentry Experiences of Adult Females

Exhibit 12. Sources of income during the 6 months before incarceration, by employment
status and group
4%
5%
2%
5%
6%
4%

Other

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

59%
33%

Held job (Non-SVORI)

53%

Illegal

67%
32%
35%

Held job (SVORI)

30%
30%
35%

Government

24%
33%
26%
27%
28%
30%

Friends*

23%
34%
19%
50%
47%
52%

Family

45%

0%

20%

40%

48%
49%

60%

80%

100%

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

receiving support from family, friends, government programs,
or other sources.
Important gender differences emerged with respect to sources
of income. When the various sources of financial support during
the 6 months before incarceration were examined, clear
differences between women and men were evident. Women
were significantly more likely than men to receive financial
support from family (49%, compared with 32%), friends (27%,
compared with 15%), and government programs (30%,
compared with 11%).
In contrast, women were significantly less likely than men to
receive income from legal employment (49%, compared with
62%), reflecting the previously discussed finding of women’s
being less likely than men to report having worked in the 6
months before incarceration. Interestingly, equal proportions of
women (45%) and men (44%) reported having supported
themselves with illegal activities. In addition, the same

38

Pre-release Experiences

“protective” effect of having a job from supporting oneself with
illegal activities (i.e., those who had a job were much less likely
to report illegal activities as a source of financial support) was
evident among both men and women.
Criminal History, Violence, Victimization, and Gang
Involvement
This subsection describes respondents’ involvement with the
adult and juvenile justice systems before incarceration and
outlines women’s experiences with both the perpetration and
victimization of violence before incarceration. The role of gang
membership is also briefly described.
Criminal History
The women reported considerable involvement with the
criminal justice system (Exhibit 13). On average, the women
were 19 years old at the time of their first arrest and had been
arrested slightly more than 10 times. Most respondents had
served time in prison or jail previously, with the women
reporting an average of 1.1 previous incarcerations in prison.
As shown in the exhibit, the SVORI and non-SVORI respondents
were similar on these dimensions of criminal history. Not
surprisingly, however, the women differed substantially from
the adult male sample. The men were, on average, 3 years
younger (16 years old) when they experienced their first arrest
and reported an average of 12.8 (as opposed to 10.6) arrests.
Interestingly, however, the incarceration experiences of men
were similar to those of women (the same proportion had
served time in prison or jail previously and the average number
of previous incarcerations in prison did not differ significantly
from that of women).
One third of the women (30% of SVORI and 36% of nonSVORI; difference not statistically significant) reported that
they had spent time in a juvenile correctional facility for
committing a crime. These respondents had been detained an
average of 4.1 times (3.7 for SVORI and 4.3 for non-SVORI:
difference not statistically significant). Although significantly
fewer female respondents reported spending time in a juvenile
facility than male respondents (33%, as opposed to 50%; p <
0.05), among those who had been detained, the average
number of times was similar for men and women.

39

Prisoner Reentry Experiences of Adult Females

Exhibit 13. Criminal
history of respondents,
by group

Variable
Age at first arrest (mean)
Times arrested (mean)
Times convicted (mean)
Ever previously incarcerated (%)
Times previously incarcerated (mean)

SVORI
19.0
9.6
5.0
0.8
1.1

NonSVORI
19.2
11.4
5.7
0.9
1.3

All
19.1
10.6
5.4
0.8
1.2

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

Exhibit 14 shows the conviction offense(s) for the current term
of incarceration that were reported by the female respondents.
The most frequently reported conviction offense was for a
property crime (43% of SVORI, 39% of non-SVORI; difference
not statistically significant). Almost a third reported a drug
(29%) or person/violent (29%) crime. Public order offenses,
which include probation and parole violations, were reported by
21% of the women. No differences in conviction offense were
evident between the SVORI and non-SVORI groups. Not
surprisingly, however, the women differed from the men in
several ways. The female sample was significantly less likely
than the male sample to be incarcerated for a person/violent
crime (29% of women, compared with 41% of men; for gender
difference, p < 0.05) and more likely to be incarcerated for a
property offense (41% of women, compared with 25% of men;
for gender difference, p <0.05). The likelihood of serving time
for a drug or public order crime did not differ significantly
between men and women.
Perpetration of Violence
Respondents were asked about their experiences with
perpetration of several types of violence during the 6 months
before incarceration, including threats of violence and using (or
threatening to use) a weapon on someone, as well as physically
harming someone by throwing something, pushing/grabbing/
shoving, or slapping/kicking/biting/hitting. Two-thirds of the
women (65% of SVORI and 67% of non-SVORI; difference not
statistically significant) indicated that they had perpetrated at
least one type of violent behavior. This percentage was similar
to that found among the male sample (68%).

40

Pre-release Experiences

Exhibit 14. Conviction offenses for current incarceration, by group

21%
Public order

24%
18%

29%
Drug

26%
32%
All
Non-SVORI
SVORI

41%
Property

39%
43%

29%
Person/Violent

31%
27%

0%

20%

40%

60%

80%

100%

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

Victimization
In addition to being asked about the perpetration of violence,
respondents were also asked whether they had been the victim
of the same acts of violence just described. More than half of
the women (62%) reported having been victimized either
through threats or use of violence during the 6 months before
incarceration, with victims experiencing an average of two
types of victimization. Women were not more likely than men
to report having been victimized before incarceration, but,
among victims, women experienced significantly more types of
victimization than men. No differences between the female
SVORI and non-SVORI groups were evident.
Gang Membership
Only six women (0.02% of the female sample) reported being a
member of a gang at the time of the pre-release interview—a
significantly lower prevalence than that observed among the
male sample (0.05%). Among the very small number of women
who reported gang membership, most (83%) considered their
gang to be their family.

41

Prisoner Reentry Experiences of Adult Females

IN-PRISON EXPERIENCES
This section describes the in-prison experiences of the female
respondents, examining several dimensions. Characteristics of
their current sentences, including sentence length, disciplinary
infractions, victimization experienced during the term of
incarceration, and contact with family during incarceration, are
explored.
Sentence Length
At the time of the pre-release interview (which was conducted,
on average, approximately 30 days before release), the women
had been incarcerated for an average of 1.7 years. The women
who were enrolled in the SVORI program had been incarcerated
for a significantly longer period of time than the non-SVORI
comparison group at the time of the pre-release interview (2.2
years for SVORI and 1.3 years for the non-SVORI group; for
difference, p < 0.05). Because of the gender differences in
offense type, it is not surprising that the women’s sentence
lengths were significantly shorter than those of the male
sample (who had reported an average sentence length of 2.5
years at the time of the pre-release interview).
Disciplinary Infractions and Administrative Segregations
The respondents were asked about disciplinary infractions they
had received and any times that they had been put in
administrative segregation during their current term of
incarceration (Exhibit 15). As shown in the exhibit, about half of
the women reported having received at least one disciplinary
infraction, and about a third had received more than one. A
third also reported being placed in administrative segregation
at least once, with very few (17%) having received
administrative segregation more than once. Consistent with
their longer term of incarceration, the female SVORI
participants were significantly more likely to report having
received disciplinary infractions and having been placed in
administrative segregation during their incarceration. Not
surprisingly (because of their longer term of incarceration),
men reported having received significantly more disciplinary
infractions and were significantly more likely to report being
placed in administrative segregation: at least 2 more times
than women.

42

Pre-release Experiences

Exhibit 15. Disciplinary
infractions and
administrative
segregations during
current incarceration, by
group

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

SVORI

Non-SVORI

All

40%
18%
42%

55%
17%
29%

48%
17%
34%

59%
19%
22%

73%
14%
13%

67%
16%
17%

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

In-Prison Victimization
Respondents were asked whether they had experienced
violence or the threat of violence during incarceration, including
someone threatening to hit them with a fist or anything else
that could hit them; someone using (or threatening to use) a
weapon on them; someone throwing anything at them that
could hurt them; someone pushing, grabbing, or shoving;
someone slapping, kicking, biting, or hitting them; or their
requiring medical attention for violent acts directed at them by
others. Forty-one percent of the women (with similar
proportions of SVORI and non-SVORI group members) reported
having experienced at least one type of victimization, with an
average of 1.0 type of victimization experienced. The likelihood
of victimization during incarceration was lower for women than
it was for men (among whom 55% reported having experienced
at least one type of victimization, with an average of 1.5 types
reported). For reported severity of victimization (derived from a
36-point victimization scale, which was based on the type and
frequency of violence experienced, with higher values indicating
greater severity), the women had an average score of 1.8 (with
no differences between the SVORI and non-SVORI groups),
whereas the men had an average of 2.8 (for gender difference,
p < 0.05).
In-Prison Work
More than half of the women (59% of both the SVORI and nonSVORI groups) indicated that they had a job in the institution
where they were incarcerated. On average, respondents with
prison jobs spent about 24 hours per week working.
Very few respondents reported having a work-release job. Only
4% of the women were on work release at the time of the pre-

43

Prisoner Reentry Experiences of Adult Females

release interview. Those with work-release jobs reported more
hours worked than those with institution jobs reported (36.9
hours per week for the work-release jobs). The SVORI and nonSVORI respondents had similar experiences with work-release
positions.
There were no gender differences in the likelihood of having a
prison job, in the likelihood of being on work release, or in the
number of hours worked at either position.
Family Contact
During the pre-release interview, respondents were asked
about the types of contact they had with their family and
friends during their current incarceration. As discussed
previously, the majority of the women had familial ties—nearly
half were either married or in a committed relationship, and
most (84%) had children. In addition, the women reported
receiving high levels of emotional support from their families at
the time of the pre-release interview, with the women and men
reporting similar levels of familial emotional support.
The women were asked about specific ways in which they
maintained contact with family and friends during their
incarceration, including telephone contact, mail contact, and inperson visits (Exhibit 16). As is evident in the exhibit, women
reported receiving more contact from family members than
from friends. The most commonly reported mode of contact
was mail, followed by phone. In-person visits were less
common, with only 57% of the women reporting any in-person
visits from family and less than 40% reporting visits from
family at least monthly). Perhaps because of their longer
sentence lengths, the women in the SVORI group were
significantly more likely to report phone contact and in-person
visits (from both family and friends) than the non-SVORI
women. Overall, in-prison contact with family and friends did
not differ between women and men (data not shown).

SERVICE NEEDS
Meeting the service needs of returning prisoners was an
integral part of SVORI programming. Information collected from
SVORI program directors revealed that, rather than focusing
resources on a particular skill or need (e.g., vocational
training), most SVORI programs attempted to meet all of the

44

Pre-release Experiences

Exhibit 16. 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
NonSVORI
SVORI
All

Contact With Friends
NonSVORI
SVORI
All

11%
17%
17%
40%
15%

25%
17%
17%
35%
7%

19%
17%
17%
37%
10%

42%
10%
12%
27%
9%

55%
18%
10%
14%
3%

50%
15%
11%
20%
6%

5%
17%
22%
45%
11%

8%
19%
17%
46%
11%

7%
18%
19%
45%
11%

19%
15%
15%
41%
11%

24%
19%
14%
31%
11%

22%
17%
14%
35%
11%

36%
19%
19%
25%
1%

49%
18%
17%
15%
0%

43%
19%
18%
19%
1%

62%
17%
8%
13%
0%

72%
17%
6%
5%
0%

68%
17%
6%
9%
0%

*p < 0.05 for test of significant difference between SVORI and non-SVORI for both family and friend contact.

needs of their target population. Consequently, documenting
the pre-release service needs from the perspective of the
returning prisoners was an important goal of the SVORI
evaluation, these data being used to determine whether
participants’ needs were indeed subsequently met. In addition,
a comparison of pre-release service needs between the SVORI
participants and the non-SVORI comparison group members
was important for assessing the comparability of the two
groups so that any observed “program effects” could be
appropriately attributed to the intervention.
Regarding the comparability goal, it is important to note that,
because the service need questions asked respondents about
their “current” needs (i.e., needs at the time of the pre-release
interview, which was conducted approximately 30 days before
release), interpreting pre-release differences between the two
groups was often difficult. Specifically, any lower need among
the SVORI group could be a direct result of some of the prerelease services that this group had already received rather
than true baseline differences between the groups: because the
pre-release interview was conducted after the respondents had

45

Prisoner Reentry Experiences of Adult Females

already enrolled in and received some pre-release services,
these service receipts presumably would have decreased their
need for such services.
In the pre-release interviews, respondents were asked about
the extent to which they needed each of 29 specific services. 7
For ease of presentation and interpretation, the individual
services were grouped into five service categories or “bundles.”
The bundles are
ƒ

services to help with the transition from prison to the
community;

ƒ

health care services (including substance abuse and
mental health);

ƒ

employment, education, and skills services;

ƒ

domestic violence-related services; and

ƒ

child-related services.

Service need bundle scores were developed from the prerelease 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 zero/one 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. 8 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). 9
Service Need Bundle Scores
This subsection reviews the bundle scores for all SVORI and
non-SVORI female respondents.
Exhibit 17 shows the service need bundle scores for all women,
and for the SVORI and non-SVORI groups. As can be seen from
the exhibit, the highest areas of need for the women were

7
8
9

46

Response options were “a lot,” “a little,” or “not at all.” Responses
were subsequently recoded to “some” and “not at all.”
These items are listed by bundle in Appendix Exhibit A-3 and are
presented bundle-by-bundle in the subsections below.
Although not presented in this report, program-level bundle scores of
service delivery were also developed from reports provided by
SVORI program directors.

Pre-release Experiences

Exhibit 17. Service need bundle scores across service bundles, by group

49
Child servicesa

All

47

Non-SVORI
SVORI

51

21
Domestic violence services

22
20

79

Employment/education/life
skills services

79
80

57
Health services

58
54

73
Transitional services

72
74

0

20

40

60

80

100

Note: Differences between SVORI and non-SVORI were not significant at the 0.05 level.
a
Of those with children younger than 18 years of age.

employment, education, and life skills services (which include
services such as more education, job training, a job for when
the respondent is released, money management skills, life
skills), followed by transition services (which include financial
assistance, public health care insurance, transportation for
when the respondent is released, assistance obtaining a driver’s
license, access to clothing and food). As can be seen, average
bundle scores for all service bundles except domestic violence
services were high. Specifically, the bundle score of 79 for the
employment/education/skills services indicates that, on
average, the women reported needing nearly 80% of the
services in this bundle. Similarly, women reported needing
nearly three quarters of the transition services and about half
of the health services and child services. There were no
differences between the SVORI and non-SVORI samples, which

47

Prisoner Reentry Experiences of Adult Females

indicates that, at the time of the pre-release interview, the
women generally had similar needs for various types of
services.
Although the data are not shown, when the service needs of
men and women were compared, very substantial gender
differences were evident. For all service bundles except child
services, women reported significantly higher needs than men.
Interestingly, although men and women both reported the
highest needs in the employment/education/life skills services,
followed by transition services, among men the third highest
need area was child services, whereas among women it was
health services. As will be discussed in more detail, the relative
importance of child services to men is likely due to men’s need
for assistance in modification of child support debt.
In the subsections that follow, each service bundle is discussed
in more detail, with attention being given to differences
between the SVORI and non-SVORI respondents, as well as to
differences between the total female and male samples.
Transition Services
The transition services bundle reflects services that can help an
individual successfully reintegrate into the community upon
release. The individual services composing this bundle, together
with the proportion of women (presented separately for the
SVORI and non-SVORI groups) who reported needing each
service, are presented in Exhibit 18. Before release, nearly all
women (99%) reported needing at least some transition
services to address immediate needs that would be
encountered on release, such as financial, public, or legal
assistance; a place to live; various identification documents;
transportation; health insurance; and access to emergency
resources, such as clothing and food. As already mentioned,
the average bundle score for the transition services was 74 for
the SVORI women and 72 for the non-SVORI women
(difference not statistically significant), indicating a very high
level of need for transition services. Overall, more than half of
the women reported needing each of these transition services,
and, for many services, well over three quarters of the women
reported need.

48

Pre-release Experiences

Exhibit 18. Self-reported need for specific transition services, by group

76%
75%
77%

Access to clothing/food

79%
77%
82%

Driver's license

77%
73%

Transportation*

82%
55%
54%
58%

Place to live

All
Non-SVORI

60%
61%
58%

Documents for
employment

SVORI
83%
81%
86%

Mentor

91%
90%

Public health care
insurance

92%
68%

Public financial assistance*

73%
61%
87%
85%
90%

Financial assistance
54%
54%
54%

Legal assistance
0%

20%

40%

60%

80%

100%

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

The highest areas of need were public health care insurance
and financial assistance. The great majority of women reported
needing assistance in these areas (91% for public health care
insurance and 87% for financial assistance). Needs were also
high for basic services, such as access to clothing and food.
Approximately 30 days before release, 76% of women reported
that they would need access to clothing banks and food
pantries when they were released. A surprisingly high
percentage of women (83%) reported that they needed a
mentor. The need for a driver’s license was also reported by
most women (79%). In addition, more than half of the women
(60%) reported needing other identification documents
necessary for obtaining employment and securing public
benefits, such as a birth certificate, Social Security card, and
photo identification card.

49

Prisoner Reentry Experiences of Adult Females

As shown in the exhibit, significant differences between the
SVORI and non-SVORI respondents were evident for two of the
transition services. Specifically, women enrolled in SVORI were
less likely to report needing public financial assistance yet more
likely to report needing transportation than their non-SVORI
counterparts.
Although the data are not shown, gender differences for specific
transition services follow the overall pattern already discussed,
in which women have higher self-reported service needs than
men. For 7 of the 10 individual transition services, women were
significantly more likely to report needing the service than men.
Gender differences were most notable for public health care
insurance (reported by 91% of women and 74% of men) and
the need for a mentor (reported by 83% of women and 60% of
men). The only services for which gender differences were not
evident were financial assistance, documents for employment,
and driver’s license.
Health Services
Respondents’ self-reported needs for health services are shown
in Exhibit 19. Overall, women ranked health-related services
third out of the six types of services included in the pre-release
interview. The average bundle score of 57 (see Exhibit 17)
indicates that, on average, the women reported needing more
than half of the services in the health bundle. Virtually all of the
women (97%) reported needing at least one health-related
service.
When specific services in the health services bundle were
examined, the most needed service by far was medical
treatment, which was reported as a need by 78% of the
women. More than half of the women also reported needing
substance use treatment (65%), mental health treatment
(56%), and an anger management program (52%). Slightly
less than one third needed to participate in a support group for
victims of sexual or physical abuse.
As shown in the exhibit, the only service need for which
significant differences occurred between the SVORI and nonSVORI respondents was for an anger management program.

50

Pre-release Experiences

Exhibit 19. Self-reported need for specific health services, by group

52%
Anger management
program*

59%
42%

All

32%
Support group for abuse
victims

Non-SVORI

34%

SVORI

29%

65%
Substance use treatment

64%
66%

56%
Mental health treatment

57%
55%

78%
Medical treatment

77%
79%

0%

20%

40%

60%

80%

100%

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

Not surprisingly, for all five health services, self-reported need
was significantly higher for women than for men (data not
shown). The need for support groups for abuse victims was
particularly discrepant among men and women, with only 4% of
men but 32% of women reporting this need. Similarly, twice as
many women reported the need for mental health treatment as
men (56% of women, compared with 25% of men).
Employment/Education/Skills Services
Exhibit 20 shows the respondents’ self-reported needs for
services related to employment, education, and skills. As
discussed, this bundle of services was ranked the highest (in
terms of needs) among the sample, with the average bundle
score of 79 indicating that the respondents, on average,
needed almost 80% of the services in this bundle. Once again,
almost all women (99%) reported needing at least one service

51

Prisoner Reentry Experiences of Adult Females

Exhibit 20. Self-reported need for employment, education, and skills services, by group

75%
Change in criminal attitudes

75%
74%
78%

Work on personal
relationships

79%
78%
74%

Life skills

74%

All

73%

Non-SVORI
SVORI

70%
Money management skills

69%
73%
95%

More education

95%
95%
83%
81%

Job

87%
83%
81%
86%

Job training

0%

20%

40%

60%

80%

100%

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

from the employment/educations/skills service bundle. It is
clear from the data in the exhibit that the most needed
individual service by far is more education, which almost all
women (95%) reported needing.
Among the remaining services, more than three fourths of the
women reported needing job training and a job itself. The
majority of women also recognized that some aspect of their
own behavior needed to change to improve their lives after
release. More than three quarters of the women indicated that
they needed to work on personal relationships and change their
attitudes related to criminal behavior. Almost as many (74%)
indicated that they needed to learn life skills, and 70% of
women reported needing to learn money management skills.
As shown in the exhibit, women in the SVORI and non-SVORI
groups did not differ in their needs for any services related to
employment, education, and skills. In addition, fewer gender

52

Pre-release Experiences

differences were found for this service bundle than for the other
service bundles. Among the seven specific employment/
education/skills services, women had significantly higher needs
for four: a job, job training, the need to work on personal
relationships, and the need to change attitudes related to
criminal behavior.
Domestic Violence Services
In the pre-release interview, women (and men) were asked
about their need for two types of domestic violence services—
batterer intervention programs and domestic violence support
groups. These two services constituted the domestic violence
service bundle. As mentioned previously, needs for services in
this bundle were much lower than those for the other five
service areas, with the women having an average domestic
violence bundle score of 21.
Seventeen percent of women (18% of SVORI and 16% of nonSVORI; difference not statistically significant) reported needing
a batterer intervention group (defined as a special program to
help people who have problems with physically abusing their
partners), and one quarter (22% of SVORI and 27% of nonSVORI, difference not statistically significant) reported needing
a domestic violence support group. Not surprisingly, both of
these proportions are significantly higher than those found
among men (among whom only 8% reported needing a
batterer intervention program and 7% reported needing a
domestic violence support group).
Child Services
Respondents who had minor children (76% of women) were
asked about their need for child-related services, including
instruction in parenting skills, child care (on release), child
support payments, modifications in the child support debt the
respondent owed, and modifications in the custody of the
respondent’s children. The average child services bundle score
among women was 49 (51 for the SVORI group and 47 for the
non-SVORI group; difference not statistically significant),
indicating that, on average, women reported needing about half
of the child-related services. Most (90%) of the women
reported needing at least one service in the bundle.
The percentage of mothers reporting need for specific childrelated services is shown in Exhibit 21. As shown in the exhibit,

53

Prisoner Reentry Experiences of Adult Females

Exhibit 21. Self-reported need for specific child services, by group

All

38%
Child care (when released) a

Non-SVORI

34%

SVORI

44%

70%
a

Parenting skills

72%
69%

40%
a

Modification of custody

39%
41%

86%

Modification in child support
debtb

91%
83%

45%
Child support paymenta

43%
48%

0%

20%

40%

60%

80%

100%

Note: Differences between SVORI and non-SVORI were not significant at the 0.05 level.
a
Of those with children younger than 18 years of age.
b
Of those who owed back child support.

the most commonly reported need in this service bundle was
modification in the child support debt owed by the women.
Importantly, however, this item was asked only of the 10% of
mothers who actually owed back child support. Among all
mothers of minor children, 70% reported needing help
developing parenting skills. Less than half of the mothers
reported needing child support payments (45%), modification
of custody arrangements for their children (40%), and child
care (when released; 38%). As shown in the exhibit, the SVORI
and non-SVORI samples had similar needs for all child-related
services.
Interestingly, the only difference between the male and female
samples in their self-reported needs for child-related services
was parenting skills. Significantly more mothers (70%)

54

Pre-release Experiences

reported needing to learn parenting skills than fathers (62%) at
the pre-release interview. Unlike most of the other types of
services (for which women had higher self-reported need),
women and men reported virtually identical levels of need for
child-related services.
Level of Need Across Services
Overall, the examination of service needs by categories of
services has demonstrated extremely high need among
incarcerated women at a time period approximately 30 days
before release. For all service bundle areas except child-related
services, women had substantially (and statistically
significantly) higher self-reported need than men. When gender
differences in the individual service items were assessed,
women had significantly higher needs than men for 19 of the
29 services measured in the pre-release interview. Very few
differences between the female SVORI and non-SVORI
respondents were evident, indicating that the two groups had
similar service needs at the time of the pre-release interviews.
As an additional examination of self-reported service need
among women, the most commonly reported service needs
were identified. The top 10 are shown in Exhibit 22. Overall
among women, the most commonly reported service need was
for more education, which was reported for 95% of women. The
need for more education was closely followed by the need for
public health care insurance (reported by 91% of women) and
financial assistance (reported by 87% of women). More than
three quarters of the women also reported needing a mentor
(83%), job training (83%), a job (83%), a driver’s license
(79%), medical treatment (78%), and work on personal
relationships (78%). Although there were some minor
differences in the order in which the top 10 needs were ranked
by the SVORI and non-SVORI respondents, the percentages
reporting each service need did not differ significantly by group
(with the exception of transportation).
In addition to the rankings that were created (assessing the
most commonly reported service needs), the respondents were
also asked to report their top two service needs. According to
this measure, the services that were most commonly reported
in the “top two” were a job (reported by 37% of the women)
and a place to live (on release; reported by 19% of the
women). Women were diverse in their identification of the most

55

Prisoner Reentry Experiences of Adult Females

Exhibit 22. Most commonly reported service needs, by group

78%
79%
78%

Work on personal
relationships

95%
95%
95%

More education
83%
81%
87%
83%
81%
86%

Job
Job training

78%
77%
79%
All
76%
Non-SVORI
75%
77%
SVORI
79%
77%
82%
77%
73%
82%

Medical treatment
Clothing/food bank
Driver's license
Transportation*

83%
81%
86%

Mentor

91%
90%
92%

Public health care insurance

87%
85%
90%

Financial assistance
0%

20%

40%

60%

80%

100%

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

important services because—other than a job, a place to live,
more education (listed in the top two by 15% of women), and a
driver’s license (listed by 12% of women)—no single services
were listed in the top two by more than 10% of the sample.
Men identified similar service needs in their rankings, with the
most frequently mentioned need being a job, followed by a
driver’s license and more education.
As a final indicator of service need, an “all services” bundle was
created to capture the level of overall need across all services
(in addition to the service bundles already described). On
average, the respondents reported needing nearly two thirds of
the all the service items (average score of 64 for both groups).
Reflecting the overall pattern of significantly greater pre-release
service need among women, the “all services” bundle score was
significantly higher for women than for men (among whom the
average “all services” bundle score was 54).

56

Pre-release Experiences

SERVICE RECEIPT
As discussed, the SVORI programs were intended to address
the high needs of returning prisoners by increasing access to
services; therefore, obtaining self-reported information on the
services actually received by SVORI participants (in addition to
parallel reports from the program directors on what services
were being delivered) was critical to understanding the manner
in which the programs were implemented. The evaluation also
documented services received by non-SVORI comparison group
members in order to determine which services appeared to be
delivered through “treatment as usual.” This evaluation was
used, in turn, to determine whether SVORI did indeed result in
enhanced service delivery.
This section presents findings on self-reported receipt of
services. In the interview (which was conducted approximately
30 days before release), participants were asked about services
they had received at any point during their incarcerations.
Because many SVORI participants had actually enrolled in the
SVORI programs (and had begun receiving services) well before
the pre-release interview, the information reported here yields
only an initial impression of program implementation. The
comparison of pre-release services received by women
participating in SVORI and their non-SVORI counterparts allows
us to determine whether SVORI programs offered more access
to services among participants than was available through
“treatment as usual” in the prisons.
Service receipt bundle scores were calculated in a manner
parallel to the service need bundle scores reported in the
previous section. Specifically, respondents were asked whether
they had received each of 36 services since they were
incarcerated, and the number of “yes” responses to the items in
a bundle was divided by the number of bundle items and
multiplied by 100. Individual bundle scores were averaged to
yield overall scores. As with the service needs bundle, scores
for the child services receipt bundle were generated only for
those respondents who reported having children younger than
18. In addition to the bundles that were created for service
needs (transition services; health services; employment.
education. and skills services; domestic violence services; and

57

Prisoner Reentry Experiences of Adult Females

child services), a sixth bundle, service coordination items, was
included for service receipt. 10
Service Receipt Bundle Scores
The service receipt bundle scores for the total female sample,
as well as for the SVORI and non-SVORI groups, are shown in
Exhibit 23. As is evident from the exhibit, the SVORI programs
appeared to substantially increase access to services for the
women. For all service areas, women enrolled in SVORI
programs reported significantly higher service receipt during
their period of incarceration than comparable women not
enrolled in SVORI. For several bundles, SVORI participants
received twice the services that the comparison group received.
The most substantial difference was with respect to
coordination services: SVORI participants received more than
70% of the services in this bundle, compared with only 30%
received by non-SVORI comparison group members. SVORI
participants also reported receiving more than half of the
employment/education/life skills services (compared with only
one quarter received by comparison group members) and
nearly half of the health services. For health services,
differences between the SVORI and non-SVORI respondents,
while statistically significant, were not as large. Services less
frequently reported were domestic violence services and child
services.
Although the data are not presented in this report, the
differences between men and women in the level of service
receipt during incarceration were also compared. Reflecting the
higher service need among women as discussed in the previous
section, women reported receiving substantially higher levels of
services than men. For all service bundles except coordination
services, the bundle scores were significantly higher for women
than for men.
The subsections that follow provide additional detail on the
items within individual service receipt bundles. Gender
differences are examined further, in addition to differences
between the female SVORI and non-SVORI groups.

10

58

These items are listed by bundle in Appendix Exhibit A-4 and are
presented bundle-by-bundle in the subsections that follow.

Pre-release Experiences

Exhibit 23. Service receipt bundle scores across service bundles, by group

14

a

Child services *

All

9

Non-SVORI

19

SVORI

12
Domestic violence services*

9
16

38

Employment/education/life
skills services*

25
55

41
Health services*

36
48

31
Transitional services*

24
41

49
Coordination services*

30
74
0

20

40

60

80

100

a

Of those with children younger than 18 years of age.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

Coordination Services
Because the SVORI programs attempted to provide needed
services to program participants, accurate assessment of
participants’ needs is essential to designing a treatment
program that will successfully meet individuals’ needs.
Assessment of needs specific to reentry is also important, in
addition to release planning and the development of a specific
reentry plan. Finally, the provision of case management is
necessary to effectively coordinate the various stages of
assessment, planning, and service delivery. The coordination
services bundle measures the receipt of these types of
“services” during incarceration.
Exhibit 24 shows the proportion of all women, as well as the
SVORI and non-SVORI groups, who reported receiving each of
the five coordination services during their terms of

59

Prisoner Reentry Experiences of Adult Females

Exhibit 24. Self-reported receipt of specific coordination services, by group

51%
Release planning*

27%
82%

42%
Reentry plan*

19%
73%
All

44%

Release needs
assessment*

Non-SVORI

26%

SVORI

69%

49%
Case manager*

33%
71%

61%
Needs assessment*

47%
81%

0%

20%

40%

60%

80%

100%

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

incarceration. Interview respondents were provided with
detailed descriptions of needs assessment and case
management; for each description, they were asked whether
they had received the service since incarceration. They were
also asked whether they had received a needs assessment that
was specifically designed to help prepare them for their release
(release needs assessment). “Release planning” indicates the
proportion of respondents who reported that they had worked
with anyone to help plan for their release, and “reentry plan”
indicates the proportion of respondents who reported that a
reentry plan had been developed for them.
As is evident from the chart, levels of receipt were high among
the SVORI participants. For each of the five services, a
significantly higher proportion of SVORI participants reported
receiving the service. This pattern is to be expected because of
the integral role of service coordination in SVORI programming
and the fact that close coordination of services is a departure
from “treatment as usual” in correctional settings. Notably,
despite the importance of these services, substantially less than
100% of respondents received them. Even among SVORI
participants, almost 20% did not receive a needs assessment,
and almost 30% had not met with a case manager during the
incarceration.

60

Pre-release Experiences

Although the overall service coordination receipt scores were
significantly higher for women than for men, when gender
differences in specific services were examined, only two
individual services differed significantly among men and
women. Women were significantly more likely to receive needs
assessments (assessing general needs) and needs assessments
specific to release.
Transition Services
The transition services help individuals successfully return to
the community. Exhibit 25 shows the proportion of all women,
as well as the SVORI and non-SVORI groups, who reported
receiving each of the 12 transition services during the current
period of incarceration. As shown in the exhibit, levels of
service receipt were typically higher among the SVORI
participants. With the exception of help finding transportation
and legal assistance, a significantly higher proportion of SVORI
participants reported receiving each service.
Exhibit 25. Self-reported receipt of specific transition services, by group
36%
31%

Help accessing
clothing/food banks*

All

43%
20%
15%

Help getting a driver's
license*

21%
18%
25%
22%
10%

Help finding transportation
Mentoring*

SVORI

27%
27%
19%

Help finding a place to live*

Non-SVORI

38%

38%
51%
44%

Help with documents*

59%
18%
15%
22%
16%
8%
26%
17%
14%
22%

Help obtaining legal
assistance
Help accessing public health
care*
Help accessing public
financial assistance*
10%
2%

Help accessing financial
assistance*

20%
63%
49%

Release preparation
classes*

83%

72%
59%

Release preparation
programs*

90%

0%

20%

40%

60%

80%

100%

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

61

Prisoner Reentry Experiences of Adult Females

The most commonly reported items were participation in
classes (83% of SVORI participants, compared with 49% of
comparison group members) and programs designed
specifically to prepare individuals for release (90% and 59% of
SVORI participants and comparison group members,
respectively). More than half of the women also reported
receiving assistance obtaining documents necessary for
employment, such as birth certificate or Social Security card
(59% and 44% of SVORI participants and comparison group
members, respectively).
Despite high levels of reported need, less than half of the
women received the remaining nine transition services. A
higher proportion of SVORI participants received seven of these
less frequently received services than comparison group
members.
A significantly higher proportion of women than men received
11 of the transition services. No difference was found in receipt
for help accessing financial assistance, which was one of the
least common services reported by both men (9%) and women
(10%).
Health Services
Exhibit 26 shows the proportion of all women, as well as the
SVORI and non-SVORI groups, who reported receiving a variety
of health services. Not surprisingly, any medical treatment
(71%) was reported more frequently than either any substance
use treatment (43%) or any mental health treatment (41%)
among the full sample of women. Less than 30% of women
reported participation in either an anger management program
or a group for victims of abuse.
Although the difference in the receipt of any medical treatment
between SVORI (69%) and non-SVORI (72%) women was not
significant, SVORI participants were more likely to receive
dental services, any mental health treatment, and any
treatment for substance use (54% of SVORI participants
compared with 35% of comparison group members).
Furthermore, a higher proportion of SVORI participants
reported participation in anger management programs and in
groups for victims of abuse.

62

Pre-release Experiences

Exhibit 26. Self-reported receipt of specific health services, by group

13%
Victims' group for abuse*

All

8%

Non-SVORI

18%

SVORI
27%

Anger management
program*

19%
37%
43%

Any substance use
treatment*

35%

Any mental health
treatment*

35%

54%
41%
48%
53%

Dental services*

45%
63%
71%
72%
69%

Any medical treatment

0%

20%

40%

60%

80%

100%

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

Corresponding to their higher levels of need, a higher
proportion of women than men reported receiving three of the
health services. Women reported higher levels of receiving any
medical treatment (71% of women, compared with 56% of
men). Furthermore, twice as many women reported receiving
any mental health treatment (41% of women, compared with
18% of men) and participating in groups for victims of abuse
(13% of women and 5% of men). While women reported
significantly greater need for anger management and substance
use treatment than men, they were not more likely to receive
such services.
Employment/Education/Skills Services
Respondents’ self-reported receipt of employment, education
and skills services is reported in Exhibit 27 for the full sample,
as well as the SVORI and non-SVORI groups. Although the
most commonly received service among the full sample of

63

Prisoner Reentry Experiences of Adult Females

Exhibit 27. Self-reported receipt of specific employment, education, and skills services, by
group

45%

Received training to change
criminal behavior attitudes*

29%
46%
30%

Received assistance with
personal relationships*

19%
39%
42%

Received other life skills
training*

25%
43%

All
Non-SVORI

23%

Received money management
services*

SVORI

10%
30%
49%

Received any educational
services*

40%
49%
38%

Received any employment
services*

26%
44%
0%

20%

40%

60%

80%

100%

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

women was education, there was still a large unmet need:
nearly all of the women reported needing education (95%), but
only half (49%) received it.
Other frequently reported services included training to change
criminal behavior attitudes (45%) and other life skills training
(42%). Notably, a significantly higher proportion of SVORI
participants than comparison group members received each of
the services in this bundle.
Differences in service receipt also emerged by gender. Women
were significantly more likely to receive money management,
other life skills, assistance with personal relationships, and any
employment services than the men.
Domestic Violence Services
Corresponding to the relatively low self-reported need, few
women received domestic violence services. Seventeen percent

64

Pre-release Experiences

of women participated in domestic violence support groups,
including twice as many SVORI (24%) as non-SVORI (12%)
women. Not surprisingly, a significantly lower proportion of
men (8%) reported such participation.
Participation in batterer intervention programs was low for all
populations, and differences in service receipt were not
significant. Six percent of women (7% of the SVORI group and
5% of non-SVORI group) and 4% of men reported involvement
in these programs.
Child Services
Although parenting classes were the most frequently reported
child-related service overall (37%), almost twice as many
SVORI mothers (50%) as non-SVORI mothers (26%)
participated in these classes. The other child services were
received by less than 15% of the mothers. SVORI mothers
were significantly more likely than non-SVORI mothers to
report receiving assistance finding child care on release from
prison (8% and 2%, respectively).
A couple of differences were found in service receipt by gender,
with significantly higher proportions of women reporting
receiving assistance modifying custody (10% of women and 3%
of men) and parenting classes (37% of women and 20% of
men).
Levels of Receipt Across Services
Overall, this examination of service receipt has shown that
SVORI participants received higher levels of services than
comparison group members. In contrast to the similarities
reported in need, a significantly higher proportion of SVORI
respondents than non-SVORI respondents reported receiving
each of the service areas. Examining the individual services
that compose these bundles has shown that a higher proportion
of SVORI than non-SVORI respondents reported receiving 97%
of the services; the difference was statistically significant for
90% of the services.
The services the women most frequently reported receiving
were participating in programs to prepare for release (72%),
receiving medical treatment (71%), taking a class specifically
for release (63%), and receiving a needs assessment (61%).
With the exception of medical treatment, SVORI participants

65

Prisoner Reentry Experiences of Adult Females

were significantly more likely to report receiving each of these
services than non-SVORI respondents (Exhibit 28).
Exhibit 28. Most commonly reported services received, by group

72%
Participated in programs
to prepare for release*

59%
90%

71%
Received any medical
treatment

72%
69%
All
Non-SVORI

63%
Took class specifically for
release*

SVORI

49%
83%

61%
Received needs
assessment*

47%
81%

0%

20%

40%

60%

80%

100%

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

Similar to the “all services” need bundle, the “all services”
receipt bundle captures the level of overall receipt across all
available services. SVORI participants reported receiving nearly
half of the services, whereas the comparison group received
only one quarter of the services (average bundle scores of 46
and 25, respectively; p < 0.01).
In addition to reporting service receipt, respondents were also
asked to identify which two services they thought were most
helpful. Overall, the most frequently reported “top two”
services were education (22%), spiritual or religious assistance
(21%), and alcohol or drug treatment (20%). Except for
parenting classes (10%), no other service was ranked in the
top two by more than 10% of the sample. SVORI and non-

66

Pre-release Experiences

SVORI respondents did not differ significantly in these
selections. Although these three services were also the most
commonly identified by men as being among the two most
helpful, a higher proportion of women (21%) than men (16%)
selected spiritual or religious assistance.
With the exception of coordination services, a significantly
higher proportion of women reported receiving each of the
service bundles than men. When the individual items in each
bundle where examined, women were found to have reported
significantly higher service receipt than men for 23 of the 36
services (64%). Reflecting these higher levels of receipt across
most areas, the “all services” receipt bundle was significantly
higher for women than for men (average bundle scores of 34
and 28, respectively).

67

Post-release
Experiences of
Returning Female
Prisoners
Findings from the three post-release interviews, which were
administered 3, 9, and 15 months after release, are presented
here. Unlike the pre-release section, which was intended to be
descriptive, this section examines reentry outcomes for women
who participated in SVORI programming as compared with
outcomes for women who received “treatment as usual.”
Consequently, potential bias associated with treatment group
membership must be adjusted for, because in most sites
women were not randomly assigned to SVORI treatment.
As detailed in Lattimore and Steffey (2009), propensity
modeling was employed to model the likelihood of SVORI
status. The propensity score weights developed from these
models were applied to the raw data; therefore, all of the data
presented in this section are weighted to adjust for selection
bias (unlike the “Pre-release Experiences of Returning Female
Prisoners” section, which presents unweighted data). Because it
is of interest to examine patterns across waves (beginning with
the pre-release interviews) in a comparable manner, weighted
Wave 1 (pre-release) data are also presented in several places
in this section. To assess whether SVORI participation had a
significant impact on a range of outcomes, the authors ran a
series of weighted bivariate regression (for continuous
outcomes) and logistic regression (for dichotomous outcomes)
models (with treatment status as the independent variable and
the outcome of interest as the dependent variable). Due to the

69

Prisoner Reentry Experiences of Adult Females

relatively small sample sizes, regression models were not
always appropriate and were occasionally excluded from the
exhibits; weighted means are presented for all outcomes. The
exhibits present regression results only when there were at
least 20 respondents (with a minimum of 10 SVORI and 10
non-SVORI respondents) in each cell. For example, at the 15month post-release interview only one SVORI and 7 non-SVORI
respondents reported needing assistance modifying child
support debt; because these cells are too small to render
logistic regression results meaningful, only weighted means are
presented.
It is also important to note that, while descriptive comparisons
of trends across time are discussed, the number of respondents
varied at each wave, and significance tests of differences across
time were not conducted. The outcome analyses were not
limited to individuals who responded to all interviews, so the
possibility cannot be ruled out that some of the differences
across time are a result of the differences in respondents across
time.

SERVICE NEEDS
This subsection focuses on self-reported service needs at each
interview wave, with service receipt being addressed in the
next subsection. The same “bundle scoring” procedures
described in the “Pre-release Experiences of Returning Female
Prisoners” section, in which 29 services were grouped into five
service bundles, were used. Findings on self-reported needs for
services in each of the five bundles over the four interview
waves are discussed here, after a brief discussion of overall
patterns across bundles and waves.
Although the data are not shown, gender differences in selfreported service need among men and women over time are
discussed.
Service Need Bundle Scores
Exhibit 29 presents the results of the weighted outcome
analyses and the weighted means for each service bundle (and
individual item) across waves. Several interesting patterns are
evident in the data.

70

—
26%
28%
27%

91%
—
41% −0.04
71% 0.09
35% −0.41

—
0.96
1.09
0.66

3.49
26% 0.82

0.79
1.15

1.04
0.83

47.88 4.02
44% −0.20

27%
26%
3.92
30%
26%

96%
34%
50%
28%

40.80
46%

59%
46%
8.04
7%
9%

−0.11
−0.02
−0.09
—
−0.32

100%
—
46% −0.51
68% −0.78
24% 0.18

46.06 −5.25
42% 0.16

62%
47%
8.12
4%
12%

64.01 −0.37
67% 0.25
65% 0.22
92%
—
55% −0.06
64% −0.17

63.64
72%
70%
93%
54%
60%

2.65
31%
31%
—
25%
25%
0.73
0.59
—
0.86
1.10

Wave 2
Non-S
Mean Est.
59.63 0.29
46% 0.14
71% 0.21
61% −0.42
79% 0.06
64% −0.31
32% −0.14
52% 0.20
66% 0.42
67% 0.22
58% −0.08
40.72 −2.19
70% 0.02
38% 0.22
38% −0.12
20% −0.09
38% −0.59

SVORI
SE
OR
Mean
2.47
59.92
23% 1.02
49%
34% 0.58
75%
24% 1.70 * 51%
40% 0.82
80%
30% 0.78
56%
23% 1.11
29%
23% 0.79
57%
27% 0.51 * 75%
28% 0.74
71%
26% 0.88
56%
2.94
38.52
27% 0.92
71%
23% 0.96
43%
24% 0.82
35%
24% 1.11
19%
23% 1.96 * 25%

—
0.33
0.33
0.37

4.46
0.32

0.28
0.27
3.25
—
0.44

3.83
0.29
0.29
—
0.27
0.28

SE
3.43
0.27
0.30
0.27
0.34
0.27
0.29
0.27
0.30
0.29
0.27
4.05
0.30
0.28
0.28
0.34
0.30

—
0.60
0.46
1.20

NA
1.17

0.90
0.98
NA
—
0.73

NA
1.28
1.24
—
0.94
0.84

1.15
1.23
0.66
1.07
0.73
0.87
1.22
1.52
1.24
0.92
NA
1.03
1.25
0.89
0.91
0.55

OR

*

96%
27%
57%
23%

38.45
41%

67%
45%
11.50
8%
15%

60.24
55%
57%
87%
54%
51%

SVORI
Mean
51.11
38%
71%
44%
67%
57%
28%
43%
64%
56%
43%
37.37
61%
40%
36%
18%
*
32%
−0.12
−0.09
0.01
−0.01
−0.27
−0.26

100%
—
32% −0.20
57% 0.00
27% −0.21

39.45 −1.01
41% 0.02

59% 0.35
40% 0.20
11.85 −0.35
8% 0.04
16% −0.07

60.36
58%
57%
87%
61%
58%

Wave 3
Non-S
Mean Est.
52.04 −0.93
37% 0.06
66% 0.22
58% −0.57
69% −0.08
58% −0.01
30% −0.11
40% 0.09
57% 0.30
60% −0.15
46% −0.11
38.27 −0.90
56% 0.22
36% 0.18
41% −0.20
18% −0.02
41% −0.38

—
0.35
0.31
0.37

5.11
0.32

0.27
0.27
3.74
0.48
0.38

4.02
0.27
0.27
0.41
0.27
0.27

SE
3.67
0.28
0.28
0.27
0.29
0.27
0.30
0.27
0.27
0.27
0.27
3.85
0.27
0.28
0.28
0.35
0.28

—
0.82
1.00
0.81

NA
1.02

1.41
1.22
NA
1.04
0.93

NA
0.91
1.01
0.99
0.76
0.77

1.06
1.25
0.56
0.92
0.99
0.90
1.09
1.35
0.86
0.89
NA
1.25
1.20
0.82
0.98
0.68

OR

70%
25%
52%
20%

34.54
37%

64%
43%
9.57
6%
14%

57.49
60%
56%
87%
46%
50%

−6.20
0.08
−0.28
−0.50
−0.50
−0.53

3.95
0.26
0.26
0.41
0.25
0.26

90%
35%
63%
31%

—
—
−0.51 0.32
−0.44 0.29
−0.57 0.34

42.90 −8.36 4.38
40% −0.09 0.30

61%
0.11 0.26
45% −0.10 0.26
10.97 −1.40 3.17
8%
—
—
14%
0.00 0.37

63.68
58%
63%
92%
59%
63%

Wave 4
SVORI Non-S
Mean Mean Est.
SE
47.67 56.33 −8.66 3.88
36%
43% −0.30 0.26
63%
65% −0.08 0.26
* 43%
64% −0.84 0.26
63%
70% −0.31 0.27
50%
60% −0.42 0.26
26%
37% −0.52 0.28
51%
48%
0.11 0.25
54%
64% −0.42 0.26
52%
61% −0.36 0.26
40%
52% −0.48 0.26
35.25 39.35 −4.10 3.90
58%
57%
0.02 0.26
38%
41% −0.11 0.26
33%
36% −0.13 0.26
20%
18%
0.12 0.32
27%
44% −0.74 0.27

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

Of those who owed back child support.

b

Of those with children younger than 18 years of age.

a

Note: Regression results not shown when cell sizes <10. Non-S = Non-SVORI; NA = not applicable. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months
post-release; Wave 4 = 15 months post-release.

Child
51.91
a
Child support payments
49%
Modification in child support
b
debt
85%
a
Modification in custody
42%
a
Parenting skills
70%
a
Child care
45%

Wave 1
SVORI Non-S
Mean Mean Est.
Transition
73.97 71.82 2.15
Legal
54%
55% 0.02
Financial
91%
85% −0.55
Public financial
60%
72% 0.53
Public health care
91%
90% −0.20
Mentor
85%
82% −0.24
Documents for employment
58%
60% 0.10
Place to live
58%
52% −0.24
Transportation
83%
72% −0.67
Driver’s license
82%
77% −0.30
Clothing/food banks
78%
76% −0.12
Health
54.94 57.42 −2.48
Medical treatment
80%
78% −0.09
Mental health treatment
56%
55% −0.04
Substance use treatment
68%
64% −0.20
Victim support group
30%
32% 0.11
Anger management
42%
59% 0.67
Employment/Education/Life
Skills
80.48 78.76 1.73
Job training
86%
82% −0.31
Job
88%
81% −0.52
Education
95%
95%
—
Money management skills
73%
70% −0.15
Life skills
73%
75% 0.10
Work on personal
relationships
78%
79% 0.04
Change in criminal attitudes 78%
74% −0.19
Domestic Violence
21.41 20.97 0.44
Batterer intervention
19%
16% −0.24
Support group
24%
26% 0.14

Exhibit 29. Weighted means and parameter estimates of the effect of SVORI on service need

—
0.60
0.65
0.57

NA
0.91

1.11
0.90
NA
—
1.00

NA
1.08
0.75
0.61
0.61
0.59 *

*
0.74
0.92
0.43 *
0.74
0.66
0.60
1.12
0.66
0.70
0.62
NA
1.02
0.90
0.88
1.13
0.48 *

OR

Post-release Experiences

71

Prisoner Reentry Experiences of Adult Females

The first pattern pertains to declining needs over time. Selfreported needs were extremely high at the pre-release
interview (conducted approximately one month before release),
declined substantially at the 3-month post-release time period,
and then continued declining gradually at the 9- and 15-month
interviews. For example, the mean bundle scores for
employment/education/life skills service needs among the
SVORI participants, which is the highest reported area of need
among the female sample, declined from 80 (at the pre-release
interview) to 57 at the 15-month post-release interview.
Substantively, the scores are interpreted to mean that, at the
time of the pre-release interview, the women reported needing
80% of the services in this bundle, whereas at the 15-month
post-release interview they needed just slightly more than half
of the employment/education/life skills services. The only
exception to the pattern of declining need was domestic
violence services, which fluctuated in a less consistent manner
at the 3-, 9-, and 15-month interviews; however, the very low
need in this area makes such trends relatively difficult to
interpret. Importantly, however, although need declined over
time, women’s levels of need for most services remained high.
The second pattern pertains to the relative importance of
different service needs. As shown in Exhibit 29, the sample of
women reported the highest needs for services related to
employment, education, and life skills. This area of need was
the highest at the pre-release interview (as discussed in the
“Pre-release Experiences of Returning Female Prisoners”
section) and remained the strongest need area across each
interview wave. The relative importance of the other service
needs was consistent over time, generally. At each interview
wave, transition services were ranked second in importance,
followed by either health services or child services (depending
on the particular interview wave). Domestic violence−related
services remained the lowest-ranked services over time.
The third pattern pertains to differences in service need
between the SVORI and non-SVORI respondents. As previously
discussed, the level of self-reported need was similar between
the SVORI and non-SVORI respondents at the time of the prerelease interview, indicating that the two groups were
comparable before release. For most of the follow-up period,
the groups continued to report similar levels of service need. At
the final interview (15-months post-release), however, the non-

72

Post-release Experiences

SVORI group reported significantly higher needs for transition
services (p < 0.05). In addition, as will be discussed, significant
differences indicating lower needs for SVORI participants were
evident for several individual services within the bundle areas.
This pattern is difficult to interpret, although it may indicate a
delayed treatment effect, in that the needs of women who
participated in SVORI programming may have been better
addressed over time (so that they no longer reported as many
needs as their situations stabilized). Another possibility
explored is that response bias contributed to this pattern.
Because response rates varied across each interview wave,
researchers considered the possibility that the results for
women who completed a particular interview were different
from the results for women who completed all interviews. The
analyses for the service need outcomes were conducted on the
subsample of women that completed all three follow-up
interviews. The results for this subsample were substantially
the same as those for the full sample at each wave. Combined
with the relatively high response rates and comparability
between SVORI and non-SVORI respondents across waves
(described earlier), response bias apparently was not a problem
and was rejected as an explanation for the difference in selfreported needs among the SVORI and non-SVORI respondents.
Although the data are not shown in the exhibit, comparisons of
service needs over time between the entire female and male
samples were also made. As was observed with the pre-release
data, women continued to have higher self-reported needs than
men, particularly for health- and family-related services (which
were significantly higher for the female sample than for the
male sample at all post-release time periods). Notably,
however, gender differences became decreasingly pronounced
over time in all service areas. For all service areas except
family services, the differences between service need as
reported by men and by women diminished (and for
employment/education and child services the gender difference
actually reversed directions so that men reported higher levels
of need than women at the 9- and 15-month interviews). The
only exception to this pattern was domestic violence services,
which remained significantly higher for women at all time
periods and evinced more pronounced gender differences over
time.

73

Prisoner Reentry Experiences of Adult Females

Transition Services
As shown in Exhibit 30, the women in the sample reported high
levels of need for transition services, which included various
forms of assistance with the reentry process. Self-reported
need for these services was highest at the time of the prerelease interview (at which point the weighted bundle scores
were 74 for the SVORI group and 72 for the non-SVORI group)
and declined with each subsequent interview, reaching their low
at the 15-month interview of 48 for the SVORI group and 56 for
the non-SVORI group. At each time period women’s most
commonly reported needs within this bundle were public health
care insurance (identified as a need by 79% of the women at
the time of the 3-month post-release interview) and financial
assistance (such as short-term loans or housing deposits),
identified as a need by 73% of the women.
Exhibit 30. Transition service needs bundle score
100
SVORI
90

Non-SVORI
All

80
74

72

73

70
60

60

60

60
56
51

52

52

52
48

50

40

30

20

10

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

74

Post-release Experiences

Overall, few differences in self-reported need for transition
services were found between the SVORI and non-SVORI
respondents; however, the finding that SVORI participants were
less likely to report needing public financial assistance at the
pre-release interview was also evident at the 9- and 15-month
interviews. In addition, as mentioned, the overall bundle scores
for transition service needs were significantly lower for the
SVORI group than for the non-SVORI group (48, compared with
56) at the time of the 15-month interview.
Health Services
Self-reported needs for health services also declined slightly at
each interview wave, for both the SVORI and non-SVORI
women. Exhibit 31 shows the bundle scores for health service
needs at each interview wave. At the time of the 15-month
post-release interview, women reported needing 37% of the
services in the bundle. When the two groups were compared,
the only individual health service for which significant
differences in need were observed was anger management.
Women who did not participate in SVORI were significantly
more likely to report needing anger management services at all
waves except the 9-month post-release interview.

75

Prisoner Reentry Experiences of Adult Females

Exhibit 31. Health service needs bundle score
100

SVORI
Non-SVORI

80

All

60

55

57

56

39

40

41

40

37

38

38

39
35

37

20

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

Employment/Education/Skills Services
The service area for which women consistently reported the
greatest need was employment, education, and skills-related
services. On average, women reported needing 80% of the
services in this bundle before release, and 60−64% of the
services after release (Exhibit 32). In terms of individual
services that were most needed, virtually all women reported
needing more education at the pre-release (95%) and 3-month
post-release (93%) interviews. Although the number decreased
slightly by 9 months post-release (87%), education remained
the most commonly reported service need (among all 28
services) at all time periods among both the SVORI and nonSVORI groups. Job training and a job itself closely followed
education as the most frequently reported service needs.
Interestingly, although the SVORI and non-SVORI respondents
were comparable in their self-reported needs within the
employment, education, and skills bundle, at the 15-month

76

Post-release Experiences

Exhibit 32. Employment/education/life skills service needs bundle score
100
SVORI
Non-SVORI
80
80

79

80

All

64

64

64
60

60

60

60

64
57

61

40

20

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

interview the need for life skills was significantly lower among
the women who had enrolled in SVORI programming than
among the women who received “treatment as usual.”
Domestic Violence Services
As shown in Exhibit 33, domestic violence services continued to
be a low ranked service need among the female sample after
the women were released, with the lowest need reported at the
3-month post-release interview (at which point the average
bundle score was only 8). Very few women reported needing a
batterer intervention program (for example, the highest
proportion reported at any time period post-release was only
8% of women, which was reported at the 9 month follow-up
interview) or a domestic violence support group (expressed as
a need by only 15% of the women at 9 months post-release) at
any time period, and there were no differences between the
SVORI and non-SVORI subgroups.

77

Prisoner Reentry Experiences of Adult Females

Exhibit 33. Domestic violence service needs bundle score
100
SVORI
90

Non-SVORI
All

80
70
60
50
40
30
21

21

21

20
12
8

10

8

12

8

12

10

11

10

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

Child Services
Among women who had minor children (three quarters of the
female sample), the 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
over the follow-up period. As shown in Exhibit 34, although
women reported needing half of the child-related services at
the time of the pre-release interview, they reported needing
only 43% at the 3-month interview, 39% at the 9-month
interview, and 38% at the 15-month interview.

78

Post-release Experiences

Exhibit 34. Child service needs bundle score
100
SVORI

90

Non-SVORI
80

All

70
60
52
50

48

50
46
41

43

43
38

40

39

39

38

35

30
20
10
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.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

Levels of Need Across Services
When women’s perceived need for services over time was
examined, women were found to have extremely high need in
the days immediately before release. By 3 months after
release, their perceptions of need had declined substantially
and continued to decline slightly over the following 6 months.
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 women were more education, public health care
insurance, and financial assistance.
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,

79

Prisoner Reentry Experiences of Adult Females

the services that were most commonly reported in the “top
two” across all waves were a job (20–39% of the women), a
place to live (15–24% of the women), more education (15–
21% of the women), and a driver’s license (13–16% of the
women).
As a final indicator of service need, an “all services” bundle,
which captured the level of overall need across all services (in
addition to the service bundles already described), was created.
On average, the respondents reported needing 46–51% of all
the service items during the post-release period (compared
with 64% of items during the pre-release period).

SERVICE RECEIPT
In the Multi-site Evaluation of SVORI, service receipt was
extensively documented. This documentation allowed us to
assess whether SVORI programs were successful in increasing
participants’ access to services beyond access under “treatment
as usual” and to examine the duration of any observed
increased service receipt. In the “Pre-release Experiences”
section, the findings indicated that the SVORI programs
substantially increased access to pre-release services for the
women. For all service areas, women enrolled in SVORI
programs reported significantly higher service receipt during
their periods of incarceration than comparable women not
enrolled in SVORI.
This section examines whether self-reported service utilization
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). The
service receipt bundles mirror the bundles created to show
service need, with the addition of a “coordination services”
bundle (measured only for service receipt), which assesses
receipt of services such as need assessments, case
management, and, for the post-release time periods, postrelease supervision.
Service Receipt Bundle Scores
The results of the weighted outcome analyses and the weighted
means for each service bundle (and individual item) across
waves are shown for the SVORI and non-SVORI groups in
Exhibit 35. Several patterns are clear in the data.

80

—
—
—
—
0.49 0.43
1.00 0.28

0%
1%
9%
28%

—
—
1.63
2.71
*

*

*
*
*
*
*
*
*
*
*

23%
3%
13%
51%

NA
5.68
5.04
3.57
4.48
2.36
2.97
NA
2.33
1.40
NA
—

24.66 28.97 3.12
10%
1.74 0.29
25%
1.62 0.24
19%
1.27 0.25
29%
1.50 0.24
40%
0.86 0.23
25%
1.09 0.24
8.46
6.91 3.03
12%
0.85 0.30
5%
0.34 0.45
9.74 9.66 2.54
2%
—
—

53.64
38%
63%
46%
64%
61%
50%
15.36
23%
7%
19.39
10%

*
*
*

25%
2.53 0.28 12.62 *
—
—
—
—
23.02 17.11 2.20 NA *
3%
—
—
—
13%
0.72 0.30 2.05 *
9%
1.13 0.34 3.10 *
14%
0.52 0.29 1.68
43%
0.61 0.23 1.85 *
10%
1.59 0.30 4.89 *
18%
0.99 0.26 2.68 *
17%
0.38 0.27 1.46
15%
0.63 0.28 1.89 *
29%
0.62 0.23 1.86 *
35.59 12.08 2.68 NA *
8%
0.82 0.34 2.27 *
18%
0.96 0.26 2.60 *
71% −0.15 0.25 0.86
45%
0.74 0.23 2.11 *
34%
0.60 0.23 1.82 *
37%
0.63 0.23 1.88 *

OR
NA
4.44
4.83

81%
—
40.13
19%
24%
25%
22%
59%
36%
37%
23%
25%
43%
47.67
17%
37%
68%
63%
49%
52%

SE
3.44
0.27
0.24

20%
7%
1%
7%

25.69
11%
21%
21%
33%
21%
48%
3.11
5%
1%
5.68
5%

68%
88%
26.33
12%
39%
30%
4%
29%
35%
29%
37%
8%
41%
24.66
8%
4%
39%
18%
32%
48%

8%
4%
7%
7%

—
—
—
—

7.39 18.30
2%
—
6%
—
7%
—
7%
1.87
6%
—
16%
1.60
0.64
2.46
1%
—
0%
—
5.40
0.29
4%
—

18%
2.25
75%
0.89
12.45 13.87
2%
—
30%
0.40
25%
0.22
4%
—
12%
1.12
7%
—
11%
1.15
12%
1.43
6%
—
16%
1.32
14.17 10.49
0%
—
3%
—
44% −0.19
3%
—
17%
0.83
19%
1.36

9.50
2.44
NA
—
1.49
1.25
—
3.05
—
3.16
4.17
—
3.74
NA
—
—
0.83
—
2.28
3.88
*
*

*
*

*
*

*

*
*
*

—
—
—
—

—
—
—
—

2.87 NA *
—
—
—
—
—
—
0.40 6.47 *
—
—
0.31 4.93 *
1.25 NA *
—
—
—
—
2.09 NA
—
—

0.31
0.37
2.31
—
0.29
0.30
—
0.34
—
0.37
0.35
—
0.31
2.36
—
—
0.27
—
0.32
0.30

Wave 2
SVORI Non-S
Mean Mean Est.
SE OR
69.35 36.50 32.85 3.82 NA *
52%
15%
1.78 0.31 5.92 *
74%
38%
1.53 0.30 4.61 *

6%
7%
4%
5%

16.90
10%
14%
13%
24%
18%
23%
4.23
5%
3%
4.98
5%

51%
80%
16.60
10%
25%
25%
7%
18%
22%
13%
22%
7%
17%
25.70
7%
5%
54%
20%
26%
42%

0%
3%
1%
1%

7.58
3%
5%
8%
10%
9%
10%
1.20
2%
1%
1.36
1%

—
—
—
—

9.32
—
—
0.57
0.97
0.79
0.97
3.03
—
—
3.62
—

16% 1.73
56% 1.15
10.19 6.41
1%
—
20% 0.27
26% −0.05
10%
—
10% 0.74
6%
—
7% 0.69
7% 1.28
3%
—
12% 0.39
17.78 7.92
0%
—
7%
—
50% 0.17
15% 0.32
14% 0.80
20% 1.03

—
—
—
—

2.57
—
—
0.42
0.37
0.41
0.36
2.03
—
—
1.68
—

0.31
0.31
2.14
—
0.33
0.31
—
0.41
—
0.44
0.42
—
0.37
2.47
—
—
0.27
0.36
0.32
0.30
*
*

*

*

*
*
*

—
—
—
—

—
—
NA *
—

NA *
—
—
1.76
2.65 *
2.21
2.65 *

5.64
3.14
NA
—
1.31
0.95
—
2.10
—
1.99
3.60
—
1.48
NA
—
—
1.19
1.37
2.24
2.79

Wave 3
SVORI Non-S
Mean Mean Est.
SE OR
55.05 29.77 25.28 4.25 NA *
39%
22% 0.80 0.29 2.23 *
59%
27% 1.34 0.28 3.83 *

12%
7%
2%
8%

14.14
6%
10%
7%
21%
18%
22%
5.96
8%
4%
6.32
7%

39%
54%
13.16
11%
20%
26%
12%
10%
17%
8%
12%
4%
12%
21.69
2%
4%
53%
18%
20%
33%

8%
5%
2%
8%

7.80
5%
7%
6%
9%
11%
11%
1.67
1%
2%
5.17
5%

18%
43%
11.68
4%
24%
27%
8%
10%
7%
8%
8%
4%
18%
18.90
2%
5%
54%
18%
18%
17%

—
—
—
—

6.34
—
0.51
—
1.04
0.60
0.87
4.29
—
—
1.15
—

1.06
0.42
1.48
—
−0.18
−0.02
0.36
−0.04
1.04
—
0.50
—
−0.49
2.79
—
—
−0.03
0.01
0.17
0.90

Wave 4
SVORI Non-S
Mean Mean Est.
39.27 25.47 13.80
36%
21%
0.74
42%
25%
0.77

—
—
—
—

2.25
—
0.46
—
0.38
0.38
0.36
2.24
—
—
1.88
—

0.30
0.26
2.02
—
0.33
0.31
0.44
0.44
0.43
—
0.45
—
0.38
2.37
—
—
0.27
0.35
0.33
0.32

—
—
—
—

NA *
—
1.66
—
2.83 *
1.83
2.39 *
NA
—
—
NA
—

2.90 *
1.53
NA
—
0.83
0.98
1.44
0.96
2.83 *
—
1.65
—
0.61
NA
—
—
0.97
1.01
1.19
2.47 *

SE OR
4.06 NA *
0.30 2.10 *
0.28 2.17 *

Note: Regression results not shown when cell sizes <10. Non-S = Non-SVORI, NA = not applicable. Wave 1 = 30 days pre-release; Wave 2 = 3 months post-release; Wave 3 = 9 months postrelease; Wave 4 = 15 months post-release.
a
Of those with children younger than 18 years of age.
b
Of those who owed back child support.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

Coordination services
Needs assessment
Case manager
Worked with someone to
reintegrate
Currently on probation/parole
Transition services
Financial assistance
Public financial assistance
Public health care
Legal assistance
Documents for employment
Mentoring
Place to live
Transportation
Driver’s license
Access to clothing/food
Health services
Victim support group
Anger management program
Medical treatment
Dental services
Mental health treatment
Substance use treatment
Employment/education/life
skills services
Money management skills
Life skills
Work on personal relationships
Change criminal attitudes
Any educational services
Any employment services
Domestic violence services
Support group
Batterer intervention
Child services
a
Child care
Modification in child support
b
debt
a
Child support payments
a
Modification in custody
a
Parenting classes

Wave 1
Non-S
Mean Est.
29.11 43.56
46%
1.49
33%
1.58

SVORI
Mean
72.67
79%
70%

Exhibit 35. Weighted means and parameter estimates of the effect of SVORI on service receipt

Post-release Experiences

81

Prisoner Reentry Experiences of Adult Females

First, it is clear that the SVORI participants reported
significantly higher service receipt than the non-SVORI
respondents for all types of services and all time periods. The
women who enrolled in SVORI programming reported higher
levels of services not only during their incarcerations (as
measured in the Wave 1 interview), but also at 3, 9, and 15
months post-release. The disparity in service receipt between
the SVORI and non-SVORI groups was particularly evident in
service areas such as coordination and employment, education,
and life skills services; for these bundles, the receipt scores for
the SVORI participants were often twice as high as those for
the women who received “treatment as usual.” Some
exceptions to the pattern of significantly higher service receipt
among the SVORI participants are evident, however. By 15
months post-release, the higher receipt of transition and health
services that had been statistically significant at all previous
interviews were no longer statistically significant. In addition,
the patterns for domestic violence and child services were
relatively unstable (with the SVORI participants reporting
higher levels of receipt but this pattern not being statistically
significant for all time periods), which is what would be
expected because of the very small numbers of women who
received these services.
Second, it is evident that, as in the pattern of self-reported
need for services declining over time, the likelihood of receiving
services declined over time. Specifically, the women were most
likely to report receiving services during their incarceration.
After they were released they were increasingly less likely to
receive services at 3, 9, and 15 months. Notably, however, this
pattern is much more evident among the SVORI participants
than among the non-SVORI respondents. For the women who
received “treatment as usual,” service receipt decreased
dramatically from Wave 1 to Wave 2. However, after the initial
decrease, the level of service receipt remained relatively stable
(and for some services, such as health services, actually
increased slightly over time). In contrast, the women who
participated in SVORI programming generally experienced
substantial decreases in service receipt from each time period
to the next. For health services, level of service receipt
appeared to stabilize after the initial decrease in the initial postrelease period. In addition, the patterns for domestic violence

82

Post-release Experiences

and child services were relatively unstable because of the small
number of women reporting receipt of such services.
The final pattern pertains to gender differences observed
among the sample. Although the data are not shown, when
differences in service receipt between the entire female and
male samples were examined, the women were found to have
reported significantly higher levels of service receipt. This
difference held not only for services received during
incarceration (as previously described), but also for all postrelease time periods. When the service receipt areas were
examined, women were found to have reported significantly
higher receipt of transition, health, employment/education/
skills, and child-related services at the 3- and 15-month time
periods (with coordination/supervision and domestic violence
services also being significantly higher for women at the 15month time period). At the 9-month time period, only transition
and health services were significantly higher for women.
In the sections that follow, individual service bundles are
discussed in more detail and important differences across
bundles and among the individual services that compose each
bundle are highlighted.
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 differ 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
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

83

Prisoner Reentry Experiences of Adult Females

supervision. Although the bundle scores are 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 is nonetheless difficult.
As is evident from Exhibit 35, services related to coordination
were the most commonly reported services among the female
sample. Women reported higher levels of coordination services
than the other service areas considered in the evaluation, and
this pattern was true at all time periods. In addition, although
likelihood of receiving coordination services followed the
general trend observed, in that the likelihood of service receipt
decreased over time, the decrease was much less dramatic for
coordination services. As shown in Exhibit 36, women still
reported relatively high levels of coordination services over
time.
Exhibit 36. Coordination services receipt bundle score
100
SVORI

90

Non-SVORI
80

All

73
69

70
60

54

51

55

50
43
40
30

39

37

33

30

29

25

20
10
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.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

84

Post-release Experiences

When specific types of services within the coordination services
bundle were examined, the most commonly reported “service”
was post-release supervision. The vast majority of both the
SVORI and non-SVORI respondents reported being on postrelease supervision at 3 months post-release. Interestingly,
80% of the SVORI participants were still on supervision at
9 months after release, compared with only 56% of the
comparison women. By 15 months post-release, the disparity
between the two groups was less evident (54%, compared with
43%) yet was still marginally significant (p < 0.10).
As shown in Exhibit 35, large differences between the SVORI
and non-SVORI groups were evident across all coordination
services. Most notably, the proportion of SVORI respondents
who reported having worked with someone to reintegrate back
into the community was more than 3 times as high as that of
non-SVORI respondents. The differences in the likelihood of
having had a needs assessment and having worked with a case
manager were also dramatic. Even 15 months after release, the
women who had participated in SVORI were significantly more
likely to report three of the four services in the coordination
services bundle.
Transition Services
The transition services bundle covers a large number of
services that can help individuals with the transition back into
society. As shown in Exhibits 35 and 37, although the overall
transition service receipt score was significantly higher for the
SVORI participants (compared with the non-SVORI group) at
the Wave 1, 2, and 3 interviews, the two groups appeared to
converge over time, culminating in no statistically significant
difference at Wave 4. At the final interview wave the only
individual service for which the difference between groups was
statistically significant was mentoring services.
Other transition services that appeared to be more commonly
received among the SVORI participants were help attaining
documents for employment and transportation, for which
significant differences were observed across most time periods.
Unfortunately, although public health care insurance and
financial assistance were among the three most frequently
reported service needs among the women (as described
previously), these services were not commonly received. As

85

Prisoner Reentry Experiences of Adult Females

Exhibit 37. Transition services receipt bundle score
100
SVORI
90

Non-SVORI
All

80
70
60
50
40

40
32

30
23

26
20

20
12
10

17

14
10

13

12

12

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

shown in Exhibit 35, less than a third of the sample reported
public health care assistance at a given interview wave, and
hardly any women (particularly those who did not participate in
SVORI) reported assistance accessing financial assistance (such
as short-term loans or housing deposits).
Health Services
Health services, which include substance abuse treatment as
well as physical health services (including medical treatment
and dental services) and mental health services and support
groups, were associated with an interesting pattern of receipt.
As shown in Exhibit 38, after the initial decrease in service
receipt during the months immediately following receipt, the
level of access appeared relatively stable over time (although it
stabilized at a very low level for both groups).

86

Post-release Experiences

Exhibit 38. Health services receipt bundle score
100
SVORI

90

Non-SVORI
80

All

70
60
50

48
42

40

36

30

26

25
20

20

22

22

18

19

Wave 3*

Wave 4

20

14

10
0
Wave 1*

Wave 2*

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

The women who participated in SVORI reported greater
utilization of health-related services at all time periods except
the 15-month post-release time period. The two areas where
the disparity in services was particularly consistent were
substance abuse treatment and mental health treatment (see
Exhibit 35).
Among the health-related services, the ones most commonly
reported were medical treatment (which actually increased
from the 3-month to the 9-month post-release period and then
remained at a relatively high level of access at the 15-month
period) and substance abuse treatment (reported by more than
one third of SVORI participants and approximately one fifth of
comparison women).

87

Prisoner Reentry Experiences of Adult Females

Employment/Education/Skills Services
Services related to employment, education, and skills were very
important to the women. As discussed previously, self-reported
needs were highest for this type of service, particularly for
education. When the actual receipt of such services was
compared with the level of need reported, a high level of unmet
need was evident. For example, although nearly all women
reported, at the pre-release interview, that they needed more
education, only 21% of the SVORI participants and 6% of the
comparison group had received any post-release educational
services at the 3-month follow-up interview.
The most commonly reported services in this service bundle
were employment services, followed by training on how to
change attitudes related to criminal behavior.
Not surprisingly, because many SVORI programs primarily
targeted employment services, consistent differences in
services in the employment, education, and skills bundle were
evident between the SVORI and non-SVORI groups. This
pattern can be observed in Exhibit 39. For all time periods, the
SVORI participants had significantly higher bundle scores for
this service area, and many differences for individual services
were evident. Even 15 months after release, the women who
participated in SVORI were significantly more likely to report
having received employment services and training on how to
change attitudes related to criminal behavior. Once again,
however, the vast majority of programming appeared to be
concentrated at the pre-release stage, with the largest levels of
service receipt being reported at the pre-release interview and
significant group differences being evident at that time point for
every service in the employment, education, and skills services
bundle.
Domestic Violence Services
Domestic violence services, which are limited to participation in
domestic violence support groups and batterer intervention
programs, were very rarely reported among the women. The
small number of women who received them were most likely to
do so while they were incarcerated. Although receipt of
domestic violence services was low, it was consistent with the
extremely low self-reported need for such services among the
sample, indicating a low level of unmet need in this area.

88

Post-release Experiences

Exhibit 39. Employment/education/life skills services receipt bundle score

100
SVORI

90

Non-SVORI
80

All

70
60

54

50
39

40
30

25

26
17

20

17
13

10

14

7

8

8

Wave 2*

Wave 3*

Wave 4*

11

0
Wave 1*

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

The overall pattern for domestic violence services is shown in
Exhibit 40. The differences between the SVORI and non-SVORI
respondents observed at the 3-month interview were significant
at the p < 0.05 level. Differences at the 15-month interview
were marginally significant (p < 0.10).
Child Services
Receipt of child-related services, such as parenting classes,
assistance finding child care, assistance modifying child support
debt, and assistance obtaining child support payments, was
also very infrequently reported among the women. The few
women who reported needing such services makes
identification of consistent trends over time difficult. As with
other service types, however, receipt of child-related services
was most likely while the women were incarcerated. After
release, receipt of such services stabilized at a very low level,
with no large differences between the SVORI and non-SVORI

89

Prisoner Reentry Experiences of Adult Females

Exhibit 40. Domestic violence services receipt bundle score
100
SVORI

90

Non-SVORI
80

All

70
60
50
40
30
20
10

15

12
8
3

1

2

4

1

3

6
2

4

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

samples (except at the 9-month interview, for which the higher
child services bundle score among the SVORI participants was
statistically significant). This pattern is shown in Exhibit 41.
Levels of Receipt Across Services
Overall, the examination of service receipt has shown that the
women who enrolled in SVORI programming received
substantially higher levels of services than women who received
“treatment as usual.” Although programming was concentrated
on the pre-release phase (i.e., levels of service receipt were
dramatically higher at that time period than at any of the postrelease time periods), SVORI appeared to increase access to
services well beyond release. Even 15 months after release, the
SVORI participants still reported significantly higher services
than the non-SVORI respondents in many service areas.

90

Post-release Experiences

Exhibit 41. Child services receipt bundle score
100
SVORI
90

Non-SVORI
All

80

70

60

50

40

30

20

10

19
15
10
6

5

6

5
1

3

6

5

6

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

The services that women 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.
Despite the increase in access, when the services women
actually received were compared with those they reported
needing, a great disparity emerged: very small proportions of
women reported receiving the services they needed most. In
addition, aggregate levels of service receipt were substantially
lower than aggregate levels of service need (across all bundles
and time periods, and among both groups).
The subsections that follow report the impact of SVORI
programming on several key domains. Detailed findings for
housing, employment, family/peer/community outcomes,

91

Prisoner Reentry Experiences of Adult Females

substance abuse and physical and mental health, and criminal
behavior/recidivism are presented.

HOUSING
Several dimensions of housing are relevant as reentry
outcomes. In the SVORI Multi-site Evaluation, three “core”
housing outcomes were housing independence, housing
stability, and the extent of challenges faced in locating housing
after release. The SVORI and non-SVORI groups were
compared on these outcomes at 3, 9, and 15 months postrelease. 11 Exhibit 42 shows the weighted proportion of women
in each group (with estimates, standard errors, and odds ratios
from the logistic regression models) who
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).

As shown in the table, no statistically significant differences
emerged between the SVORI and non-SVORI respondents
along these core housing dimensions, indicating that SVORI
programming did not significantly improve the post-release
housing experiences for returning women prisoners. The
differences are graphically depicted in Exhibits 43–45. As is
evident from the exhibits, the SVORI and non-SVORI groups
were similarly likely to be classified as housing independent, as
having stable housing, and as having experienced no housing
challenges.

11

92

Unlike service needs and receipt, for which Wave 1 values were
displayed (for the purpose of comparing all 4 time periods), several
of the outcomes presented in the remaining subsections were not
measured in a parallel manner at the pre-release interview and
therefore are only presented for Waves 2-4.

OR
0.83
0.85
1.30

Wave 3
SVORI Non-S
Mean Mean Est.
0.78
0.77
0.05
0.64
0.73 −0.41
0.72
0.74 −0.08
SE
0.32
0.29
0.30

OR
1.05
0.67
0.93

Wave 4
SVORI Non-S
Mean Mean Est. SE OR
0.82
0.73 0.52 0.34 1.68
0.65
0.65 0.01 0.28 1.01
0.82
0.76 0.38 0.33 1.46

Note: Non-S = Non-SVORI. Differences between SVORI and non-SVORI were not significant at the 0.05 level. Wave 2 = 3 months post-release; Wave 3 = 9 months post-release;
Wave 4 = 15 months post-release.

Housing independence
Housing stability
No housing challenges

SVORI
Mean
0.65
0.76
0.82

Wave 2
Non-S
SE
Mean Est.
0.69 −0.18 0.29
0.79 −0.17 0.32
0.77
0.26 0.33

Exhibit 42. Weighted means and parameter estimates of the effect of SVORI on housing outcomes

Post-release Experiences

93

Prisoner Reentry Experiences of Adult Females

Exhibit 43. Self-reported housing independence since release/last interview
100%
SVORI
Non-SVORI

90%

All

82%
78%

80%

77%

78%

77%
73%

70%

69%
65%

67%

60%

50%

40%

30%

20%

10%

0%
Wave 2

Wave 3

Wave 4

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

Also of interest in the exhibits are the variable temporal
patterns observed among the three core housing dimensions.
For example, while housing independence improved gradually
over the post-release follow-up period (with two thirds of the
women being classified as “housing independent” at the
3-month period, with a gradual improvement until
approximately three quarters were classified as “housing
independent” at the final interview period), 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 time period). This
pattern may be because the 9- and 15-month post-release
interviews had longer reference periods (6 months) than the 3month post-release interview (3 months) so that respondents
had more opportunities to experience instability during the 9and 15-month interviews. The pattern observed for the
measure of housing challenges, which is perhaps the broadest
94

Post-release Experiences

Exhibit 44. Self-reported housing stability
100%
SVORI
90%

80%

Non-SVORI
79%
76%

All
77%
73%
68%

70%

65%

64%

65%

65%

60%

50%

40%

30%

20%

10%

0%
Wave 2

Wave 3

Wave 4

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

measure of difficulty in finding quality housing, indicates that
the time period from 3 to 9 months post-release was the period
during which women experienced the most challenges, with
their situation appearing to improve by the 15 month postrelease time period.
In addition to the three core housing measures, several other
dimensions of housing were measured in the SVORI Multi-site
Evaluation. Examining the other housing measures illuminates
the women’s overall post-release housing experiences. For
example, one of the individual measures used to create the
“housing independence” measure was whether the respondent
lived in her 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. When the women’s housing situation
was examined, no differences between SVORI and comparison
women were evident, with the two groups being equally likely
to live in their own house or apartment (roughly 33% of both

95

Prisoner Reentry Experiences of Adult Females

Exhibit 45. Self-reported lack of housing challenges since release/last interview
100%
SVORI
Non-SVORI

90%

All

82%
80%

77%

82%

80%

79%
72%

74%

76%
73%

70%

60%

50%

40%

30%

20%

10%

0%
Wave 2

Wave 3

Wave 4

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

groups at 3 months post-release, 36% of both groups at 9
months post-release, and 42% of both groups at 15 months
post-release), to live with someone else (roughly 50% of both
groups at 3 months post-release, 45% of both groups at 9
months post-release, and 45% of both groups at 15 months
post-release), or to be homeless, living in a shelter, or without
a set place to live (roughly 4% of both groups at 3 months
post-release, 6% of both groups at 9 months post-release, and
3% of both groups at 15 months post-release).
For both groups, the “best” housing situation was evident at the
15 month post-release time period, not only according to their
reported housing situation but also according to their
perceptions. At each interview, women were asked whether the
place where they currently lived was better, worse, or about
the same as the last place they lived. The percentage of women
who reported that the place where they were currently living
was better than the place where they used to live reached its

96

Post-release Experiences

highest point (61% for both groups) at the 15-month postrelease interview, further evidencing that women’s housing
situations improved over time.
The other individual components of housing independence also
showed improvements (for both the SVORI and non-SVORI
groups) during the post-release follow-up period. Specifically,
the percentage of women who contributed to housing costs
increased from 59% at 3 months post-release to 73% at 15
months post release, and the percentage of women who
reported that their own name was on the mortgage or lease
increased from 27% (3 months post-release) to 41%
(15 months post-release).
In addition to the ways women’s housing situations changed
during the post-release observation period, how their postrelease housing status compared with the time period
immediately preceding their incarceration is of interest. Among
the core housing outcomes examined, the only individual
measures for which preincarceration information (measured at
the pre-release interview) is available are the woman’s housing
situation (i.e., whether she lived in her 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).
Interestingly, this comparison suggests that the housing
situations among the women were better after release than in
the 6 months before incarceration. For example, before
incarceration, 20% reported as their primary housing situation
that they were homeless, living in a shelter, or had no set place
to live. In contrast, after incarceration, at all time periods, less
than 7% reported being homeless, living in a shelter, or being
without a set place to live). This pattern is shown in Exhibit 46.
In addition to housing independence, stability, and the extent
of housing challenges experienced, which are the core housing
outcomes, other dimensions of housing were documented in the
post-release interviews. One important dimension pertains to
the people with whom the women lived. Depending on the
interview wave, 86–94% of the women were living with
someone else. The most commonly reported categories of
people with whom the women reported living were as follows:
ƒ

their children or stepchildren, as reported by 32%, 34%,
and 39% of the women at the 3-, 9-, and 15-month
interviews, respectively

97

Prisoner Reentry Experiences of Adult Females

Exhibit 46. Self-reported being homeless, living in a shelter, or being without a set place to
live
100%

SVORI

90%

Non-SVORI
All

80%

70%

60%

50%

40%

30%

20%

20%

21%

20%

10%

6%

4%

6%

6%

6%
3%

1%

3%

3%

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

98

ƒ

their husbands or boyfriends, as reported by 20%, 26%,
and 31% of women at the 3-, 9-, and 15-month
interviews, respectively

ƒ

their mothers or stepmothers, as reported by 24%,
17%, and 18% of the women at the 3-, 9-, and 15month interviews, respectively

ƒ

their fathers or stepfathers, as reported by 15%, 11%,
and 9% of the women at the 3-, 9-, and 15-month
interviews, respectively

ƒ

siblings, as reported by 12%, 8%, and 10% of the
women at the 3-, 9-, and 15-month interviews,
respectively

ƒ

other family members, as reported by 12%, 6%, and
7% of the women at the 3-, 9-, and 15-month
interviews, respectively

ƒ

friends, as reported by 9%, 14%, and 14% of women at
the 3-, 9-, and 15-month interviews, respectively

Post-release Experiences

The coresidence patterns were similar for the SVORI and nonSVORI respondents, with the only significant difference being
that the non-SVORI group was significantly more likely to
report living with a boyfriend at the 9-month interview.
Neighborhood quality was a final relevant dimension of housing.
At each post-release interview the women 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 (data not shown), indicating that the women had
similar perceptions of the neighborhoods in which they were
living at each time period at which they were interviewed.
A comparison of the post-release housing situations between
the male and female subsamples suggests several interesting
patterns. No large or consistent differences were evident in the
three core housing outcomes (housing independence, stability,
and extent of challenges experienced). However, women were
more likely than men to experience housing challenges at the
9-month post-release time period (72% of women, compared
with 84% of men, reported not experiencing housing
challenges, p < 0.05). Other gender differences in living
arrangements were also identified. When gender differences in
respondents’ housing situations were examined, women were
significantly more likely than men to have reported living in
their own houses or apartments at both the 3- and 15-month
time periods (while men were significantly more likely than
women to live in another person’s house or apartment—a
finding that was significant at all three time periods). Women
were significantly more likely to be homeless at the 3- and 9month interviews. Interestingly, men were significantly more
likely to contribute to housing expenses at all post-release time

99

Prisoner Reentry Experiences of Adult Females

periods. In terms of with whom the respondents resided, at all
post-release time periods men were significantly more likely
than women to live with their mothers or stepmothers, sisters,
and brothers. In contrast, women were significantly more likely
than men to live with their children at all time periods.

EMPLOYMENT
Employment was a critical reentry outcome for the SVORI
Multi-site Evaluation. Not only did the SVORI programs report
employment as a major focus (confirmed by the higher service
receipt scores for employment-related services as consistently
reported by the SVORI participants), but the women identified
employment services as a major area of need. Several core
employment outcomes were examined:
ƒ

current support of oneself with a job

ƒ

the number of months worked during the reference
period

ƒ

work 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

ƒ

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

The results for these outcomes at 3, 9, and 15 months postrelease are shown in Exhibit 47.
As shown in the exhibit, several findings suggesting more
positive employment outcomes for women who enrolled in
SVORI were evident. Notably, the women who participated in
SVORI programming were significantly more likely than the
non-SVORI respondents to report having supported themselves
with a job at the 15-month time period. As shown in Exhibit 48,
this significant difference is due not only to the steadily
increasing employment observed for SVORI participants, but
also to the sharp decline in employment observed at the
15-month time period for comparison group members.

100

0.49
1.38
0.32
0.72
0.74
0.26

0.53
1.57
0.24
0.88
0.89
0.41

0.16
1.04
0.68

0.16
0.19
−0.40
0.13
0.47
0.37

0.27
0.19
0.38

SE

NA
2.83 *
1.98

1.17
NA
0.67

OR

2.65
0.90
0.41

0.61
3.72
0.42
2.29
0.73
0.33

0.56
3.25
0.46
0.36
1.16
0.34

0.17
0.47
−0.17

Wave 3
SVORI Non-S
Mean Mean Est.
OR

0.30 NA
0.44 3.20 *
0.33 1.41

0.28 1.18
0.36 NA
0.32 0.85

SE

2.76
0.90
0.42

0.68
3.74
0.44
2.10
0.74
0.37

0.45
2.93
0.42

0.66
1.09
0.23

0.95
0.81
0.12

Wave 4
SVORI Non-S
Mean Mean Est.

OR

0.30 NA *
0.46 2.97 *
0.32 1.26

0.28 2.59 *
0.36 NA *
0.32 1.13

SE

Note: Non-S = Non-SVORI, NA = not applicable. Wave 2 = 3 months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

Currently supported self
with job
Number of months worked
Worked each month
Number of months worked
same job
Was receiving formal pay
Job provided benefits

Wave 2
SVORI Non-S
Mean Mean Est.

Exhibit 47. Weighted means and parameter estimates of the effect of SVORI on employment outcomes

Post-release Experiences

101

Prisoner Reentry Experiences of Adult Females

Exhibit 48. Self-reported currently supporting self with job
100%
SVORI
90%

Non-SVORI
All

80%

68%

70%
61%
60%

56%
53%

50%

49%

59%

57%

51%
45%

40%

30%

20%

10%

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

Findings also indicate that, at the 15-month time period,
women who had enrolled in SVORI had also worked
significantly more months since the last interview and more
months at the same job than comparison group members.
These findings are shown graphically in Exhibits 49 and 50.
Once again, it is clear that comparison group members
experienced a “dropoff” in employment at the 15-month postrelease time period, whereas the SVORI participants did not.
In addition, as shown in Exhibit 51, at all follow-up time
periods, the SVORI participants were significantly more likely to
report receiving formal pay for their job, with no evidence of
temporal trends for either group. The SVORI participants were
also slightly more likely to report that their job provided
benefits at the 3-month post-release time period (p = 0.07),
although, as shown in Exhibit 52, this gap narrowed over time

102

Post-release Experiences

Exhibit 49. Self-reported number of months worked since release/last interview
4.0
SVORI

3.7

3.7

Non-SVORI

3.5

All

3.5

3.3

3.2
2.9

3.0

2.5

2.0
1.6
1.5

1.4

1.5

1.0

0.5

0.0
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.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

and was not statistically significant at the 9- and 15-month
interviews.
Overall, the findings for the core employment outcomes
indicate that SVORI programming was associated with
significant improvements on a variety of employment
outcomes, including likelihood of working at all, duration of
employment, employment stability, and employment quality (in
terms of working at jobs for which the women were paid
formally and which offered benefits). The program effects for
outcomes associated with employment status and duration of
employment were evident only at the 15-month post-release
time period—a time at which comparison women appeared to
experience a stark decline in employment. The findings for
outcomes associated with job quality suggest that the jobs held
by the SVORI participants may be higher-quality jobs than
those held by their non-SVORI counterparts. In addition to the
findings presented in Exhibit 47, that the SVORI participants

103

Prisoner Reentry Experiences of Adult Females

Exhibit 50. Self-reported number of months worked at same job since release/last
interview
3.0
SVORI

2.8

Non-SVORI

2.7

All

2.5

2.5

2.4

2.3
2.1
2.0

1.5

1.0

0.9
0.7

0.8

0.5

0.0
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.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

were more likely than the comparison women to report that
they considered their current job to be permanent at the 15month interview further supports this possibility (82% of
SVORI, as opposed to 67% of non-SVORI; p < 0.05).
Interestingly, however, when the SVORI and non-SVORI groups
were compared on job satisfaction and stress, no significant
differences were observed (data not shown).
In addition to identifying program effects, examining how
women’s post-release employment situations compare to their
pre-incarceration employment is of interest. For several of the
employment outcomes, the existence of preincarceration
measures allowed women’s trajectories from 6 months before
incarceration through 15 months after release to be explored.
When the percentage of women who reported supporting
themselves with a job was examined, an analysis of all waves

104

Post-release Experiences

Exhibit 51. Self-reported receipt of formal pay for current job
100%

90%

SVORI
Non-SVORI

90%

89%

90%

All
83%

83%

82%
80%
74%

73%

74%

Wave 2*

Wave 3*

Wave 4*

70%

60%

50%

40%

30%

20%

10%

0%

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

of data indicated that women’s employment in the immediate
post-release period was similar to that in the time period 6
months before incarceration (in the 6 months before
incarceration, 48% of women supported themselves with a job,
compared with 51% in the initial 3 months post-release time
period); however, the employment levels reached at the 9 and
15 month time periods (59% and 57%, respectively) were
notably higher than the preincarceration employment levels.
When job quality was considered, as measured by the jobs’
offering health insurance benefits, no substantial differences
across the time periods were evident. Similarly low percentages
of women reported that their jobs provided benefits at the 6month pre-incarceration time period (32%) as at the postrelease time periods (28−33%).

105

Prisoner Reentry Experiences of Adult Females

Exhibit 52. Self-reported having a job with benefits
100%
SVORI
Non-SVORI

90%

All

80%

70%

60%

50%
41%

42%

41%

40%

37%
34%

30%

40%
37%

33%

26%

20%

10%

0%
Wave 2

Wave 3

Wave 4

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

In addition to the core employment outcomes that have been
discussed, several other dimensions of employment are
relevant. At the time that women were asked whether they
supported themselves with a job, they were also asked about
other sources of financial support, including illegal activities and
support from a government program. SVORI participants were
more likely than comparison group members to report having
received money from a government program (once again, with
this difference being statistically significant only at the 3-month
post-release time period; p < 0.01). Interestingly, SVORI
participants were also significantly less likely than comparison
group members to report receipt of financial support from
family members at all post-release time periods and, at the
3-month time period, from friends. For all women, reliance on
family members appeared to be highest at the 3-month postrelease time period (data not shown), with such support
steadily decreasing over the post-release time periods. Reliance

106

Post-release Experiences

on support from friends appeared to be more stable across the
time periods.
When the post-release employment experiences of men and
women were compared, men were found to fare better than
women on several employment outcomes. At all time periods,
men were significantly more likely than women to report that
they currently supported themselves with a job. At the 3- and
9-month time periods, men also reported significantly more
months worked. Men were also significantly more likely to
report that their jobs offered benefits at the 3- and 9-month
time periods. Interestingly, however, women were significantly
more likely to report that their jobs provided formal pay,
although this difference was observed only at the 15-month
time period.
Very few men and women reported having received money
from illegal activities, across all three post-release time periods,
with no gender differences evident. In addition, unlike the
findings from the pre-release interview, which indicated that
women were more likely to receive financial support from
family and friends during the 6 months before incarceration,
post-release data showed no gender differences in the
likelihood of receiving such support, except that women were
significantly more likely to report support from friends at the
15-month post-release time period. As in the pre-incarceration
findings, at all post-release time periods women were
significantly more likely than men to report having received
financial support from a government program.

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
family-related measures were identified as key outcomes in the
SVORI Multi-site Evaluation. Several aspects of family
relationships were examined, including family emotional
support, family instrumental support, quality of intimatepartner relationships, and quality of relationship with children.
Not surprisingly, given the lack of emphasis on family-related
services among the SVORI programs, no differences were

107

Prisoner Reentry Experiences of Adult Females

observed in these measures between the SVORI and nonSVORI groups at any of the post-release time periods. In
addition, little variation was observed (for both groups) over
the three post-release time periods in these scales, suggesting
that perceptions of family support and relationship quality
remained stable over time.
When the women’s family relationships over time were
examined, some interesting patterns became evident. Although
the proportions of women who reported being currently married
(12–14%) were similar across all time periods (including the
preincarceration time period), women appeared more likely to
report intimate partnerships at all post-release time periods
than during the preincarceration time period. In addition, there
was substantial variation in women’s likelihood of living with a
spouse or current romantic partner at the various time periods.
Women were most likely to report living with their spouse or
partner at the 6-month preincarceration time period (67%);
they were least likely to report living with a spouse or partner
at the 3-month post-release time period (38%), which was
followed by large, steady increases at the 9- (49%) and 15month (58%) time periods.
Women’s experiences as mothers during the post-release time
period are also of interest because a great majority of the
women were mothers. When the women’s likelihood of
reporting that they had primary care responsibilities for their
children was examined, the immediate post-release time period
was evidently particularly challenging for women in this
respect. Although more than half of the mothers reported that
they had primary care responsibilities for their children during
the 6 months before incarceration, only 39% of mothers
reported this level of responsibility at the 3-month post-release
interview, with this percentage steadily increasing at the 9(47%) and 15-month (52%) interviews. As mentioned, the
women were also increasingly more likely to report living with
their children with each post-release time period, suggesting
that it takes time for women to resume the level of care for
their children that they reported during the time period
immediately before incarceration.
Not surprisingly, men and women differed in many ways along
these dimensions of family relationships. Although no gender
differences in marital status were evident, men were

108

Post-release Experiences

significantly more likely to live with their spouses or romantic
partners at 3 months after release. Interestingly, women
scored higher on the summary scale measuring quality of the
intimate partnership at the 3-month post-release time period,
but this difference was not evident 9 or 15 months after
release. No differences were observed between men and
women on the measures of emotional or instrumental support
received by family members.
At all time periods, women were significantly more likely than
men to report having primary care responsibilities for their
children (and, as already mentioned, more likely to live with
their children). In addition, at all time periods women had
significantly higher values on the scale that was created to
measure the quality of the respondent’s relationship with
children (which reflects how much time the respondents spent
with their children and the extent of their involvement in the
children’s lives).
Peer Relationships
Although less researched, particularly for women, peer
relationships may also 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 several 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 deemed appropriate to measure
the “negative exposure” reported by the women from both
friends with whom they spent time and from individuals with
whom they lived. Consequently, a summary measure was
included that reflects the extent to which the respondent
reported living with people who had ever been incarcerated,
used illegal drugs, engaged in any other illegal activity, or used
alcohol in their presence, and reflects the extent to which the
responded reported 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 results for this composite measure of “negative exposure,”

109

Prisoner Reentry Experiences of Adult Females

which is coded such that higher values indicate less negative
exposure, are shown in Exhibit 53.
Exhibit 53. Negative peer exposure (0–14, <better)
14.0
SVORI
Non-SVORI
12.0

11.7
11.2

11.5

All
11.1

10.9

11.0

11.0

10.8

10.9

10.0

8.0

6.0

4.0

2.0

0.0
Wave 2

Wave 3

Wave 4

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

Although at 3 months post-release the SVORI participants had
slightly less exposure to peers or companions who had negative
influence than did the non-SVORI respondents (p < 0.10), this
difference was not significant at the other follow-up periods,
indicating that the women who enrolled in SVORI did not have
better peer exposure outcomes than the comparison women.
For all women, criminogenic exposure was lowest in the
immediate post-release period, appearing to worsen slightly
over time. Indeed, at the 15-month post-release time period,
only 63% of SVORI participants and 53% of comparison group
members (difference not statistically significant) reported that
all or most of their close friends were friends that they could
“hang out with” and know that they would not “get into
trouble.”

110

Post-release Experiences

Positive peer support was also measured. Specifically, the
women 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

ƒ

financial support

Interestingly, the summary measure of peer instrumental
support was significantly higher for the SVORI participants than
for the non-SVORI group at both the 9- and 15-month time
periods, suggesting that women who participated in SVORI
were more likely to count on tangible support from their
friends.
When gender differences for the peer relationship variables
were examined, men and women did not differ in terms of
negative peer exposure or the level of instrumental support
they received from their peers.
Community Involvement
To assess the extent to which women became involved in their
communities after release, women 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); (5) participated in the activities of a church,
mosque, temple, or other religious group; or (6) served in a
Neighborhood Watch or tenant patrol program. Responses were
summed to create a summary measure of “civic action.”
Because most respondents had very little involvement in the
types of civic action listed here, with the exception of church
activities, the summary measure excluded church activities.
As shown in Exhibit 54, civic action was low for both groups,
with only marginal differences between the SVORI and nonSVORI groups at the 3- and 15-month time periods (p < 0.10).
This pattern indicates that the women who received SVORI

111

Prisoner Reentry Experiences of Adult Females

Exhibit 54. Self-reported high civic action since release/last interview
100%
SVORI
Non-SVORI

90%

All
80%

70%

60%

50%

40%

30%

35%

24%
20%

20%

30%

28%

25%

24%

20%

15%

10%

0%
Wave 2

Wave 3

Wave 4

Note: Difference between SVORI and non-SVORI were not significant at the 0.05 level. Wave 1 = 30 days prerelease; Wave 2 = 3 months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.

programming reported slightly more involvement in their
communities than comparable women who did not participate in
SVORI, during the post-release time period. For both groups,
the extent of civic action appeared to increase slightly over the
post-release follow-up period.
At all time periods, men and women had virtually identical
levels of civic action (based on the measure which excluded
church attendance).

SUBSTANCE USE AND PHYSICAL AND
MENTAL HEALTH
Substance Use
Substance use outcomes were measured both by means of selfreport during all follow-up interviews and by oral fluids drug
tests administered to nonincarcerated respondents at the

112

Post-release Experiences

3- and 15-month interviews. The results for the core substance
use outcomes are shown in Exhibit 55.
Self-reported substance use was generally lower for the SVORI
group than for the non-SVORI group; however, these
differences were not statistically significant (although past-30day drug use at the 15-month interview was marginally
significant, p < 0.10). Interestingly, though, when the outcome
that reflects either self-reported or confirmed (by drug tests)
use was examined, the results indicated that the SVORI
participants were significantly less likely to use drugs from
release to 3 months post-release (p < 0.01) and, when the
measure was limited to past-30-day use, less likely to have
used during the previous 30 days at both the 3- and 15-month
post-release time periods.
The patterns for substance use, based on the combined selfreport and drug test measures, are shown graphically in
Exhibits 56 and 57. Of interest is that, not only did the
difference between the number of SVORI participants and the
number of comparison group members who were “clean”
decrease over time, but also fewer women in both groups were
clean at 15 months post-release, suggesting increasing
substance use over time for both groups.
When individual drugs used among the women were examined,
the only consistent difference between the SVORI and nonSVORI respondents was that, at 15 months post-release, the
SVORI participants self-reported significantly lower cocaine use
than the non-SVORI respondents (87% of SVORI participants
did not use cocaine, as opposed to 73% of non-SVORI
respondents; p < 0.05). The drug test results also confirmed
lower cocaine use among the SVORI participants, at both 3 and
15 months post-release (data not shown).
When drug use between men and women during the postrelease follow-up period was compared, no gender differences
were evident. The male and female samples reported
equivalent levels of substance use and tested positive (based
on the drug tests) at similar rates.

113

114
0.75
0.85
0.81

0.87
0.49
0.52

0.78
0.87
0.85

0.92
0.67
0.72

0.86

0.76

0.52

0.29

0.17

0.18

0.29

0.29

0.50

0.38

0.40

0.33

SE

2.35 *

2.15 *

1.69

1.34

1.18

1.19

OR

NA

NA

0.76

0.69

0.69

0.60

NA

NA
NA

NA

0.82 −0.37

0.73 −0.18

0.71 −0.11

0.61 −0.03

Wave 3
SVORI Non-S
Mean Mean Est.

NA

NA

0.35

0.30

0.30

0.28

SE

NA

NA

0.69

0.84

0.89

0.97

OR

0.55

0.46

0.82

0.75

0.77

0.63

0.41

0.38

0.74

0.63

0.67

0.55

0.56

0.33

0.49

0.55

0.53

0.32

Wave 4
SVORI Non-S
Mean Mean Est.

OR

0.27 1.75 *

0.27 1.39

0.34 1.63

0.30 1.74

0.31 1.69

0.28 1.38

SE

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

Note: Non-S = Non-SVORI. NA = not applicable; drug tests were not performed at Wave 33. Wave 2 = 3 months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months
post-release.

No self-reported drug
use
No self-reported drug
use other than
marijuana or steroid
No self-reported drug
use in past 30 days
No self-reported drug
use other than
marijuana or steroids
past 30 days
No self-reported drug
use or positive drug
tests
No self-reported drug
use or positive drug
tests past 30 days

Wave 2
SVORI Non-S
Mean Mean Est.

Exhibit 55. Weighted means and parameter estimates of the effect of SVORI on self-reported substance use outcomes

Prisoner Reentry Experiences of Adult Females

Post-release Experiences

Exhibit 56. No drug use reported or detected since release/last interview

100%
SVORI
Non-SVORI

90%

All

80%

70%

67%
58%

60%

50%

49%

46%
42%
38%

40%

30%

20%

10%

0%
Wave 2*

Wave 4

Note: Information gathered from self-report and drug test. Wave 1 = 30 days pre-release; Wave 2 = 3 months postrelease; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

115

Prisoner Reentry Experiences of Adult Females

Exhibit 57. No drug use reported or detected in past 30 days
100%
SVORI
Non-SVORI

90%

All
80%
72%
70%
63%
60%

55%
52%
49%

50%
41%
40%

30%

20%

10%

0%
Wave 2*

Wave 4*

Note: Information gathered from self-report and drug test. Wave 1 = 30 days pre-release; Wave 2 = 3 months postrelease; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

Physical Health
Physical health services were not a major programmatic focus
among the SVORI programs; therefore, no core physical health
outcomes were identified as being relevant for analysis of
program effects in the evaluation. Of interest, however, is the
physical health status of the women during the post-release
time period. Several dimensions were measured in the postrelease 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

116

Post-release Experiences

activities done as a result of physical health, and pain that
interferes with normal work).
Based on the composite measure reflecting the number of
physical health conditions experienced by the women, the
women reported an average of one condition, with no
differences being evident between the SVORI and non-SVORI
groups and no temporal trend being apparent during the postrelease follow-up period. Likewise, the SVORI and non-SVORI
respondents scored similarly on the SF-12 physical health scale,
with no evidence of temporal trends. As shown in Exhibit 58,
the physical health functioning of both groups remained quite
stable over the entire follow-up period. Additionally, because
the exhibit includes the pre-release time period, it is evident
that no major differences in women’s health status occurred
throughout the entire observation period.
When asked to rate their overall physical health, somewhat
more than a third of women reported their health as being
“excellent” or “very good,” with little variability over the three
post-release time periods (and no differences between the
SVORI and non-SVORI groups). Interestingly, however, as
shown in Exhibit 59, overall perceptions of health appeared to
be slightly higher at the pre-release interview than at the postrelease time periods.
Consistent with the pre-release findings showing that
perceptions of overall physical health were significantly lower
for women than for men, at all post-release time periods
significantly lower proportions of women than men rated their
health as “excellent” or “very good.” Not surprisingly, women
also had significantly lower SF-12 scores and reported a
significantly higher number of physical health diagnoses than
men at all post-release time periods.
Mental Health
Because mental health issues were extremely prevalent among
the female sample (as described in the “Pre-release
Experiences” section), 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).

117

Prisoner Reentry Experiences of Adult Females

Exhibit 58. Physical health scale (12-Item Short-Form Health Survey)
60

SVORI
Non-SVORI

50

All

50
48

49

49

48

48

49

49

49

50
48

49

40

30

20

10

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

According to these key outcomes, there is no evidence that
SVORI programming had an impact on the mental health status
of women. In addition, as in the pattern observed for physical
health, the mental health status of women remained fairly
stable over time. This pattern is shown graphically for the GSI
outcome in Exhibit 60.
Exhibit 61 shows women’s overall assessments of their mental
health, with the pre-release time point shown as a reference
point. As in the pattern indicated by the SF-12 and GSI
outcomes, women’s mental health status appeared quite stable
over time, with no differences between the SVORI and nonSVORI respondents.

118

Post-release Experiences

Exhibit 59. Self-reported overall physical health is very good or excellent
100%
SVORI

90%

Non-SVORI
All

80%

70%

60%

50%
44%
41%
40%

43%

41%
36%

36%

37%

36%

39%

40%

40%

33%
30%

20%

10%

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

Consistent with the data from the pre-release interview, gender
differences in mental health status remained pronounced for
the entire follow-up period. On both the SF-12 mental health
scale and the GSI, women scored significantly lower than men
at all post-release time periods. In addition, at all post-release
time periods, significantly lower proportions of women rated
their overall mental health status as “excellent” or “very good.”

CRIMINAL BEHAVIOR AND RECIDIVISM
In the SVORI evaluation, several measures were used to
determine program effects on desistance from criminal activity.
These measures include a combination of self-reported and
official measures of criminal behavior. Core criminal
behavior/recidivism outcomes based on “unofficial” (i.e., selfreported) data sources are shown in Exhibit 62.

119

Prisoner Reentry Experiences of Adult Females

Exhibit 60. Global Severity Index (45–225, >better)
225
SVORI
Non-SVORI

205

All

185

165

155

155

155

154

159
151

153

153

156

145

125

105

85

65

45
Wave 2

Wave 3

Wave 4

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

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 or
using a weapon. The responses were summed to create the
summary measure. According to this outcome, the women in
the SVORI group were less likely to perpetrate violence than
the women in the comparison group at the 15-month postrelease time period. Exhibit 63 graphically illustrates the
pattern for this outcome, with a significant difference
(p < 0.05) evident between the groups at the final interview
point. Interestingly, when a measure that was parallel to the
perpetration measure but assessed victimization was examined,
no differences between the two groups were evident (data not
shown).

120

Post-release Experiences

Exhibit 61. Self-reported overall mental health “very good” or “excellent”
100%
SVORI

90%

Non-SVORI
All

80%

70%

60%

50%

40%
33%

35%
32%

32%

34%

35%

33%

33%

30%

33%

31%

29%

30%

20%

10%

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

The second core measure of criminal behavior/recidivism was
compliance with conditions of supervision. This outcome is
critical because the majority of women reported being under
post-release supervision throughout the follow-up period. As
discussed previously, more than three quarters of the women
were under post-release supervision at the time of the 3-month
post-release interview. At the 9-month interview, 80% of the
SVORI participants (compared with 56% of the non-SVORI
respondents) were still under post-release supervision. At the
final interview wave, 54% of SVORI participants and 43% of
comparison women were currently under post-release
supervision.

121

122
0.91
0.94 −0.12
0.86 −0.48

0.93
0.93
0.80

0.18

0.83 −0.35
0.82 −0.27

0.78
0.78

0.26

0.71

0.76

0.37

0.55

0.55

0.37
0.34

0.31

SE

0.62

0.89

1.20

0.70
0.76

1.30

OR

0.68

0.78

0.90

0.59
0.69

0.62

0.07

Est.

Wave 3

0.72 −0.17

0.86 −0.53

0.93 −0.27

0.76 −0.79
0.75 −0.31

0.60

SVORI Non-S
Mean Mean

0.29

0.34

0.46

0.36
0.30

0.27

SE

0.84

0.59

0.77

0.45 *
0.73

1.08

OR

0.65

0.77

0.96

0.68
0.70

0.68

1.26

0.35
0.31

0.53

0.65

0.00

0.78 −0.10

0.87

0.60
0.64

0.56

Wave 4
SVORI Non-S
Mean Mean Est.

Note: Non-S = Non-SVORI. Wave 2 = 3 months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.
*p < 0.05 for test of significant difference between SVORI and non-SVORI.

No perpetration of
violence
Complied with
conditions of
supervision
No criminal behavior
No violent or weapons
crimes
Not reincarcerated at
follow up
Not booked or
reincarcerated since
release/last interview

Wave 2
SVORI Non-S
Mean Mean Est.

Exhibit 62. Weighted means and parameter estimates of the effect of SVORI on self-reported recidivism outcomes

0.26

0.30

0.52

0.40
0.27

0.27

SE

1.00

0.91

3.53 *

1.43
1.36

1.70 *

OR

Prisoner Reentry Experiences of Adult Females

Post-release Experiences

Exhibit 63. Self-reported no perpetration of violence since release/last interview
100%
SVORI
90%

Non-SVORI
All

80%

76%
71%

74%
68%

70%
62%
60%

60%

62%

61%
56%

50%

40%

30%

20%

10%

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

As shown in Exhibit 64, among the women who were under
post-release supervision, the SVORI participants were less
likely to report that they had complied with the conditions of
their supervision at all time periods, with the difference being
statistically significant for the 9-month post-release time
period. Compared with 76% of comparison women, only 59%
of SVORI participants reported (at the 9-month post-release
interview) that they had never failed to comply with any
conditions of their supervision. Potential explanations and
implications of this finding are discussed in greater detail in the
conclusions.

123

Prisoner Reentry Experiences of Adult Females

Exhibit 64. Self-reported complying with conditions of supervision since release/last
interview
100%
SVORI
Non-SVORI

90%

All

83%
80%

80%

78%

76%

70%

66%

65%
60%

59%

60%

68%

50%

40%

30%

20%

10%

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

Interestingly, as shown in the exhibit, the pattern for
supervision compliance reversed at the 15-month time period
(with the SVORI participants reporting higher compliance than
the non-SVORI respondents), although the difference was not
statistically significant.
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 Exhibit 62: 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

124

Post-release Experiences

prostitution, soliciting, shoplifting, or disorderly conduct) and
self-reported involvement in violent or weapons offenses.
As shown in the exhibit, results for self-reported criminal
behavior are inconsistent. Differences in any crime are not
statistically significant (and appear to be unstable, because the
SVORI participants appear to have been less likely to report not
committing any crimes at the 3- and 9 month-interviews yet
more likely to report not committing any crimes at the 15month interview). This pattern is shown in Exhibit 65. However,
when subset to violent or weapons crimes only (Exhibit 66), the
results indicate that, at the 15-month post-release time period,
the women who participated in SVORI were significantly more
likely to report not having committed any violent or weapons
crimes than the women in the comparison group (p < .0.05).
Exhibit 65. Self-reported not committing any crimes since release/last interview
100%
SVORI
Non-SVORI

90%
82%
80%

78%

All
80%
75%
72%

69%

70%

70%
67%
64%

60%

50%

40%

30%

20%

10%

0%
Wave 2

Wave 3

Wave 4

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

125

Prisoner Reentry Experiences of Adult Females

Exhibit 66. Self-reported not committing any violent or weapons crimes since release/last
interview

SVORI

100%
93%

91%

92%

90%

Non-SVORI
93%

All

96%
92%

91%
87%

90%

80%

70%

60%

50%

40%

30%

20%

10%

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

The final criminal behavior/recidivism outcomes based on selfreported data are whether the respondent had not been
reincarcerated at the time of her follow-up interview and a
composite measure reflecting whether the respondent had not
been reincarcerated at the time of the follow-up interview and
reported that she had not been booked into jail or prison (for
24 hours or more) during the reference period. The latter
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. As shown
in Exhibits 67 and 68, no significant differences on these
outcomes were observed between SVORI and comparison
women for any follow-up period. In other words, SVORI and
comparison women appeared equally likely to be reincarcerated
during the follow-up period.

126

Post-release Experiences

Exhibit 67. Not reincarcerated at follow-up interview
100%
93%

94%

SVORI

93%

Non-SVORI

90%

86%

All
81%

78%

80%

77%

78%

77%

70%

60%

50%

40%

30%

20%

10%

0%
Wave 2

Wave 3

Wave 4

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

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, as
described in Lattimore and Steffey, 2009) and reincarceration
in state prison (obtained from the state Departments of
Corrections). The core recidivism measures based on official
records are shown in Exhibit 69.
The findings for rearrest (shown graphically in Exhibit 70)
indicate that the women in the SVORI and non-SVORI groups
were equally likely to be rearrested within 3 and 6 months of
release but that the SVORI participants were significantly less
likely to be rearrested within 9, 12, 15, and 21 months of
release. When type of rearrest (considering person/violent
crimes, property crimes, drug crimes, public order crimes, and
other crimes) at the 21- and 24-month time periods was
examined, no significant differences were evident.

127

Prisoner Reentry Experiences of Adult Females

Exhibit 68. Not booked or reincarcerated since release/last interview
100%
SVORI
90%

Non-SVORI

86%

All

83%
80%

80%
72%
68%

70%

70%
65%

65%

65%

60%

50%

40%

30%

20%

10%

0%
Wave 2

Wave 3

Wave 4

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

The findings for reincarceration (shown in Exhibit 71) indicate
that the SVORI and non-SVORI respondents were equally likely
to be reincarcerated within 3, 6, and 9 months of release but
that the SVORI participants were significantly more likely to be
reincarcerated within 12, 15, 21, and 24 months of release.
When gender differences in recidivism were examined, the data
suggested that women were less likely to recidivate than men.
Although no gender differences were evident during the
immediate post-release time period (i.e., the first 3 months
after release), at both the 9- and 15-month time periods
women were less likely to report having committed a
violent/weapon crime, less likely to report having been
incarcerated at the time of interview, and less likely to report
having been incarcerated during the reference period or at the
time of interview. Interestingly, no gender differences were
evident in self-reported perpetration of violence or compliance
with supervision conditions.

128

Post-release Experiences

Exhibit 69. Official measures of recidivism

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

SVORI
Mean

NonSVORI
Mean

Est.

SE

OR

12%
18%
25%
29%
33%
44%
49%
7%
17%
15%
32%
4%
8%
19%
16%
34%
2%

14%
26%
36%
42%
49%
59%
60%
12%
22%
21%
40%
7%
12%
24%
21%
41%
7%

−0.16
−0.46
−0.52
−0.56
−0.67
−0.63
−0.46
−0.62
−0.26
−0.41
−0.36
−0.59
−0.48
−0.31
−0.36
−0.30
−1.11

0.34
0.30
0.26
0.25
0.24
0.24
0.24
0.46
0.29
0.31
0.24
0.58
0.40
0.28
0.30
0.24
0.58

0.85
0.63
0.59
0.57
0.51
0.53
0.63
0.54
0.77
0.66
0.70
0.56
0.62
0.73
0.69
0.74
0.33

4%
10%
15%
24%
30%
36%
41%

2%
8%
11%
14%
17%
21%
22%

0.76
0.24
0.38
0.69
0.75
0.74
0.87

0.76
0.40
0.33
0.29
0.26
0.25
0.24

2.13
1.27
1.46
1.99
2.13
2.09
2.38

*
*
*
*

*
*
*
*

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

When gender differences in official measures of recidivism were
examined, the data showed that significantly lower proportions
of women than men were arrested at all time periods after the
first 3 months of release (e.g., significant differences in
likelihood of rearrest were observed for 6, 9, 12, 15, 21, and 24
months post-release). When the types of crimes for which
individuals were arrested were considered, gender differences
were most pronounced for person/violent and drug crimes—
both of which were significantly higher for men. By official
measures of reincarceration, women were also significantly less
likely than men to be reincarcerated within 9, 12, 15, 21, and
24 months of release.

129

Prisoner Reentry Experiences of Adult Females

Exhibit 70. Months to first rearrest
100%
SVORI

90%

Non-SVORI

80%
70%
60%

59%

60%

49%

49%

50%

44%

42%
40%

36%
33%

29%

30%

26%

25%

18%

20%

14%

12%
10%
0%
3

6

9

12*

15*

21*

24

*p < 0.05 for test of significant difference between SVORI and non-SVORI.
Exhibit 71. Months to first reincarceration
100%
90%

SVORI
Non-SVORI

80%
70%
60%
50%

41%
40%

36%
30%

30%

24%

20%
10%

15%
10%
4%

8%

11%

21%
14%

22%

17%

2%

0%
3

6

9

12*

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

130

15*

21*

24

Conclusions
The findings in this report substantiate prior research indicating
that female prisoners returning to their communities are a
population with extremely high needs (see, e.g., Mallik-Kane &
Visher, 2008). The women included in the SVORI Multi-site
Evaluation had numerous physical and mental health problems,
extensive substance abuse histories, substantial experience
with the criminal justice system, extensive exposure to drug or
criminally involved family members and peers, and substantial
housing challenges. Furthermore, when asked about their
needs for specific services, the women reported extremely high
levels of service need across most service areas.
SVORI funding offered correctional agencies an opportunity to
intervene by providing a range of services designed to facilitate
successful reentry for returning prisoners. Although almost all
of the adult SVORI programs served both women and men, the
extent to which programming was customized for women is
unclear. Among the 11 impact sites that served both men and
women, the great majority of participants identified for
inclusion in the impact study were male and, according to
information gathered during site visits, little difference was
apparent between the program models (service delivery
approaches) employed for men and those for women. In other
words, gender-specific programming was not emphasized.
Although the customization of programming for women was not
emphasized across impact sites, the programs included
components that have been found to be important for
successful reentry for women (e.g., housing and employment
services; La Vigne, Brooks, & Shollenberger, 2008; O’Brien,
2001). Importantly, the programs were successful in increasing
the services provided to female participants. The women who
participated in SVORI reported substantially higher levels of

131

Prisoner Reentry Experiences of Adult Females

service receipt than comparable women not enrolled in SVORI,
across almost all types of services. In addition, the higher
levels of service receipt found for SVORI participants persisted
even after their release (with significantly higher levels of
service receipt documented at 3, 9, and 15 months postrelease), which suggests that the SVORI programs were better
than “treatment as usual” at linking women with services in the
community. The sustained post-release service delivery is
particularly notable when compared with the pattern observed
among male participants, for whom differences in service
receipt, as reported by SVORI participants and comparison
group members, generally were not observed beyond 9 months
post-release.
Importantly, however, the levels of services that female SVORI
participants received, although a significant improvement over
“treatment as usual,” failed to match their high levels of need.
Levels of service receipt were substantially lower than service
need, indicating that very small proportions of women received
the services they needed. Even though women’s perceptions of
service need declined over the post-release follow-up period,
the extent of services they received declined concomitantly, so
levels of unmet need remained extremely high.
Nevertheless, even though service receipt was low in
comparison with the women’s high needs, participation in
SVORI was indeed linked to improved reentry outcomes in
several dimensions. Employment and substance abuse were the
domains for which the most consistent positive outcomes were
observed.
With respect to employment outcomes, which were strongest at
the 15-month post-release time period, women who enrolled in
SVORI programs were more likely to report that they were
supporting themselves with a job, were more likely to have
worked significantly more months, had significantly higher
employment stability (as measured by months worked at the
same job), were more likely to receive formal pay, and were
less likely to report having received money from illegal
activities (although this finding was observed only for the 3month post-release time period). The successful impact on
employment is to be expected because employment was the
programmatic focus most commonly reported by the SVORI
program directors—a finding confirmed by the high levels of

132

Conclusions

employment service receipt reported by the women who
participated in SVORI. That, for many dimensions of
employment, differences were not statistically significant until
the 15-month time period is, however, somewhat surprising.
This pattern may have emerged because, while employment
“success” steadily increased over time for the SVORI group, it
“dropped off” for the non-SVORI group at the 15-month postrelease time period. Overall, the positive improvements across
a variety of employment domains support the conclusion that
the SVORI programs were effective in improving employment
outcomes.
The findings for substance abuse similarly suggest positive
outcomes for SVORI participants. Using a rigorous measure of
substance use, which combines self-reported data and oral
fluids drug test results, it was found that at both 3 and 15
months after release, women who enrolled in SVORI were
significantly less likely than the comparison women to have
used drugs during the reference period and during the 30 days
preceding assessment. This finding is particularly important
because of the extensive substance abuse histories reported by
the women and the fact that, for women overall, abstinence
from substance use appears to become increasingly more
difficult throughout the post-release follow-up period. The
women who participated in SVORI reported much higher receipt
of substance abuse treatment services than comparison women
at all time periods, which suggests that the substance abuse
services delivered through SVORI had an impact on postrelease abstinence from use.
Although the SVORI programs positively influenced women’s
employment and substance use outcomes, they did not appear
to have an impact on several relevant reentry domains. SVORI
programming was not associated with successful housing
outcomes (which gradually improved for all women throughout
the post-release follow-up period), family and peer
relationships (which remained relatively stable throughout the
post-release period), physical health (which remained relatively
stable throughout the post-release period), and mental health
(which also remained relatively stable throughout the postrelease period). For these domains, women enrolled in SVORI
and women receiving “treatment as usual” had similar
outcomes. When levels of service receipt for services directly
relevant to these outcomes are reviewed (e.g., assistance

133

Prisoner Reentry Experiences of Adult Females

finding a place to live, medical treatment, domestic violence
services, and child services), these findings are not necessarily
surprising, because no significant differences were observed
between the SVORI and non-SVORI groups (or, for domestic
violence and child services, such a small number of women
reported service receipt that minimum cell size requirements
for regression analyses were not met). In addition, SVORI
program director reports of programmatic emphasis confirm
that family, physical health, and mental health services were
not major foci for their adult programs. This is unfortunate
given the importance prior research has placed on the
importance of familial relationships in the reentry process (e.g.,
La Vigne et al., 2008; O’Brien, 2001; Richie, 2001) and the
high levels of both physical and mental health problems
reported by returning women.
While conclusions about the effectiveness of SVORI are
relatively consistent for the domains just discussed, they are
more mixed for criminal behavior and recidivism outcomes.
Women who participated in SVORI programs had better
outcomes than comparison subjects for several dimensions of
criminal behavior/recidivism. Specifically, SVORI participants
reported significantly fewer violent/weapon crimes (at 15
months post-release) and reported significantly less
perpetration of violence (at 15 months post-release) than
comparison group members. 12 SVORI-enrolled women also had
significantly lower rearrest rates (within 9, 12, 15, and 21
months of release), based on official arrest records, than their
non-SVORI counterparts. Although the differences in arrest
rates were not significantly different at 3, 6, and 24 months
post-release, the pattern is the same; fewer SVORI than
comparison women were arrested. Despite these findings,
SVORI participants were significantly less likely to report that
they complied with their supervision conditions (at 9 months
post-release) and had significantly higher reincarceration rates
(within 12, 15, 21, and 24 months of release) according to
official corrections data.
Two explanations for the findings of lower rearrest and higher
reincarceration are plausible. First, the role of site-level effects
cannot be ruled out. As discussed in the “Design and Methods”
12

134

When commission of any crimes was examined, no significant
differences were observed in the percentages of the SVORI and
non-SVORI respondents who reported committing a crime.

Conclusions

section, the women were not evenly distributed across the 11
sites and, within a site, the SVORI and non-SVORI groups were
not evenly distributed. This pattern is exacerbated by the fact
that the overall number of women is relatively small (n = 357).
Consequently, state-level policies or practices may have had an
impact on the reincarceration findings yet be unrelated to other
outcomes. For example, state reincarceration rates, postrelease supervision rates, post-release supervision intensity,
and violation rates may have had a major impact on
reincarceration.
The disproportionate composition of the SVORI and non-SVORI
groups within the impact sites may at least partly explain the
apparently contradictory pattern observed for rearrest and
reincarceration if site practices or policies are associated with
reincarceration (which is almost certainly the case). For
example, reincarceration rates varied substantially across
states. Because many states contained women only (or
primarily) from either the SVORI or the non-SVORI groups, the
ability to distinguish between program effects and state
practices was confounded. Several analyses were conducted in
an effort to examine this hypothesis (for example, some
analyses were conducted excluding Indiana, which included half
of the comparison women; other analyses included Indiana as a
control variable), but similar results were found. In the end, the
ability to examine this issue was limited by the distribution of
cases across site.
A second possible explanation is that SVORI program
participants were more likely than comparison subjects to have
been at risk for post-release supervision revocation—because
either they were more likely to be on supervision, or they were
subject to more conditions of supervision. The SVORI
participants were more likely than the non-SVORI respondents
to report being under post-release supervision at all time points
(with the difference being particularly dramatic at 9 months
post-release, when 80% of the SVORI participants compared to
only 56% of the comparison women reported being on
supervision). Supervision clearly carries with it the threat of
violation (for failure to comply with supervision requirements)
that may result in reincarceration. Consequently, SVORI
program participants may have been at higher risk of
revocation than comparison subjects.

135

Prisoner Reentry Experiences of Adult Females

To the extent that violations (and revocations) can occur
without an associated arrest, technical revocations would
explain why reincarceration (yet not rearrest) rates were higher
for the SVORI group. However, the administrative data
obtained from some probation and parole agencies was limited
(and did not consistently include supervision start/end dates
and violation data), which constrained the options for exploring
this possibility. The self-reported data that were therefore
relied upon for supervision status (1) are available only for
women who completed a particular follow-up interview (unlike
the official rearrest and reincarceration data, which were
obtained for the full pre-release sample) and (2) reflect only
the women’s “current” (i.e., at the time of the interview)
supervision status. With these constraints in mind, the
reincarceration outcome models were estimated, controlling for
self-reported 3-month post-release supervision status. Similar
results were found—significantly higher reincarceration rates for
the SVORI group than for the non-SVORI group.
A related consideration is that the conditions of supervision for
the SVORI program participants were more substantial than for
the comparison subjections—meaning, essentially, that overall
supervision intensity could have contributed to the higher
reincarceration rate of SVORI participants. Although primarily
focused on the pre-release phase, some SVORI programs hired
specialized parole officers to provide post-release supervision to
SVORI participants. In some sites, the intent was for these
parole officers to have reduced caseloads, with the goal of
providing more intense supervision. It is possible, therefore,
that SVORI participants experienced greater supervision
intensity (being subject to more supervision conditions, more
drug testing, more frequent meetings, etc.), which made
compliance with supervision more difficult and increased the
chance that any noncompliance would be detected. The finding
that SVORI participants were more likely to self-report that
they had failed to comply with their supervision conditions
(found at 9 months post-release) may support this hypothesis
(in that their supervision conditions could have been more
difficult to comply with), although it is also possible that they
simply were less compliant than the non-SVORI group.
When the models for the outcomes of reincarceration were run,
controlling for self-reported number of supervision conditions at
3 months post-release (the time period most likely to reflect

136

Conclusions

post-release supervision following release from the instant
incarceration), the results for treatment status did not change.
However, when the same model was run controlling for selfreported number of supervision conditions at 15 months postrelease (the interview wave closest to the first time period at
which significant differences in reincarceration were identified
for the two groups), treatment status was no longer significant.
This finding suggests that the number of supervision conditions
required for SVORI participants at the 15-month post-release
time period may have contributed to their greater likelihood of
reincarceration (potentially through technical revocations not
associated with an arrest). Further investigation into the role of
supervision intensity (based on both self-reported and official
probation/parole data) is warranted.
Despite the mixed results found for criminal behavior/recidivism
outcomes, the positive findings observed for employment and
substance use support the conclusion that the SVORI programs
were successful in improving some reentry outcomes for
women. The data have shown that enhanced access to a
variety of reentry services resulted in modest improvements.
These promising outcomes suggest that the reentry efforts
initiated through SVORI funding provide a strong foundation for
future reentry efforts. Indeed, many SVORI grantee agencies
indicated in 2007 that they were continuing the efforts initiated
through SVORI funding, expanding and updating their programs
to reflect evolutions in thinking about prisoner reentry that
have taken place in the years since the SVORI funds were
awarded in 2002.
Several policy and practice implications for service providers
and agencies working with returning female prisoners are
suggested by the findings. First, the current evaluation’s
detailed documentation of service areas for which women
reported high needs can be used for effective planning and
service delivery. Women consistently identified services related
to employment, education, and skills as their top area of need.
The individual services identified as a need by the most women
were more education, job training, and a job itself. Large
proportions of women also reported needing public health care
insurance and financial assistance. The identification of these
“high-need” service areas can be used by supervision and
community-based service providers who work with formerly
incarcerated women.

137

Prisoner Reentry Experiences of Adult Females

Second, because of the variety of challenges that returning
women prisoners face, particularly with respect to mental and
physical health problems, extensive family responsibilities, and
lack of employment experience (as compared with that of
reentering male prisoners), effective coordination of services is
necessary. Services should be not only available but also
accompanied by close coordination among populations with
extremely high levels of need, so that the effectiveness of
services may be maximized. Appropriate identification of needs,
treatment planning, and follow-up may be particularly
important for women. Indeed, Bloom has also highlighted the
importance of coordinating a range of comprehensive and
collaborative services and even includes it as a “guiding
principle” of gender-responsive strategies for women offenders
(Bloom et al., 2003, p. 82).
Finally, the temporal patterns observed during the follow-up
periods in the SVORI evaluation may have implications for
supervision and service delivery. For some outcomes, declines
over time (i.e., the further “out” from release) appear to be the
natural course of events. For example, substance use and
criminal behavior/recidivism appeared to worsen over time for
both SVORI participants and comparison group members. For
other outcomes, however, improvements over time seem to be
the predominant pattern. For example, women’s housing
situation, employment quality, resumption of primary care
responsibilities for their children, and extent of community
involvement appear to gradually improve over time, indicating
that many women “find their footing” in these dimensions.
These findings may be used to help identify the appropriate
point (and area of need) at which to intervene. The time period
around 9 months after release appears to be a particularly
challenging time for women, according to the data. For
example, the extent of housing challenges faced (including
homelessness) was higher at 9 months post-release than at 3
months or 15 months. In addition, self-reported drug use was
highest at the 9-month post-release time period. Future work
that not only extends the post-release follow-up period beyond
15 months but also documents women’s reentry experiences at
more frequent intervals would further inform decisions about
the most influential points of intervention.

138

References
Bloom, B., Owen, B., & Covington, S. (2003). Genderresponsive strategies: Research, practice, and guiding
principles for women offenders. Washington, DC: U.S.
Department of Justice, National Institute of Corrections.
Covington, S. S. (2003). A woman’s journey home: Challenges
for female offenders. In J. Travis & M. Waul (Eds.),
Prisoners once removed: The impact of incarceration
and reentry on children, families, and communities (pp.
67–103). Washington, DC: Urban Institute Press.
Giordano, P. C., Cernkovich, S. A., & Rudolph, J. L. (2002).
Gender, crime, and desistance: Toward a theory of
cognitive transformation. American Journal of Sociology,
107, 990–1064.
Glaze, L. E. (2008). Parents in prison and their minor children.
Washington, DC: U.S. Bureau of Justice Statistics.
Koons, B. A., Burrow, J. D., Morash, M., & Bynum, T. (1997).
Expert and offender perceptions of program elements
linked to successful outcomes for incarcerated women.
Crime & Delinquency, 43(4), 512–532.
La Vigne, N. G., Brooks, L. E., & Shollenberger, T. L. (2008).
Women on the outside: Understanding the experiences
of female prisoners returning to Houston, Texas.
Washington, DC: Urban Institute.
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.

139

Prisoner Reentry Experiences of Adult Females

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.
Lindquist, C. (2005). Reentry research in action:
Implementation of SVORI programs. Research Triangle
Park, NC: RTI International.
Mallik-Kane, K., & Visher, C. A. (2008). Health and prisoner
reentry: How physical, mental, and substance abuse
conditions shape the process of reintegration.
Washington, DC: Urban Institute.
Morash, M., Bynum, T. S., & Koons, B. A. (1998). Women
offenders: Programming needs and promising
approaches. Washington, DC: National Institute of
Justice.
O’Brien, P. (2001). Just like baking a cake: Women describe the
necessary ingredients for successful reentry after
incarceration. Families in Society, 82, 287–295.
O’Brien, P., & Harm, N. J. (2002). Women’s recidivism and
reintegration: Two sides of the same coin. In J. FigueiraMcDonough & R. C. Sarri (Eds.), Women at the margins:
Neglect, punishment, and resistance (pp. 295–315).
Binghamton, NY: Haworth Press.
Richie, B. (2001). Challenges incarcerated women face as they
return to their communities: Findings from life history
interviews. Crime & Delinquency, 47, 368–389.
Rubin, D. B. (2006). Matched sampling for causal effects. New
York: Cambridge University Press.
Sabol, W. J., & Couture, H. (2008). Prison inmates at midyear
2007. Washington, DC: U.S. Bureau of Justice Statistics.
Seiter, R. P., & Kadela, K. R. (2003). Prisoner reentry: What
works, what does not, and what is promising. Crime &
Delinquency, 49(3), 360-388.
Strategic Advantages (2000). Symptom Assessment-45
Questionnaire (SA-45). New York: Multi-Health Systems.
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.
West, H. C., & Sabol, W. J. (2008). Prisoners in 2007.
Washington, DC: U.S. Bureau of Justice Statistics.

140

References

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., 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.

141

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

Total All Cases

Case Disposition—Eligible Cases
Completed
Interview completed
Released Early
R released before Wave 1 interview
Refused
Final refusal by R, guardian or other
Access Denied
Access to R denied by prison
Other Non-Interview
R absconded
Physically/mentally incapable
Other non-interview
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
Other ineligible
Total Ineligible Cases

SVORI
N
%
264
44.82%
SVORI
% of
N
Eligible

Non-SVORI
N
%
325
55.18%
Non-SVORI
% of
N
Eligible

All Cases
N
%
589
100%
All Cases
% of
N
Eligible

153

69.86%

204

68.69%

357

69.19%

48

21.92%

66

22.22%

114

22.09%

12

5.48%

26

8.75%

38

7.36%

2

0.91%

1

0.34%

3

0.58%

0
0.00%
0
0.00%
0
0.00%
297
100.00%
Non-SVORI
% of
N
Eligible

1
2
1
516

1
0.46%
2
0.91%
1
0.46%
219
100.00%
SVORI
% of
N
Eligible

0.19%
0.39%
0.19%
100.00%
All Cases
% of
N
Eligible

2
1

4.44%
2.22%

5
0

17.86%
0.00%

7
1

9.59%
1.37%

12
1
2
24
3
45

26.67%
2.22%
4.44%
53.33%
6.67%
100.00%

10
0
3
9
1
28

35.71%
0.00%
10.71%
32.14%
3.57%
100.00%

22
1
5
33
4
73

30.14%
1.37%
6.85%
45.21%
5.48%
100.00%

Note: R = respondent.

A-1

A-2

Characteristic
Demographics and Housing
Age at incarceration
Age at pre−release (Wave 1) interview
White
Black
Hispanic
Multiracial/other
Born in United States
English was primary language
Homeless/shelter/no set place to live before incarceration
Employment History
Ever held a job
Employed during 6 months before incarceration
Source of support 6 months before incarceration: Family
Source of support 6 months before incarceration: Friends
Source of support 6 months before incarceration: Government
Source of support 6 months before incarceration: Illegal income
Source of support 6 months before 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 before incarceration

Exhibit A-2. Respondent characteristics, by group

(0.22)
(0.50)
(0.50)
(0.45)
(0.46)
(0.50)
(0.21)
(16.14)
(8.41)
(0.44)
(0.42)
(0.47)
(0.49)
(0.42)
(0.12)

(7.17)
(6.85)
(0.50)
(0.49)
(0.24)
(0.29)
(0.11)
(0.18)
(0.41)

0.14 (0.35)
0.71 (0.45)

0.95
0.53
0.49
0.27
0.30
0.45
0.04
39.55
10.16
0.74
0.78
0.32
0.62
0.23
0.01

357
357
357
357
357
357
357
187
179
188
188
185
357
357
357
357
357

29.69
31.41
0.44
0.41
0.06
0.10
0.99
0.97
0.22

Full Sample
Mean (SD)

357
357
357
357
357
357
357
357
357

N

(0.26)
(0.50)
(0.50)
(0.41)
(0.43)
(0.50)
(0.21)
(13.91)
(6.64)
(0.39)
(0.29)
(0.48)
(0.45)
(0.43)
(0.11)

(6.62)
(6.13)
(0.50)
(0.48)
(0.27)
(0.29)
(0.11)
(0.19)
(0.39)

0.17 (0.38)
0.71 (0.45)

0.93
0.51
0.46
0.21
0.25
0.50
0.05
39.49
9.42
0.82
0.91
0.36
0.71
0.25
0.01

28.86
31.10
0.48
0.35
0.08
0.09
0.99
0.96
0.19

SVORI
Mean (SD)

(0.19)
(0.50)
(0.50)
(0.47)
(0.47)
(0.49)
(0.21)
(17.59)
(9.44)
(0.46)
(0.47)
(0.46)
(0.50)
(0.41)
(0.12)

(7.51)
(7.34)
(0.49)
(0.50)
(0.22)
(0.30)
(0.10)
(0.17)
(0.43)

0.12 (0.32)
0.72 (0.45)

0.96
0.54
0.50
0.32
0.34
0.42
0.04
39.58
10.66
0.69
0.68
0.30
0.55
0.21
0.01

30.31
31.64
0.41
0.45
0.05
0.10
0.99
0.97
0.24

Non−SVORI
Mean (SD)

(continued)

1.38
−0.07

−1.31
0.58
−0.76
−2.31
−1.84
1.63
0.86
−0.04
−1.03
2.02
4.14
0.86
3.09
0.84
−0.13

−1.90
−0.76
1.32
−1.78
1.11
−0.21
−0.29
−0.51
−1.04

t−statistic

Prisoner Reentry Experiences of Adult Females

Characteristic
Family and Peers (continued)
Currently married or in steady relationship
Lived with spouse/partner before incarceration
Had any living children
Number of children (only respondents with children)
Number of children (respondents with and without children)
Had child or children younger than 18
Primary care responsibilities for any children younger than 18
years old 6 months before incarceration
Number of children younger than 18 supported 6 months before
incarceration
Required to pay child support 6 months before incarceration
Made court-ordered child support payments 6 months before
incarceration
Court order for support changed while incarcerated
Owed back child support
Dollar amount of back child support owed
State had forgiven/decreased back child support
Had people in life considered family
Had a family member who had been convicted of a crime
Had a family member who had been in a correctional facility
Had a family member who had had problems with drugs/alcohol
Family emotional support scale (0–30: > more support)
Had a friend (before incarceration) who had been convicted of a
crime
Had a friend (before incarceration) who had been in a correctional
facility
Had a friend (before incarceration) who had had problems with
drugs or alcohol
Physical and Mental Health
Physical health scale (>better)
Mental health scale (>better)
0.76 (0.43)
0.74 (0.44)
0.79 (0.41)
49.11 (12.17)
43.80 (12.65)

329
339
351
351

(0.49)
(0.47)
(0.30)
−8639
(0.19)
(0.17)
(0.41)
(0.40)
(0.39)
(5.86)

324

0.39
0.30
0.91
6688
0.04
0.97
0.78
0.80
0.82
20.96

0.71 (1.10)
0.13 (0.34)

124
269
36
33
32
24
27
357
341
342
344
345

0.54 (0.50)

270

(0.50)
(0.48)
(0.37)
(1.65)
(1.83)
(0.43)

Full Sample
Mean (SD)
0.48
0.65
0.83
2.79
2.33
0.76

N
353
171
357
298
357
357

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

(0.50)
(0.48)
(0.36)
(1.60)
(1.78)
(0.41)

(0.51)
(0.46)
(0.23)
−6760
(0.24)
(0.16)
(0.41)
(0.39)
(0.38)
(6.12)

50.47 (11.50)
46.05 (12.66)

0.80 (0.40)

0.70 (0.46)

0.78 (0.42)

0.45
0.28
0.95
6219
0.06
0.97
0.79
0.81
0.83
21.14

0.80 (1.25)
0.17 (0.37)

0.55 (0.50)

0.48
0.66
0.85
2.78
2.37
0.79

SVORI
Mean (SD)
(0.50)
(0.48)
(0.38)
(1.69)
(1.87)
(0.44)

(0.48)
(0.49)
(0.38)
−12073
0.00
(0.17)
(0.42)
(0.41)
(0.40)
(5.66)

48.10 (12.57)
42.14 (12.42)

0.79 (0.41)

0.78 (0.42)

0.75 (0.43)

0.31
0.33
0.85
7625
0.00
0.97
0.77
0.78
0.81
20.83

0.64 (0.95)
0.11 (0.31)

0.54 (0.50)

0.49
0.65
0.82
2.80
2.30
0.73

Non−SVORI
Mean (SD)

(continued)

1.81
2.89

0.10

−1.47

0.66

0.83
−0.34
0.95
−0.37
1.00
0.18
0.45
0.56
0.59
0.49

0.79
1.39

0.14

−0.02
0.17
0.66
−0.07
0.32
1.32

t−statistic

Appendix A — Data Tables

A-3

A-4

Characteristic
Physical and Mental Health (continued)
Received treatment for mental health problem before this
incarceration
Global Severity Index (45–225: >worse)
Positive Symptom Total (0–45: >worse)
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 had asthma
Was receiving treatment for asthma
Was taking prescription for asthma
Ever had diabetes
Currently had diabetes
Was receiving treatment for diabetes
Was taking prescription for diabetes
Ever had heart trouble
Currently had heart trouble
Was receiving treatment for heart trouble
Was taking prescription for heart trouble
Ever had high blood pressure
Currently had high blood pressure
Was receiving treatment for high blood pressure

N

357
357
357
357
357
357
356
357
357
357
356
357
357
357
357
80
80
356
356
24
24
357
355
26
26
357
355
40

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

0.50
79.01
17.39
9.57
10.26
7.37
8.94
9.48
9.59
7.72
7.62
8.48
0.37
0.28
0.22
0.74
0.71
0.10
0.07
0.79
0.67
0.11
0.07
0.38
0.35
0.17
0.11
0.78

(0.50)
(31.55)
(11.33)
(4.41)
(4.96)
(3.63)
(4.85)
(4.69)
(4.38)
(4.08)
(3.63)
(4.16)
(0.48)
(0.45)
(0.42)
(0.44)
(0.46)
(0.30)
(0.25)
(0.41)
(0.48)
(0.32)
(0.26)
(0.50)
(0.49)
(0.37)
(0.32)
(0.42)

Full Sample
Mean (SD)

0.44
74.90
16.08
9.28
9.39
7.04
8.46
9.28
9.11
7.28
7.15
7.91
0.46
0.27
0.20
0.73
0.77
0.09
0.04
0.83
0.83
0.08
0.06
0.56
0.44
0.14
0.09
0.79

(0.50)
(29.85)
(11.09)
(4.34)
(4.58)
(3.33)
(4.48)
(4.59)
(4.25)
(3.79)
(3.39)
(3.91)
(0.50)
(0.44)
(0.40)
(0.45)
(0.43)
(0.28)
(0.20)
(0.41)
(0.41)
(0.27)
(0.24)
(0.53)
(0.53)
(0.35)
(0.29)
(0.43)

SVORI
Mean (SD)

0.55
82.09
18.37
9.78
10.91
7.62
9.30
9.62
9.95
8.05
7.97
8.92
0.31
0.29
0.25
0.74
0.68
0.11
0.09
0.78
0.61
0.14
0.08
0.29
0.29
0.19
0.13
0.77

(0.50)
(32.50)
(11.43)
(4.46)
(5.14)
(3.83)
(5.10)
(4.78)
(4.45)
(4.27)
(3.77)
(4.30)
(0.46)
(0.45)
(0.43)
(0.44)
(0.47)
(0.31)
(0.28)
(0.43)
(0.50)
(0.34)
(0.28)
(0.47)
(0.47)
(0.39)
(0.34)
(0.43)

Non−SVORI
Mean (SD)

(continued)

−1.96
−2.14
−1.89
−1.06
−2.89
−1.50
−1.60
−0.68
−1.79
−1.76
−2.11
−2.28
2.90
−0.44
−1.10
−0.06
0.82
−0.70
−1.92
0.28
0.98
−1.81
−0.90
1.30
0.74
−1.35
−1.10
0.12

t−statistic

Prisoner Reentry Experiences of Adult Females

Characteristic
Physical and Mental Health (continued)
Was taking prescription for high blood pressure
Ever had arthritis
Currently had arthritis
Was receiving treatment for arthritis
Was taking prescription for arthritis
Ever had chronic back pain
Currently had chronic back pain
Was receiving treatment for chronic back pain
Was taking prescription for chronic back pain
Ever had tuberculosis
Tuberculosis was currently active
Ever diagnosed as being HIV positive or having AIDS
Was receiving treatment for HIV/AIDS
Was taking prescription for HIV/AIDS
Ever had hepatitis B or C
Currently had hepatitis B or C
Was receiving treatment for hepatitis B or C
Was taking prescription for hepatitis B or C
Was wearing glasses or corrective lenses
Needed eyeglasses
Currently used a hearing aid
Needed 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

N
40
357
357
44
44
357
357
76
76
357
357
357
6
6
355
352
39
39
357
207
357
356
357
281
281
281
281
281

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

0.78
0.14
0.12
0.18
0.30
0.24
0.21
0.20
0.25
0.03
0.00
0.02
0.83
0.67
0.12
0.11
0.08
0.03
0.41
0.31
0.00
0.05
0.79
0.22
0.15
0.03
0.26
0.02

(0.42)
(0.35)
(0.33)
(0.39)
(0.46)
(0.43)
(0.41)
(0.40)
(0.44)
(0.18)
0.00
(0.13)
(0.41)
(0.52)
(0.33)
(0.31)
(0.27)
(0.16)
(0.49)
(0.46)
0.00
(0.22)
(0.41)
(0.41)
(0.36)
(0.18)
(0.44)
(0.13)

Full Sample
Mean (SD)
0.79
0.13
0.10
0.20
0.33
0.24
0.22
0.24
0.33
0.04
0.00
0.01
1.00
1.00
0.12
0.10
0.07
0.00
0.42
0.27
0.00
0.06
0.79
0.16
0.16
0.04
0.21
0.02

(0.43)
(0.34)
(0.30)
(0.41)
(0.49)
(0.43)
(0.41)
(0.44)
(0.48)
(0.19)
0.00
(0.08)
0.00
0.00
(0.32)
(0.30)
(0.26)
0.00
(0.49)
(0.45)
0.00
(0.24)
(0.41)
(0.37)
(0.37)
(0.20)
(0.41)
(0.13)

SVORI
Mean (SD)
0.77
0.15
0.14
0.17
0.28
0.25
0.21
0.16
0.19
0.03
0.00
0.02
0.80
0.60
0.13
0.12
0.08
0.04
0.40
0.34
0.00
0.04
0.79
0.26
0.14
0.03
0.30
0.02

(0.43)
(0.36)
(0.35)
(0.38)
(0.45)
(0.43)
(0.41)
(0.37)
(0.39)
(0.17)
0.00
(0.16)
(0.45)
(0.55)
(0.34)
(0.33)
(0.28)
(0.20)
(0.49)
(0.47)
0.00
(0.21)
(0.41)
(0.44)
(0.35)
(0.16)
(0.46)
(0.14)

Non−SVORI
Mean (SD)

(continued)

0.12
−0.57
−1.28
0.22
0.39
−0.07
0.11
0.86
1.47
0.51
NA
−1.42
NA
NA
−0.31
−0.59
−0.19
−1.00
0.31
−1.00
NA
0.64
0.04
−2.19
0.31
0.74
−1.61
−0.14

t−statistic

Appendix A — Data Tables

A-5

A-6

Characteristic
Physical and Mental Health (continued)
Ever received care for: Depression/dysthymia
Ever received care for: Drug abuse/dependence
Ever received care for: Obsessive−compulsive 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 was receiving treatment: Alcohol abuse/dependence
Currently was receiving treatment: Anxiety disorder
Currently was receiving treatment: Attention deficit hyperactivity
disorder
Currently was receiving treatment: Bipolar disorder
Currently was receiving treatment: Conduct disorder
Currently was receiving treatment: Depression/dysthymia
Currently was receiving treatment: Drug abuse/dependence
Currently was receiving treatment: Obsessive−compulsive
disorder
Currently was receiving treatment: Oppositional defiant disorder
Currently was receiving treatment: Posttraumatic stress disorder
Currently was receiving treatment: Phobia (social or specific)
Currently was receiving treatment: Schizophrenia
Currently was receiving treatment: Other problem/diagnosis
Currently was not receiving treatment for any condition
Doctor prescribed medication for emotional/psychological problem
during this incarceration
Received the prescribed medication
Any victimization (6 months before incarceration)
Victimization severity before incarceration (0–30: >worse)
Any victimization (during incarceration)
Victimization severity during incarceration (0–36: >worse)
0.02
0.15
0.00
0.13
0.12
0.01
0.01
0.04
0.01
0.05
0.04
0.60
0.33
0.96
0.62
5.69
0.41
1.83

260
260
260
260
260
260
260
260
260
260
260
260
357
118
357
357
357
357

(0.47)
(0.20)
(0.49)
(7.38)
(0.49)
(3.63)

(0.11)
(0.09)
(0.19)
(0.11)
(0.21)
(0.20)
(0.49)

(0.12)
(0.36)
0.00
(0.34)
(0.32)

(0.47)
(0.50)
(0.16)
(0.10)
(0.25)
(0.12)
(0.25)
(0.39)
(0.27)
(0.26)
(0.23)

Full Sample
Mean (SD)
0.33
0.43
0.02
0.01
0.07
0.01
0.07
0.18
0.08
0.07
0.05

N
281
281
281
281
281
281
281
281
281
260
260

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

0.37
0.95
0.62
5.97
0.40
1.86

0.01
0.01
0.05
0.03
0.05
0.06
0.50

0.03
0.13
0.00
0.22
0.14

0.38
0.34
0.02
0.02
0.08
0.02
0.07
0.17
0.12
0.06
0.09

(0.48)
(0.23)
(0.49)
(8.08)
(0.49)
(3.70)

(0.10)
(0.10)
(0.21)
(0.17)
(0.21)
(0.23)
(0.50)

(0.17)
(0.34)
0.00
(0.42)
(0.35)

(0.49)
(0.48)
(0.16)
(0.13)
(0.28)
(0.16)
(0.25)
(0.38)
(0.32)
(0.25)
(0.29)

SVORI
Mean (SD)

0.31
0.97
0.62
5.49
0.42
1.81

0.01
0.01
0.03
0.00
0.05
0.03
0.66

0.01
0.16
0.00
0.07
0.10

0.29
0.50
0.03
0.01
0.06
0.01
0.07
0.19
0.06
0.08
0.03

(0.46)
(0.18)
(0.49)
(6.83)
(0.50)
(3.58)

(0.11)
(0.08)
(0.18)
0.00
(0.21)
(0.18)
(0.47)

(0.08)
(0.37)
0.00
(0.26)
(0.30)

(0.45)
(0.50)
(0.16)
(0.08)
(0.23)
(0.08)
(0.25)
(0.39)
(0.23)
(0.27)
(0.16)

Non−SVORI
Mean (SD)

(continued)

1.13
−0.60
−0.03
0.59
−0.43
0.12

−0.29
0.24
0.55
1.75
0.01
0.86
−2.69

1.23
−0.77
NA
3.30
1.00

1.64
−2.73
−0.01
0.78
0.85
1.20
−0.09
−0.30
1.73
−0.43
2.14

t−statistic

Prisoner Reentry Experiences of Adult Females

Substance Use
Ever drank any type of alcoholic beverage
Age at first drink
Used alcohol 30 days before 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 before this incarceration
Number of drugs used 30 days before this incarceration
Used drugs other than marijuana and steroids 30 days before this
incarceration
Ever used sedatives
Age first used sedatives
Used sedatives 30 days before this incarceration
Age last used sedatives
Ever used tranquilizers
Age first used tranquilizers
Used tranquilizers 30 days before this incarceration
Age last used tranquilizers
Ever used stimulants
Age first used stimulants
Used stimulants 30 days before this incarceration
Age last used stimulants
Ever used pain relievers
Age first used pain relievers
Used pain relievers 30 days before this incarceration
Age last used pain relievers
Ever used methadone
Age first used methadone
Used methadone 30 days before this incarceration

Characteristic

0.60
0.25
18.57
0.11
25.39
0.36
20.92
0.17
26.03
0.23
17.60
0.09
25.08
0.35
20.16
0.17
26.88
0.10
26.44
0.03

357
357
91
357
51
357
129
356
69
356
82
356
49
357
124
356
64
356
34
356

(0.49)
(0.44)
(5.70)
(0.31)
(7.21)
(0.48)
(6.70)
(0.37)
(6.73)
(0.42)
(4.29)
(0.29)
(6.65)
(0.48)
(6.38)
(0.37)
(7.66)
(0.29)
(6.87)
(0.16)

(0.19)
(4.53)
(0.50)
(7.59)
(0.24)
(2.96)
(0.47)
(1.93)

Full Sample
Mean (SD)
0.96
14.47
0.53
27.24
0.94
4.12
0.68
1.71

N
357
341
355
153
357
357
357
357

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

0.58
0.29
18.71
0.12
24.93
0.39
20.65
0.15
24.95
0.27
17.46
0.09
25.39
0.37
19.74
0.17
25.83
0.09
24.50
0.03

0.97
13.86
0.51
26.37
0.95
4.44
0.66
1.63
(0.49)
(0.46)
(5.95)
(0.32)
(7.33)
(0.49)
(6.33)
(0.36)
(5.62)
(0.45)
(4.47)
(0.28)
(6.61)
(0.49)
(5.43)
(0.38)
(7.01)
(0.29)
(5.53)
(0.18)

(0.16)
(4.59)
(0.50)
(6.98)
(0.21)
(3.04)
(0.48)
(1.82)

SVORI
Mean (SD)

0.61
0.23
18.43
0.10
25.92
0.34
21.16
0.18
27.28
0.20
17.73
0.10
24.67
0.33
20.52
0.16
27.79
0.10
27.80
0.02

0.95
14.94
0.54
27.98
0.93
3.89
0.70
1.77
(0.49)
(0.42)
(5.51)
(0.30)
(7.20)
(0.48)
(7.05)
(0.39)
(7.73)
(0.40)
(4.15)
(0.30)
(6.85)
(0.47)
(7.11)
(0.37)
(8.19)
(0.30)
(7.51)
(0.14)

(0.22)
(4.44)
(0.50)
(8.04)
(0.25)
(2.89)
(0.46)
(2.01)

Non−SVORI
Mean (SD)

(continued)

−0.59
1.47
0.23
0.44
−0.49
0.95
−0.43
−0.75
−1.45
1.52
−0.28
−0.40
0.37
0.77
−0.70
0.18
−1.02
−0.22
−1.40
0.74

1.15
−2.20
−0.66
−1.31
0.93
1.74
−0.72
−0.69

t−statistic

Appendix A — Data Tables

A-7

A-8

Characteristic
Substance Use (continued)
Age last used methadone
Ever used anabolic steroids
Age first used anabolic steroids
Used anabolic steroids 30 days before this incarceration
Age last used anabolic steroids
Ever used marijuana
Age first used marijuana
Used marijuana 30 days before this incarceration
Age last used marijuana
Ever used hallucinogens
Age first used hallucinogens
Used hallucinogens 30 days before this incarceration
Age last used hallucinogens
Ever used cocaine
Age first used cocaine
Used cocaine 30 days before this incarceration
Age last used cocaine
Ever used heroin
Age first used heroin
Used heroin 30 days before this incarceration
Age last used heroin
Ever used amphetamines
Age first used amphetamines
Used amphetamines 30 days before this incarceration
Age last used amphetamines
Ever used inhalants
Age first used inhalants
Used inhalants 30 days before this incarceration
Age last used inhalants

Full Sample
Mean (SD)
29.72 (6.79)
0.00 0.00
0.00
0.00 0.00
0.00
0.90 (0.93)
14.73 (14.04)
0.43 (0.52)
25.05 (23.47)
0.44 (0.50)
17.94 (4.40)
0.05 (0.22)
21.80 (5.13)
0.75 (0.44)
20.31 (6.34)
0.43 (0.50)
26.92 (7.65)
0.22 (0.41)
22.35 (5.77)
0.06 (0.24)
24.64 (5.72)
0.35 (0.48)
18.66 (5.50)
0.18 (0.38)
25.48 (7.57)
0.18 (0.39)
17.77 (5.72)
0.01 (0.07)
20.17 (6.51)

N
25
356
0
356
0
357
319
354
165
357
156
356
138
357
265
356
110
357
77
357
56
357
125
357
62
357
65
357
63

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

28.00 (4.50)
0.00 0.00
0.00
0.00 0.00
0.00
0.90 (0.30)
14.49 (3.68)
0.41 (0.49)
23.89 (7.10)
0.48 (0.50)
18.04 (4.71)
0.06 (0.24)
21.65 (5.05)
0.74 (0.44)
19.25 (5.07)
0.37 (0.48)
26.23 (7.61)
0.27 (0.44)
21.32 (5.14)
0.06 (0.24)
23.81 (5.38)
0.41 (0.49)
18.87 (6.46)
0.18 (0.39)
26.66 (8.04)
0.22 (0.41)
18.06 (6.45)
0.00 0.00
20.82 (7.23)

SVORI
Mean (SD)
30.69 (7.76)
0.00 0.00
0.00
0.00 0.00
0.00
0.89 (0.31)
14.92 (3.62)
0.45 (0.50)
26.01 (7.83)
0.40 (0.49)
17.85 (4.12)
0.04 (0.21)
21.95 (5.23)
0.75 (0.43)
21.08 (7.04)
0.48 (0.50)
27.66 (7.70)
0.18 (0.38)
23.53 (6.28)
0.06 (0.24)
25.75 (6.07)
0.31 (0.46)
18.45 (4.34)
0.17 (0.38)
23.96 (6.75)
0.16 (0.36)
17.47 (4.95)
0.01 (0.10)
19.47 (5.66)

Non−SVORI
Mean (SD)

(continued)

−0.95
NA
NA
NA
NA
0.30
−1.06
−0.71
−1.81
1.54
0.26
0.62
−0.34
−0.24
−2.47
−2.21
−0.98
2.04
−1.70
0.00
−1.26
2.02
0.43
0.28
1.40
1.43
0.41
−1.42
0.82

t−statistic

Prisoner Reentry Experiences of Adult Females

Characteristic
Current Incarceration and Criminal History
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
Wave 1 conviction offense(s) category: Drug crime
Drug dealing/manufacturing
Drug possession
Other drug offense
Wave 1 conviction offense(s) category: Public order crime
Wave 1 conviction offense(s) category: Other crime
Current incarceration for probation or parole violation
Current incarceration for probation violation
Current incarceration for parole violation
Parole violation: Technical violation
Parole violation: New crime
Age at first arrest
Number of lifetime arrests
Number of lifetime convictions
Number of lifetime convictions/age at incarceration
Ever locked up in a juvenile correctional facility for committing a
crime

N

356

357
356
356
356
356
356
356
356
356
356
356
356
356
356
356
356
356
356
356
357
357
357
75
75
347
329
333
333

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

(2.06)
(0.46)
(0.27)
(0.31)
(0.21)
(0.16)
(0.21)
(0.49)
(0.24)
(0.37)
(0.16)
(0.41)
(0.21)
(0.45)
(0.33)
(0.38)
(0.15)
(0.41)
(0.11)
(0.44)
(0.24)
(0.40)
(0.49)
(0.49)
(6.07)
(9.27)
(5.64)
(0.21)
0.33 (0.47)

1.71
0.29
0.08
0.11
0.04
0.03
0.05
0.41
0.06
0.16
0.03
0.22
0.04
0.29
0.12
0.18
0.02
0.21
0.01
0.26
0.06
0.20
0.63
0.37
19.14
10.58
5.36
0.19

Full Sample
Mean (SD)
(2.13)
(0.45)
(0.27)
(0.28)
(0.24)
(0.18)
(0.21)
(0.50)
(0.29)
(0.38)
(0.18)
(0.43)
(0.21)
(0.47)
(0.38)
(0.38)
(0.18)
(0.38)
(0.14)
(0.46)
(0.22)
(0.44)
(0.50)
(0.50)
(5.71)
(8.52)
(5.11)
(0.20)
0.30 (0.46)

2.22
0.27
0.08
0.08
0.06
0.03
0.05
0.43
0.09
0.17
0.03
0.24
0.05
0.32
0.17
0.17
0.03
0.18
0.02
0.31
0.05
0.25
0.54
0.46
19.02
9.56
4.95
0.18

SVORI
Mean (SD)
(1.92)
(0.46)
(0.27)
(0.34)
(0.18)
(0.14)
(0.22)
(0.49)
(0.18)
(0.37)
(0.14)
(0.41)
(0.21)
(0.44)
(0.28)
(0.39)
(0.12)
(0.43)
(0.07)
(0.42)
(0.25)
(0.36)
(0.45)
(0.45)
(6.34)
(9.75)
(6.02)
(0.21)
0.36 (0.48)

1.33
0.31
0.08
0.13
0.03
0.02
0.05
0.39
0.03
0.16
0.02
0.21
0.04
0.26
0.09
0.19
0.01
0.24
0.00
0.23
0.07
0.16
0.74
0.26
19.24
11.36
5.68
0.19

Non−SVORI
Mean (SD)

(continued)

−1.32

4.11
−0.63
−0.01
−1.32
1.06
0.74
−0.15
0.80
2.14
0.31
0.74
0.64
0.06
1.22
2.23
−0.42
1.07
−1.37
1.20
1.74
−0.63
2.25
−1.79
1.79
−0.33
−1.75
−1.19
−0.72

t−statistic

Appendix A — Data Tables

A-9

A-10

Note: NA = not applicable.

Characteristic
Current Incarceration and Criminal History (continued)
Number of times in juvenile lockup (only those who reported ever
being locked up)
Number of times in juvenile lockup (all respondents)
Ever been in jail/prison more than 24 hours at one time
Number of times sent to prison (only those who reported ever
having been in prison)
Number of times sent to prison (all respondents)
Any disciplinary infractions during this incarceration
One disciplinary infraction during this incarceration
Two or more disciplinary infractions during this incarceration
Placed in administrative segregation during this incarceration
Was currently a gang member
Considered gang to be family
Relatives were members of the gang
Any perpetration of violence (6 months before incarceration)
4.50 (5.87)
1.44 (3.93)
0.83 (0.37)
1.53
1.20
0.52
0.17
0.34
0.33
0.02
0.83
0.50
0.66

112
349
357
293
352
356
356
356
357
356
6
6
357

(3.03)
(2.05)
(0.50)
(0.38)
(0.48)
(0.47)
(0.13)
(0.41)
(0.55)
(0.47)

Full Sample
Mean (SD)

N

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

1.42
1.14
0.60
0.18
0.42
0.41
0.02
0.67
0.67
0.65

(2.15)
(2.01)
(0.49)
(0.39)
(0.49)
(0.49)
(0.14)
(0.58)
(0.58)
(0.48)

3.84 (4.00)
1.10 (2.75)
0.80 (0.40)

SVORI
Mean (SD)

1.60
1.25
0.45
0.17
0.29
0.27
0.01
1.00
0.33
0.67

(3.53)
(2.09)
(0.50)
(0.37)
(0.45)
(0.44)
(0.12)
0.00
(0.58)
(0.47)

4.91 (6.78)
1.70 (4.61)
0.86 (0.35)

Non−SVORI
Mean (SD)

−0.55
−0.48
−2.79
0.38
2.63
−2.72
0.35
−1.00
0.71
−0.26

−1.06
−1.52
−1.36

t−statistic

Prisoner Reentry Experiences of Adult Females

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 training
Job
More education
Money management skills
Life skills
Work on personal relationships
Change in 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

Service

0.78
0.56
0.65
0.32
0.52
0.83
0.83
0.95
0.70
0.74
0.78
0.75
0.17 (0.38)
0.25 (0.43)
0.45
0.86
0.40
0.70
0.38

356
356
357
357
357
357
355
357
357
356
357
356
357
357
267
29
267
270
269

(0.5)
(0.35)
(0.49)
(0.46)
(0.49)

(0.37)
(0.37)
(0.22)
(0.46)
(0.44)
(0.41)
(0.44)

(0.41)
(0.50)
(0.48)
(0.47)
(0.50)

0.54
0.87
0.68
0.91
0.83
0.60
0.55
0.77
0.79
0.76

(0.50)
(0.34)
(0.47)
(0.29)
(0.37)
(0.49)
(0.50)
(0.42)
(0.41)
(0.43)

Full Sample
Mean (SD)

356
355
354
356
357
356
357
356
356
357

N

(0.35)
(0.34)
(0.21)
(0.45)
(0.45)
(0.42)
(0.44)

(0.41)
(0.50)
(0.48)
(0.46)
(0.49)

(0.50)
(0.30)
(0.49)
(0.27)
(0.35)
(0.49)
(0.50)
(0.38)
(0.38)
(0.42)

0.48
0.83
0.41
0.69
0.44

(0.50)
(0.38)
(0.49)
(0.47)
(0.50)

0.18 (0.39)
0.22 (0.42)

0.86
0.87
0.95
0.73
0.73
0.78
0.74

0.79
0.55
0.66
0.29
0.42

0.54
0.90
0.61
0.92
0.86
0.58
0.58
0.82
0.82
0.77

SVORI
Mean (SD)

Exhibit A-3. Proportion of respondents who reported needing specific services, by group

(0.39)
(0.39)
(0.23)
(0.47)
(0.44)
(0.41)
(0.43)

(0.42)
(0.50)
(0.48)
(0.47)
(0.49)

(0.50)
(0.36)
(0.45)
(0.30)
(0.39)
(0.49)
(0.50)
(0.45)
(0.42)
(0.43)

0.43
0.91
0.39
0.72
0.34

(0.50)
(0.30)
(0.49)
(0.45)
(0.47)

0.16 (0.37)
0.27 (0.45)

0.81
0.81
0.95
0.69
0.74
0.79
0.75

0.77
0.57
0.64
0.34
0.59

0.54
0.85
0.73
0.90
0.81
0.61
0.54
0.73
0.77
0.75

Non−SVORI
Mean (SD)

0.76
−0.56
0.44
−0.57
1.68

0.53
−1.12

1.23
1.55
0.35
0.80
−0.21
−0.26
−0.14

0.34
−0.42
0.35
−0.88
−3.31

−0.09
1.58
−2.45
0.77
1.06
−0.46
0.67
2.15
1.27
0.36

t−statistic

Appendix A — Data Tables

A-11

A-12

Service
Coordination Services
Needs assessment
Case manager
Release needs assessment
Reentry plan
Release planning
Transition Services
Release preparation programs
Release preparation classes
Help accessing financial assistance
Help accessing public financial assistance
Help accessing public health care
Help obtaining legal assistance
Help with documents
Mentoring
Help finding transportation
Help finding a place to live
Help getting a driver’s license
Help accessing clothing/food banks
Health Services
Any medical treatment
Medical treatment for physical health
Preventive medical services
Dental services
Prescription meds for physical health
Info on accessing health care
Any mental health treatment
Individual mental health counseling
Group mental health counseling
Info on accessing mental health care
Any substance use treatment
Alcoholics Anonymous or Narcotics
Anonymous
Drug education
Group substance use counseling
Individual substance use counseling
0.61
0.49
0.44
0.42
0.51
0.72
0.63
0.10
0.17
0.16
0.18
0.51
0.22
0.21
0.27
0.20
0.36
0.71
0.55
0.55
0.53
0.58
0.26
0.41
0.29
0.13
0.34
0.43
0.36
0.30
0.21
0.16

357
356
357
357
357
357
356
357
357
357
357
356
357
357
357
357
357
355
357
357
357
356
357
357
357
357
357

(0.48)
(0.46)
(0.41)
(0.36)

(0.46)
(0.50)
(0.50)
(0.50)
(0.49)
(0.44)
(0.49)
(0.46)
(0.34)
(0.47)
(0.50)

(0.45)
(0.48)
(0.30)
(0.38)
(0.37)
(0.38)
(0.50)
(0.41)
(0.41)
(0.44)
(0.40)
(0.48)

(0.49)
(0.50)
(0.50)
(0.49)
(0.50)

Full Sample
Mean (SD)

352
355
349
350
357

N

0.43
0.44
0.29
0.24

0.69
0.57
0.54
0.63
0.58
0.35
0.48
0.36
0.14
0.46
0.54

0.90
0.83
0.20
0.22
0.26
0.22
0.59
0.38
0.25
0.38
0.27
0.43

0.81
0.71
0.69
0.73
0.82

(0.50)
(0.50)
(0.46)
(0.43)

(0.47)
(0.50)
(0.50)
(0.48)
(0.50)
(0.48)
(0.50)
(0.48)
(0.35)
(0.50)
(0.50)

(0.31)
(0.38)
(0.40)
(0.42)
(0.44)
(0.41)
(0.49)
(0.49)
(0.43)
(0.49)
(0.45)
(0.50)

(0.40)
(0.46)
(0.47)
(0.45)
(0.38)

SVORI
Mean (SD)

Exhibit A-4. Proportion of respondents who reported receiving specific services, by group

0.30
0.20
0.15
0.10

0.72
0.54
0.56
0.45
0.59
0.20
0.35
0.24
0.12
0.24
0.35

0.59
0.49
0.02
0.14
0.08
0.15
0.44
0.10
0.18
0.19
0.15
0.31

0.47
0.33
0.26
0.19
0.27

(0.46)
(0.40)
(0.36)
(0.30)

(0.45)
(0.50)
(0.50)
(0.50)
(0.49)
(0.40)
(0.48)
(0.43)
(0.32)
(0.43)
(0.48)

(0.49)
(0.50)
(0.16)
(0.34)
(0.28)
(0.36)
(0.50)
(0.30)
(0.39)
(0.39)
(0.36)
(0.46)

(0.50)
(0.47)
(0.44)
(0.39)
(0.44)

Non−SVORI
Mean (SD)

(continued)

2.50
4.83
3.18
3.41

−0.70
0.46
−0.43
3.48
−0.25
2.99
2.60
2.47
0.73
4.42
3.50

7.22
7.38
5.18
2.05
4.39
1.65
2.73
6.31
1.54
4.02
2.90
2.36

7.14
7.64
8.81
12.05
12.62

t−statistic

Prisoner Reentry Experiences of Adult Females

Service
Health Services (continued)
Any substance use treatment (continued)
Residential substance use counseling
Methadone
Detox
Information on accessing substance use
treatment
Anger management program
Groups for victims of abuse
Employment/Education/Skills Services
Any employment services
Employment readiness program
Job training program
Talk with potential employer
Advice about job interviewing
Advice about answering questions about
criminal history
Advice about how to behave on the job
Contacts in community to find a job
Resume development assistance
More education
Money management skills
Life skills
Work on personal relationships
Change in 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
0.13 (0.33)
0.01 (0.07)
0.01 (0.12)
0.29 (0.46)
0.27 (0.44)
0.13 (0.33)
0.38
0.23
0.15
0.10
0.32
0.32
0.30
0.27
0.24
0.49
0.23
0.42
0.30
0.45
0.06 (0.24)
0.17 (0.38)
0.02
0.14
0.10
0.37
0.05

356
356
357
356
357
357
357
355
357
357
357
357
357
357
357
357
357
357
357
357
357
357
270
36
270
270
270

(0.15)
(0.35)
(0.30)
(0.48)
(0.21)

(0.47)
(0.46)
(0.45)
(0.43)
(0.5)
(0.42)
(0.49)
(0.46)
(0.50)

(0.48)
(0.42)
(0.35)
(0.29)
(0.47)

Full Sample
Mean (SD)

N

(0.50)
(0.50)
(0.50)
(0.48)
(0.49)
(0.49)
(0.48)
(0.50)
(0.48)

(0.50)
(0.48)
(0.42)
(0.35)
(0.50)

0.03
0.25
0.12
0.50
0.08

(0.18)
(0.44)
(0.32)
(0.50)
(0.28)

0.07 (0.26)
0.24 (0.43)

0.46
0.44
0.42
0.35
0.60
0.40
0.64
0.45
0.66

0.52
0.35
0.23
0.14
0.47

0.47 (0.50)
0.37 (0.49)
0.18 (0.39)

0.20 (0.40)
0.01 (0.11)
0.03 (0.18)

SVORI
Mean (SD)

Exhibit A-4. Proportion of respondents who reported receiving specific services, by group (continued)

(0.41)
(0.40)
(0.36)
(0.36)
(0.49)
(0.30)
(0.43)
(0.39)
(0.46)

(0.44)
(0.34)
(0.28)
(0.24)
(0.41)

0.01
0.00
0.09
0.26
0.02

(0.12)
(0.00)
(0.28)
(0.44)
(0.14)

0.05 (0.23)
0.12 (0.32)

0.22
0.20
0.16
0.16
0.40
0.10
0.25
0.19
0.29

0.26
0.13
0.08
0.06
0.22

0.17 (0.37)
0.19 (0.39)
0.08 (0.28)

0.07 (0.25)
0.00 (0.00)
0.00 (0.00)

Non−SVORI
Mean (SD)

1.04
2.52
0.77
4.35
2.26

0.70
3.00

4.87
5.07
5.64
4.22
3.79
6.56
8.03
5.43
7.36

5.15
4.73
3.71
2.58
5.13

6.22
3.90
2.70

3.60
1.42
2.27

t−statistic

Appendix A — Data Tables

A-13