Svori Reentry Evaluation Methodology Dec 2009

• Topics: Prisoner Privileges, Release and Reentry

Download original document:
Brief thumbnail

This text is machine-read, and may contain errors. Check the original document to verify accuracy.

The Multi-site Evaluation of
SVORI: Methodology and
Analytic Approach
THE MULTI-SITE EVALUATION OF THE SERIOUS AND VIOLENT OFFENDER REENTRY INITIATIVE

December 2009

Pamela K. Lattimore
RTI International
3040 East Cornwallis Road
Research Triangle Park, NC 27709
Phone: (919) 485-7759
Fax: (919) 485-2617
Lattimore@rti.org

Danielle M. Steffey
RTI International
3040 East Cornwallis Road
Research Triangle Park, NC 27709
Phone: (919) 485-7759
Fax: (919) 485-2617
Steffey@rti.org

This project was supported by Grant No. 2004-RE-CX-002 awarded by the National Institute of Justice, Office of Justice Programs, U.S.
Department of Justice. Points of view in this document are those of the authors and do not necessarily represent the official position or policies
of the U.S. Department of Justice.

Acknowledgments
The Multi-site Evaluation of the Serious and Violent Offender
Reentry Initiative (SVORI) was supported by grant number
2004-RE-CX-002 from the National Institute of Justice (U.S.
Department of Justice) and was conducted by RTI International
and the Urban Institute. Points of view are those of the authors
and do not necessarily represent those of the U.S. Department
of Justice.
Principal Investigators
Pamela K. Lattimore, RTI International
Christy A. Visher, University of Delaware and Urban Institute
Report Authors
Pamela K. Lattimore, RTI International
Danielle M. Steffey, RTI International
Staff Contributors
Kelle Barrick, RTI International
Susan Brumbaugh, RTI International
Alex Cowell, RTI International
Debbie Dawes, RTI International
Chris Lindquist, RTI International
Mark Pope, RTI International
John Roman, Urban Institute
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.
RTI and the Urban Institute thank the SVORI program
directors, other program and research staff from the SVORI

iii

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/.

Abstract
Statement of Purpose
The Serious and Violent Offender Reentry Initiative (SVORI)
funded 69 agencies in 2003 to develop programs to improve
criminal justice, employment, education, health, and housing
outcomes for released prisoners. These programs were to
conduct assessments and provide participants with programs
and services during and after incarceration. The Multi-site
Evaluation of SVORI was funded by the National Institute of
Justice to examine the extent to which SVORI program
participation improved access to appropriate, comprehensive,
integrated services and resulted in better outcomes.
Research Subjects
The study included 1,697 adult males (863 SVORI participants;
834 comparison men), 357 adult females (119 SVORI; 134
non-SVORI), and 337 juvenile males (108 SVORI; 131 nonSVORI). The study participants had extensive criminal and
substance use histories, low levels of education and
employment skills, and high levels of need across a range of
services (e.g., education, driver’s license, substance abuse
treatment, and job training).
Study Methods
The impact evaluation included interviews 30 days pre-release
and 3, 9, and 15 months post-release. Data from state
agencies and the National Criminal Information Center
documented post-release recidivism. Propensity score
techniques were used to improve the comparability between
the SVORI and non-SVORI groups. Weighted analyses
examined the treatment effects of SVORI program
participation.

Major Findings
This report documents the procedures used to identify and
recruit evaluation sites and subjects for the Multi-site
Evaluation of SVORI. The investigation of potential sources of
bias, and adjustments to the data using propensity score
models, are also discussed. This report is a companion volume
to three reports that describe outcomes for the adult male,
adult female, and juvenile male evaluation participants; a
report that presents an economic analysis for five programs;
and a report that provides a summary and synthesis of the
Multi-site Evaluation findings.
Conclusions
Sixteen SVORI programs in 14 states were included in the
Multi-site Evaluation of SVORI. These programs included 12
adult and 4 juvenile programs that represented a set of
programs that were diverse in approach and geographically
distributed. The programs were representative of all SVORI
programs along many dimensions, although they were
purposively selected. The impact sites did vary from the nonimpact sites with respect to planning to have larger enrollments
and being further along in terms of full implementation—two of
the criteria that were used during the site selection process.
A total of 4,354 cases were fielded. Of these, 2,391 studyeligible men, women, and boys completed the Wave 1 (prerelease) interview. The remaining cases included 718 cases
released before interviews could be scheduled, 635 cases that
were ineligible for the evaluation, 370 refusals, 192 cases that
were not released before the end of the data collection period
for the Wave 2, 3-month post-release follow-up interview
(declared ineligible for the evaluation), and 48 other
noninterviews. Follow-up response rates were 61%, 64%, and
68% for the 3-, 9-, and 15-month post-release interviews,
respectively.
Propensity score techniques were used to improve the
comparability between the SVORI and non-SVORI groups.
Weighted analyses were used to examine the treatment effects
of SVORI program participation with respect to outcomes in
housing, employment, family/peer/community involvement,
substance use, physical and mental health, and criminal
behavior and recidivism.

Contents
Section
Abstract
Executive Summary

Page
v
ES-1

Introduction

Site Selection and Enlistment

Site Selection .............................................................. 9
Identification of Comparison Population Pools and
Evaluation Participants

Random-Assignment Sites ...........................................26
Quasi-Experimental Site-Level Designs ..........................26
Evaluation Eligibility ....................................................37
Data Collection Procedures

Implementation: SVORI Grantee and Program
Director Surveys ....................................................41
Impact: Interviews with SVORI Program Participants
and Comparison Subjects........................................44
Impact: Administrative Data.........................................47
Case Flow and Threats to Validity

Case Flow ..................................................................57
Selection Bias.............................................................68
Nonresponse ..............................................................71
Attrition.....................................................................72
Propensity Score Models

Propensity Score Model Development for Adult Male
Subjects ...............................................................81
Model Development for Adult Female Subjects ................88
Model Development for Juvenile Male Subjects ................92
Conclusions and Next Steps

References

vii

viii

Appendix A. Implementation Evaluation Protocols
and Data Collection Instruments

A-1

Appendix B. Impact Site Program Descriptions

B-1

Appendix C. Comparisons of Impact and Non–
impact Sites

C-1

Appendix D. MOU Template

D-1

Appendix E. Consent Forms and Interview
Materials

E-1

Appendix F. Administrative Data Protocols and
Data Transfer Protocol

F-1

Exhibits
Number

Page

ES-1. SVORI program logic model and evaluation
framework ......................................................... ES-3
ES-2. Completed interviews by wave, all waves, and any
waves by demographic group ............................... ES-8
1.

SVORI program logic model and evaluation
framework .............................................................. 3

Impact evaluation site selection criteria ......................12

Planned SVORI program capacity ..............................12

Geographic targeting for SVORI participation by
SVORI programs .....................................................13

Selection of impact evaluation programs: reasons
for program exclusion ..............................................16

Original programs selected for the impact
evaluation ..............................................................18

Configuration of selected SVORI Multi-site
Evaluation impact sites ............................................18

Program sizes among impact and non–impact
sites (as reported by program directors).....................20

Implementation status among adult impact and
non–impact sites.....................................................21

10.

Random assignment to SVORI program and
identification of potential evaluation participants .........27

11.

Quasi-experimental identification of potential
evaluation participants .............................................28

12.

Options for the identification of comparison
population pools by program geographic
composition ...........................................................31

13.

Evaluation selection criteria for SVORI and nonSVORI comparison groups by site ..............................33

14.

Identification of evaluation respondents .....................38

15.

Data sources for the Multi-site Evaluation of
SVORI ...................................................................42

16.

Structure of parsed arrest records file ........................53

17.

Transformed arrest records data ...............................53

18.

Outcomes of Wave 1 interview attempts with
eligible respondents ................................................59

19.

Case disposition of ineligible cases (Wave 1 data
collection) ..............................................................61

20.

Wave 2 (3-month post-release) interview case
flow ......................................................................62

21.

Wave 3 (9-month post-release) interview case
flow ......................................................................64

22.

Wave 4 (15-month post-release) interview case
flow ......................................................................65

23.

Completed interviews by wave, all waves, and any
waves by demographic group ...................................65

24.

Completed interviews by wave, by demographic
group and site ........................................................67

25.

t-statistics comparing means of SVORI and nonSVORI groups by demographic group .........................70

26.

Nonresponse pattern for three waves of follow-up
interviews, adult males ............................................73

27.

Results of nonordered multinomial logistic
regression examining nonresponse, adult males ..........74

28.

Variables included in the SVORI propensity score
models and numbers of observations with missing
values ...................................................................78

29.

Parameter estimates for the SVORI propensity
score models for the adult male sample generated
using PROC MI........................................................82

30.

Final propensity model for adult male sample ..............83

31.

ˆ distributions for all adult
Characteristics of p

males, adult male SVORI participants, and nonSVORI comparisons .................................................84

32.

Boxplot of p-hat distributions for SVORI (SVORI =
1) and non- SVORI (SVORI = 0) adult males ..............84

33.

t-statistics and standardized differences from the
comparison of mean values (unweighted) for
SVORI to Non-SVORI for selected variables from
the Wave 1 interview ...............................................86

34.

Summary of balance checks using two-sample tstatistics and percentage of standardized
differences .............................................................87

35.

Balance checks for Wave 1 data based on
propensity score weighted regression of the

SVORI indicator on each of the variables (adult
males)...................................................................88
36.

Balance checks for Waves 2, 3, and 4 data based
on propensity score weighted regression of the
variable on a SVORI indicator (adult males) ................89

37.

Final propensity model for adult female sample ...........90

38.

ˆ distributions for all adult
Characteristics of p
females, adult female SVORI participants, and
non-SVORI comparisons ..........................................91

39.

Boxplot of p-hat distributions for SVORI (SVORI =
1) and non- SVORI (SVORI = 0) adult females ............91

40.

Balance checks for Waves 1 through 4 adult
female data, based on PATE-weighted regressions .......92

41.

Final propensity model for juvenile male sample ..........93

42.

ˆ distributions for all juvenile
Characteristics of p
males, juvenile male SVORI participants, and
juvenile male non-SVORI comparisons .......................94

43.

Boxplot of p-hat distributions for SVORI (SVORI =
1) and non-SVORI (SVORI = 0) juvenile males............94

44.

Balance checks for Waves 1 through 4 juvenile
male data, based on PATE-weighted regressions..........95

C-1. Program director turnover among impact and
non–impact sites................................................... C-1
C-2a. Program characteristics of adult impact and non–
impact sites ......................................................... C-2
C-2b. Program characteristics of juvenile impact and
non–impact sites................................................... C-2
C-3a. Outcome foci among adult impact and non–impact
sites.................................................................... C-3
C-3b. Outcome foci among juvenile impact and non–
impact sites ......................................................... C-3
C-4a. Mean proportion of SVORI offenders receiving prerelease services in adult sites, by site type (as
reported by program directors) ............................... C-4
C-4b. Mean proportion of SVORI offenders receiving
post-release services in adult sites, by site type
(as reported by program directors) .......................... C-5
C-4c. Mean proportion of SVORI offenders receiving prerelease services in juvenile sites, by site type (as
reported by program directors) ............................... C-6
C-4d. Mean proportion of SVORI offenders receiving
post-release services in juvenile sites, by site type
(as reported by program directors) .......................... C-7

C-5a. Involvement and contributions of agencies and
community-based organizations (CBOs) to SVORI
programs in adult sites, by site type (as reported
by program directors) ............................................ C-8
C-5b. Involvement and contributions of agencies and
community-based organizations (CBOs) to SVORI
programs in juvenile sites, by site type (as
reported by program directors) ............................... C-9
C-6a. Support and resistance by individual stakeholders
to SVORI programs in adult sites, by site type (as
reported by program directors) ............................. C-10
C-6b. Support and resistance by individual stakeholders
to SVORI programs in juvenile sites, by site type
(as reported by program directors) ........................ C-11
C-7a. Geographic targeting among adult impact and
non–impact sites................................................. C-12
C-7b. Geographic targeting among juvenile impact and
non–impact sites................................................. C-12
C-8a. Enhancements to pre-release services in adult
sites, by site type (as reported by program
directors) ........................................................... C-13
C-8b. Enhancements to post-release services in adult
sites, by site type (as reported by program
directors) ........................................................... C-14
C-8c. Enhancements to pre-release services in juvenile
sites, by site type (as reported by program
directors) ........................................................... C-15
C-8d. Enhancements to post-release services in juvenile
sites, by site type (as reported by program
directors) ........................................................... C-16

xii

Executive Summary
This volume describes the methods and analytic approaches
that were employed in conducting the Multi-site Evaluation of
the Serious and Violent Offender Reentry Initiative (SVORI).
SVORI was a collaborative federal effort of the U.S. Department
of Justice (DOJ), Department of Labor (DOL), Department of
Education (DoEd), Department of Housing and Urban
Development (HUD), and Department of Health and Human
Services (HHS). The initiative responded to emerging research
findings that suggested that providing incarcerated individuals
with comprehensive, coordinated services based on needs and
risk assessments could result in improved post-release
outcomes (e.g., see Lattimore, 2007; National Research
Council, 2007; Re-entry Policy Council, 2005; Visher, 2007).
In 2003, DOJ, DOL, DOEd, HUD, and HHS provided more than
$100 million in grant funds to states to develop, enhance, or
expand programs to facilitate the reentry of adult and juvenile
offenders returning to communities from prisons or juvenile
detention facilities. SVORI funded agencies to develop
programs to improve criminal justice, employment, education,
health, and housing outcomes for released prisoners. Sixty-nine
agencies received federal funds ($500 thousand to $2 million
over 3 years) to develop 89 programs. Grantees were to use
their SVORI funding to create a three-phase continuum of
services for returning serious or violent prisoners—services 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. In addition to the funding, SVORI encouraged
agencies to coordinate with correctional and community
partners and services.

ES-1

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

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

In spring 2003, the National Institute of Justice awarded RTI
International, a nonprofit research organization, and the Urban
Institute, a nonpartisan economic and social policy research
organization, a grant to evaluate programs funded by SVORI.
The 6-year evaluation involved an implementation evaluation of
all 89 SVORI programs, an intensive impact evaluation of 12
adult and 4 juvenile programs, and an economic analysis on a
subset of the impact sites (see Lattimore et al., 2005). The goal
of the Multi-site Evaluation of SVORI 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 local nature of the SVORI programs and the expectation
that programs would tailor services to meet individual needs
meant that the intervention to be evaluated was not a program
in the typical conceptualization of the term (e.g., a residential
drug program or a cognitive-behavioral program). Instead,
SVORI was a funding stream that agencies used to expand and
enhance existing programs or to develop and implement new
programs. Further, individuals not in SVORI programs also
generally received some services. Thus, although the
components of the individual programs were identified and the
extent of service receipt was measured, the Multi-site
Evaluation of SVORI was not designed to examine the impact of
specific services or combinations of services. The evaluation
was designed to determine whether individuals who
participated in enhanced reentry programming, as measured by
their enrollment in SVORI programs, had improved post-release
outcomes.

This report presents the methods and analytic approach for the
evaluation. Results from the impact and economic evaluations
are presented in the separate reports as noted below.

ES-2

Executive Summary

RESEARCH DESIGN
The evaluation framework is shown in Exhibit ES-1. This
framework shows the SVORI logic model and the evaluation
components. The SVORI program model identifies SVORI
funding, technical assistance (TA), and requirements as inputs
that, in combination with local resources in the sites
(throughputs), yield a set of services and programming
(outputs) that are expected to improve the outcomes for SVORI
participants, as well as to improve the state and local systems
that provide these services and programs. Community and
individual participant characteristics influence these
throughputs, outputs, and outcomes.
Exhibit ES-1. SVORI program logic model and evaluation framework

Community Context

Offender Context
Population Characteristics
Criminal History
Mental & Physical Health
Substance Abuse
Education/Training/Work Experience
Family Ties

Population Characteristics
Unemployment Rates
Service Availability
Residential Stability
Post-release Supervision Structure

Inputs:
The SVORI
Federal Funding &
Other Resources
Technical Assistance
Federal Grant
Requirements

Throughputs
Local Partnership
Formation &
Functioning
State & Local
Resources

Outputs:
Implementation
In-Prison
Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services

Community
Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services

Post-Supervision

Outcomes
Offender
Community Involvement
Employment
Family Contact/Stability
Health/Mental Health
Housing
Recidivism
Substance Use
Supervision Compliance

Systems
Rearrest Rates
Reincarceration Rates
Systems Change

Community Reintegration
Activities

Evaluation
Components

Implementation Assessment

Impact
Evaluation
Cost-Benefit Analysis

ES-3

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

SVORI was an outcomeor goal-oriented initiative
that specified outcomes,
or goals, that should be
achieved by programs
that were developed
locally.

The SVORI program model shows that SVORI was an outcomeor goal-oriented initiative that specified outcomes, or goals,
that should be achieved by programs that were developed
locally. The initiative imposed few restrictions on the grantees.
Criteria specified by the federal partners for the local programs
were the following:

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

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

Program participants were to be serious or violent
offenders.

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

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

Needs and risk assessments should guide the provision
of services and programs to participants.

Operating within these broad guidelines, each program was
locally designed along a variety of dimensions, including the
types of services offered, the focus on pre-release and postrelease components, and the types of individuals to be served.
Programs varied in terms of what was being provided, when,
and to whom. Also, because services were to be delivered to
individuals based on their specific needs and risk factors,
individuals participating in a SVORI program could receive
different types and amounts of services depending upon their
particular needs. Thus, one challenge for the evaluation was to
attempt to characterize SVORI.
As mentioned above, SVORI was not a specific program or set
of services but rather a funding stream that agencies used to
expand and enhance existing programs or to develop and
implement new programs. SVORI program participants were
expected to receive services that directly responded to
individual deficits identified through needs and risk
assessments, while non-SVORI comparison subjects received
treatment as usual.

ES-4

Executive Summary

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

The reentry services provided were mostly services intended to
improve intermediate outcomes that have been correlated with
recidivism—for example, employment services to improve
employment, substance abuse treatment to reduce use, and
cognitive programs to address criminal thinking. The underlying
logic model suggests that improvements in these outcomes will
lead to reductions in criminal behavior—for example, having a
job reduces recidivism by 10%. Thus, the SVORI program
participants (and, to a lesser extent, the non-SVORI
respondents) received a variety of different services, each of
which could affect one or more intermediate outcomes that, in
turn, could affect recidivism. Little theoretical or empirical
guidance exists for the correct specification of such a complex
recidivism model, and, thus, the approach to the outcome
analyses was to test first-order effects of SVORI program
participation on each of the identified outcomes including
recidivism. In particular, the Multi-site SVORI Evaluation was
intended to answer the following research questions:

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

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

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

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

The Multi-site Evaluation of SVORI included an implementation
assessment (to document the programming delivered across
the SVORI programs) and an impact evaluation (to determine
the effectiveness of programming). Sixteen programs were
included in the impact evaluation, comprising 12 adult
programs and 4 juvenile programs located in 14 states (adult
only unless specified): Colorado (juveniles only), Florida
(juveniles only), Indiana, Iowa, Kansas (adults and juveniles),
Maine, Maryland, Missouri, Nevada, Ohio, Oklahoma,
Pennsylvania, South Carolina (adults and juveniles), and
Washington. The impact evaluation included pre-release
interviews (conducted approximately 30 days before release
from prison) and a series of follow-up interviews (conducted at

ES-5

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

3, 9, and 15 months post-release). Nearly 2,400 prisoners
returning to society—some of whom received SVORI
programming and some of whom received “treatment as usual”
in their respective states—were included in the impact
evaluation. An economic analysis was also conducted in five of
the impact sites to assess the extent to which program benefits
exceeded costs; findings from that study are reported
separately (Cowell, Roman, & Lattimore, 2009).
A site-specific research design was developed for each impact
site. In two sites (Iowa and Ohio), the programs randomly
assigned individuals to their SVORI programs. In the remaining
sites, comparison groups were developed by identifying the
criteria that local site staff used to identify individuals eligible
for enrollment in their SVORI program (including such factors
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.
Data collection consisted of four waves of in-person, computerassisted interviews, oral swab drug tests conducted in
conjunction with two of the follow-up interviews, and
administrative data obtained from state correctional agencies
and the National Crime Information Center (NCIC) at the
Federal Bureau of Investigation (FBI). 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 pre-prison
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. Interviews were conducted in the
community, and in jails or prisons for those who were
reincarcerated. 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,

ES-6

Executive Summary

supervision and criminal history, service needs, and service
receipt. The interview instruments were developed through an
extensive process involving substantive domain experts and the
use of existing, validated measures and scales. Oral swab drug
tests were conducted during the 3- and 15-month interviews
for respondents who were interviewed in a community setting.
The interview and drug test data were supplemented with
arrest data obtained from the NCIC and with administrative
records obtained from state correctional and juvenile justice
agencies. These data provided information on criminal history
and recidivism. In some instances, the administrative records
were supplemented with data obtained from online criminal
history databases.
A total of 4,354 cases were fielded for inclusion in the Multi-site
Evaluation of SVORI. Wave 1 (30 days pre-release) interviews
were obtained for 2,583 cases (59%) 1. Of these, 192 cases
were dropped because the respondents were not released
during the period when the first post-release interview was
being conducted. Thus, 2,391 individuals composed the final
sample—1,697 adult men (863 SVORI and 834 non-SVORI),
357 adult females (153 SVORI and 204 non-SVORI), and 337
juvenile males (152 SVORI and 185 non-SVORI). The
remaining cases included 718 cases released before interviews
could be scheduled, 635 cases that were ineligible for the
evaluation, 370 refusals, and 48 other noninterviews. Among
eligible subjects approached for interviews, refusal rates were
reasonably low—12% for adult men, 7% for adult women, and
8% for juvenile males.
Most of the noninterviews (718) were due to individuals’ being
released before their Wave 1 (pre-release) interview could be
scheduled and completed. Release before an interview could be
completed was problematic primarily during the early stages of
data collection and was addressed by identifying potential
respondents earlier relative to their expected release date. No

The 4,354 cases do not include cases that were fielded during the
initial months of the evaluation for populations that were dropped
from the study, including juvenile girls, northern Nevada site
respondents, Maine juvenile subjects, and Virginia adults. All
populations except Virginia adults were excluded because of
insufficient case flow; the Virginia site was dropped because of
logistical difficulties in identifying and interviewing comparison
subjects.

ES-7

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

information was uncovered that suggested that the “early
releases” affected individuals in the two study groups
differently. Furthermore, although the release was “early,” the
difference between actual and expected release dates was only
a few days.
All eligible cases were fielded for each follow-up wave with few
exceptions. Overall, response rates for follow-up interviews
increased over time for all groups. Furthermore, one or more
interviews were obtained with at least 80% of all subjects.
Exhibit ES-2 shows, for each demographic group, the
percentages of subjects who participated in (1) Wave 2
interviews, (2) Wave 3 interviews, (3) Wave 4 interviews,
(4) all 3 follow-up interviews, and (5) at least one follow-up
interview.
Exhibit ES-2. Completed interviews by wave, all waves, and any waves by demographic
group

Completed Interviews (% of Fielded Interviews)
Interview(s)

Adult Males

Adult Females

Juvenile Males

Wave 2

58.0%

68.4%

70.0%

Wave 3

61.0%

70.9%

Wave 4

65.6%

77.3%

73.6%

All 3 follow-ups

42.3%

54.9%

54.3%

Any follow-up

79.3%

87.1%

87.2%

Note: Wave 2 = 3 months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.

Although the response rates were reasonable, the possibility
remains that attriters differed from those who completed the
follow-up interviews. As preliminary evidence that the attrition
was random or affected the SVORI and non-SVORI groups
similarly, comparisons of the SVORI and non-SVORI groups
found them to be similar at each wave on a range of
characteristics. Results from models that examined differences
between groups with respect to response also suggested that
SVORI program participation was not related to whether an
individual responded.
Propensity score techniques were used to improve the
comparability between the SVORI and non-SVORI groups.
Weighted analyses were used to examine the treatment effects
of SVORI program participation with respect to outcomes in
housing, employment, family/peer/community involvement,

ES-8

Executive Summary

substance use, physical and mental health, and criminal
behavior and recidivism.

KEY FINDINGS
Evaluation findings are presented in the following reports:

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

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

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

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

ES-9

Introduction
This volume describes the methods and analytic approaches
that were employed in conducting the Multi-site Evaluation of
the Serious and Violent Offender Reentry Initiative (SVORI).
SVORI was a collaborative federal effort of the U.S.
Departments of Justice, Labor, Education, Housing and Urban
Development, and Health and Human Services. The initiative
responded to emerging research findings that suggested that
providing individuals with comprehensive, coordinated services
based on needs and risk assessments could result in improved
post-release outcomes (e.g., see Lattimore, 2007; National
Research Council, 2007; Re-entry Policy Council, 2005; Visher,
2007).
In 2003, SVORI provided more than $100 million in grant
funding to state agencies to develop programs to improve
outcomes for adults and juveniles offenders released from
prisons and juvenile detention facilities. Sixty-nine agencies
(departments of correction and juvenile justice) received $500
thousand to $2 million one-time awards to develop reentry
programs over a grant period not to exceed 3 years. These
agencies used these grant funds to develop 89 SVORI programs
that provided services to participants over a three-phase
continuum that began during incarceration, continued postrelease during supervision, and extended into continuing
community integration post-supervision. These programs were
intended to improve criminal justice, employment, education,
health (including substance abuse and mental health), and
housing outcomes (see Lattimore, Visher, Brumbaugh,
Lindquist & Winterfield, 2005; Lattimore, Visher, & Steffey,
2008; Winterfield, Lattimore, Steffey, Brumbaugh, & Lindquist,
2006).
In 2003, the National Institute of Justice (NIJ) funded RTI
International and the Urban Institute to plan and conduct a
1

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Multi-site Evaluation of SVORI. The Multi-site Evaluation of
SVORI began in May 2004, after the completion of a 1-year
evaluation design and planning period. 2 This volume describes
site selection, identification of comparison population pools,
data collection procedures and response rates, nonresponse,
and development of propensity score models. 3 Results of the
evaluation are provided separately in Lattimore, Steffey, and
Visher ; Lindquist, Barrick, Lattimore, and Visher ; and
Hawkins, Dawes, Lattimore, and Visher . The remainder of this
section provides a brief overview of SVORI and the evaluation
design.
The evaluation framework is shown in Exhibit 1. This framework
shows the SVORI logic model and the evaluation components.
The SVORI program model identifies SVORI funding, technical
assistance (TA), and requirements as inputs that, in
combination with local resources in the sites (throughputs),
yield a set of services and programming (outputs) that are
expected to improve the outcomes for SVORI participants, as
well as to improve the state and local systems that provide
these services and programs. Community and individual
participant characteristics influence these throughputs, outputs,
and outcomes.

SVORI was an outcomeor goal-oriented initiative
that specified
outcomes/goals that
should be achieved by
programs that were
developed locally.

The SVORI program model shows that SVORI was an outcomeor goal-oriented initiative that specified outcomes/goals that
should be achieved by programs that were developed locally.
The initiative imposed few restrictions on the grantees. Criteria
specified by the federal partners for the local programs were
the following:

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

Planning and design work, including documentation of program
characteristics, were conducted between May 2003 and December
2004 under NIJ award 2003-RE-CX-K101. Continued documentation
of SVORI program progress, the impact evaluation, economic
analysis, and dissemination activities began in June 2004 under NIJ
award 2004-RE-CX-002.
All research activities in support of this evaluation were conducted in
accordance with the approval and oversight of an RTI Institutional
Review Board (IRB). RTI operates three IRBs under Federalwide
Assurance (FWA) granted by the Office of Human Research
Protections (FWA #3331, effective until March 5, 2012).

Introduction

Exhibit 1. SVORI program logic model and evaluation framework

Community Context

Offender Context
Population Characteristics
Criminal History
Mental & Physical Health
Substance Abuse
Education/Training/Work Experience
Family Ties

Population Characteristics
Unemployment Rates
Service Availability
Residential Stability
Post-release Supervision Structure

Inputs:
The SVORI
Federal Funding &
Other Resources
Technical Assistance
Federal Grant
Requirements

Throughputs
Local Partnership
Formation &
Functioning
State & Local
Resources

Outputs:
Implementation
In-Prison
Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services

Community
Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services

Post-Supervision

Outcomes
Offender
Community Involvement
Employment
Family Contact/Stability
Health/Mental Health
Housing
Recidivism
Substance Use
Supervision Compliance

Systems
Rearrest Rates
Reincarceration Rates
Systems Change

Community Reintegration
Activities

Implementation Assessment

Evaluation
Components

Impact
Evaluation
Cost-Benefit Analysis

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

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

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

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

Needs and risk assessments should guide the provision
of services and programs to participants.

Operating within these broad guidelines, each program was
locally designed along a variety of dimensions, including the
types of services offered, the focus on pre-release and post-

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

release components, and the type(s) of individuals to be
served. Programs varied in terms of what was being provided,
when, and to whom. Grantees also identified the locations
where the program would be provided both pre- and postrelease. Thus, a SVORI program could be narrowly focused on a
single institution pre-release serving participants who were
returning to a single community post-release or could be
implemented throughout the correctional (or juvenile justice)
system serving participants who were to be released statewide.
A combination of multiple (but not all) institutions and multiple
(but not all) communities was the modal configuration. Finally,
because services were to be delivered to individuals based on
their specific needs and risk factors, individuals participating in
a SVORI program could receive different types and amounts of
services depending upon individual needs. 4 Thus, one challenge
for the evaluation was to attempt to characterize SVORI.
The local nature of the SVORI programs and the expectation
that programs would tailor services to meet individual needs
meant that the intervention to be evaluated was not a program
in the typical conceptualization of the term (e.g., a residential
drug program or a cognitive behavior program). Instead,
SVORI was a funding stream that agencies used to expand and
enhance existing programs or to develop and implement new
programs. SVORI program participants were expected to
receive services that directly responded to individual deficits
identified through needs and risk assessments. A further
complication for the evaluation was the reality that individuals
not in SVORI programs generally also received some services.
The reentry services provided were mostly services intended to
improve intermediate outcomes that have been correlated with
recidivism—for example, employment services to improve
employment, substance abuse treatment to reduce use, and
cognitive programs to address criminal thinking. The underlying
logic model suggests that improvements in these outcomes will
lead to reductions in criminal behavior—for example, having a
job reduces recidivism by 10%. Thus, the SVORI program

Specific details on the planned characteristics of individual programs
are available in the National Portrait of SVORI (Lattimore et al.,
2004). Also see Lattimore et al. (2005), Winterfield et al. (2006),
and Lindquist and Winterfield (2005) for information on the delivery
of services and programs by the SVORI programs, along with
information on barriers to implementation.

Introduction

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

participants (and, to a lesser extent, the non-SVORI
respondents) received a variety of different services, each of
which could affect one or more intermediate outcomes that, in
turn, could affect recidivism. Little theoretical or empirical
guidance exists for the correct specification of such a complex
recidivism model, so the approach to the outcome analyses was
to test first-order effects of SVORI program participation on
each of the identified outcomes including recidivism. In
particular, although the components of the individual programs
were identified and the extent of service receipt was measured,
the Multi-site Evaluation of SVORI was not designed to examine
the impact of specific services or combinations of services. The
evaluation was designed to determine whether participation in
enhanced reentry programming, as measured by enrollment in
SVORI programs, resulted in increased service receipt and
better post-release outcomes.
The Multi-site SVORI Evaluation was intended to answer the
following research questions:

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

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

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

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

To address these questions, the SVORI evaluation included an
implementation assessment, an impact evaluation, and a costbenefit component.

The implementation assessment component, while not a
comprehensive process evaluation, provided context for
the impact evaluation. The implementation assessment
relied on three surveys of SVORI program directors in
2004, 2005, and 2006 in which data were collected that
characterized the 89 SVORI programs (69 grantees) and
identified the extent to which SVORI programs increased
access to a broad array of services and promoted
systems change. A fourth survey was conducted in 2007
that gathered information on efforts to sustain the

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

The Multi-site
Evaluation of SVORI
was intended to answer
the following research
questions:
● To what extent did
SVORI lead to more
coordinated and
integrated services
among partner
agencies?

SVORI programs. A major product of the implementation
assessment was The National Portrait of SVORI
(Lattimore et al., 2004). Other results were presented in
Lattimore et al. (2005) and Winterfield et al. (2006) and
in a series of Reentry Research in Action brief reports,
including Lindquist and Winterfield (2005) and Lindquist
(2005).

The impact evaluation assessed the effectiveness of
SVORI by comparing key outcomes of those who
participated in SVORI programming with those of
individuals comparable to SVORI participants but who
did not participate in SVORI. This evaluation component
was based on a longitudinal study of adult male, adult
female, and juvenile male returning prisoners in a
subset of sites. (Juvenile females were initially included
in the longitudinal study but were subsequently excluded
due to extremely small numbers.) The impact evaluation
was based on data collected during four waves of inperson interviews and administrative recidivism data.
The interviews were conducted about 1 month before
release from prison (or detention facilities for juvenile
participants), and 3, 9, and 15 months post-release.

The economic analysis examined the return on the
SVORI investment and included both a cost-benefit and
cost-effectiveness analysis. A subset of the impact sites
were selected to study the relative costs and benefits of
SVORI. The approach taken by the Multi-site Evaluation
of SVORI primarily focused on identifying the additional
costs of providing services to SVORI participants and
then assessed the extent to which these additional funds
resulted in improved outcomes. The economic analysis,
including data collections procedures, is described in a
separate report (Cowell et al., 2009).

● To what extent did
SVORI participants
receive more
individualized and
comprehensive
services than
comparable, nonSVORI offenders?
● To what extent did
reentry participants
demonstrate better
recidivism,
employment, health,
and personal
functioning outcomes
than comparable,
non-SVORI
offenders?
● To what extent did
the benefits derived
from SVORI
programming exceed
the costs?

The Multi-site Evaluation of SVORI was designed to address
several challenges, including: (1) the variety of local program
models and treatment populations, (2) a quasi-experimental
design that posed challenges in terms of the identification of
appropriate comparison subjects, (3) a need to be able to
recruit and retain a sample of adequate size to assure sufficient
statistical power given expected “small to moderate” effect
sizes, and (4) the analysis of non-experimental data.
SVORI site teams were integral in collecting the information
necessary for this study. A site team consisted of a site lead
and a site liaison. The site lead was a senior evaluation
researcher who assessed the evaluability of a program and

Introduction

worked with the selected impact program and the state agency
to secure a research agreement. The site liaison collected
information on the programs, maintained contact with the
program directors, and, in the impact sites, coordinated the
acquisition of study subjects and access to prisons and local
jails for the field interviewers.
This methodology report describes the methods and analytic
approach that were employed to conduct the Multi-site
Evaluation of SVORI. The following sections describe the
procedures and criteria that were employed to select the adult
and juvenile impact sites and to identify comparison
populations in these sites. Data collection procedures are
described next, including the procedures for the program
director surveys, the four waves of interviews with SVORI
program participants and comparison subjects, and the
collection and processing of administrative data from state
agencies and the Federal Bureau of Investigation (FBI). The
subsequent section addresses issues related to threats to
validity, including selection, nonresponse, and attrition. The
next section describes the development of the propensity score
models that were used to adjust for differences between SVORI
participants and non-SVORI comparison subjects and the final
section describes the analytic methods applied to examine the
impact of SVORI.

Site Selection and
Enlistment
This section describes the procedures that were followed to
identify sites for the impact evaluation and the activities
associated with enlisting programs to participate in the impact
evaluation. The first objective required the development of an
extensive database of information describing the 69 SVORI
grantees and 89 SVORI programs, conduct of site visits,
analysis, and final selection (in concert with the National
Institute of Justice [NIJ]). The second required working closely
with each department of correction and juvenile justice to
comply with each agency’s requirements for the conduct of
research in their facilities. These efforts are described in this
section.

SITE SELECTION
The initial goal was to identify a set of sites that would provide
a total of about 2,000 SVORI participants and 2,000
comparison subjects over a Wave 1 (pre-release) interview
period of 12 months while providing geographic and
programmatic diversity. The 2,000 subjects in each group were
to include 1,000 men, 500 women, and 500 juveniles. 5
In developing criteria for site selection, the focus was on
identifying factors that would provide the best assurance that a
program would be evaluable. Six criteria were identified to
guide site selection:
1. Program has clearly defined elements and goals.
5

These targets were not met. The final dataset of eligible respondents
consisted of data for 1,697 men, 357 women, and 337 boys.
Additional information is provided in subsequent sections of this
report.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

2. Program is implemented (or is likely to be
implemented).
3. Program target population is accessible and of sufficient
size.
4. Appropriate comparison population is available and
accessible for inclusion in the study.
5. Administrative data are of good quality and available for
the evaluation.
6. Program is amenable to and able to participate in the
evaluation.
The strategy implemented to identify the impact programs was
based on the following successive data collection activities:
1. Review of SVORI grantee proposals and work plans and
follow-up telephone interviews with program directors to
obtain information not gleaned from the review and
clarification and updates on the programs’ status
2. Visits to the sites of a selected subset of programs
3. Review of all information to develop a list of
recommended programs for inclusion in the impact
evaluation that was submitted to NIJ for approval
Additional information on the conduct of each of these steps is
provided below.
Review of Work Plans and Follow-Up Telephone
Interviews

The first step in assessing the programs for inclusion in the
impact evaluation was a document review by SVORI site
liaisons. The site liaisons reviewed the work plans submitted by
the SVORI grantees and extracted descriptive information that
was entered into a database within the project management
information system (SVORIMIS) that was accessible through an
internal project Web site.
Subsequently, semi-structured telephone interviews were
conducted with all grantees to collect additional information and
confirm the number of programs being supported by each
grant. A total of 89 distinct SVORI programs were identified.6
Follow-up telephone calls with program directors were
conducted, as needed, to obtain clarifying information. Copies
6

A grant was determined to be supporting different programs if
(1) the program director identified multiple programs or
(2) programs were operating at different locations headed by
different program directors.

Site Selection and Enlistment

of the data collection protocol and the work plan review form
are included in Appendix A.
This review produced information on the characteristics of the
SVORI program target population, status of implementation,
program components and services, the capacity and willingness
to participate in the evaluation, the availability of sufficient
treatment and appropriate comparison populations, and
additional information on program goals and activities. Other
information focused on agency involvement in SVORI,
management and oversight of the project, and plans to conduct
a local evaluation. Information from these sources was entered
by site liaisons into SVORIMIS.
Once the initial data were gathered, the sites were examined
using the site selection criteria. Exhibit 2 shows the criteria for
site selection. Implementation, target population size,
comparison subject availability, and willingness to participate
were key factors that were considered.
Reports generated from the project database revealed that the
69 grantees were operating a total of 89 programs, including
37 programs targeting juveniles, 45 programs targeting adults
only, and 7 programs targeting both adults and juveniles (sites
that are focused primarily on adult offenders but that include
offenders younger than 18 years of age if they are housed in
adult facilities). For the evaluation, these combination programs
were included with the adult programs. Most programs were
provided to both males and females, although 17 were provided
only to males and one was provided only to females. Sixty-five
of the programs specified that they were identifying a broad
segment of their serious and violent offenders for participation;
the remainder of the programs was focusing on special-needs
populations, including those with substance abuse problems or
co-occurring disorders, or those charged as sex offenders.
(Program descriptions are available in the National Portrait of
SVORI, Lattimore et al., 2004).

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 2. Impact evaluation site selection criteria

As shown in Exhibit 3, 34 of the 87 (39%) programs that
reported planned to serve fewer than 150 participants over the
entire period of their grant (3 years). From a programmatic
standpoint, concentrating resources on a few participants
suggests a higher likelihood of a strong treatment effect.
However, from the standpoint of fielding interview teams, low
case flow has substantial cost implications and, thus, small
programs were excluded from consideration as impact sites.
Exhibit 3. Planned
SVORI program capacity

Adult/Combination
Juvenile
Total

Program Size
Fewer Than 100
101–150
10
5
12
7
22
12

151+
34
19
53

Total
49
38
87

Note: Two programs did not provide estimates.

As understanding of the program configurations developed,
additional dimensions of program characterization emerged.
Specifically, the programs varied considerably in how broadly
based geographically they were during both the pre-release and
the post-release phases. In particular, in some sites, programs
were provided in only one prison/facility; in other sites,

Site Selection and Enlistment

programs were provided in multiple—but not all—institutions;
and, in a few sites, SVORI programs were available in all
facilities (i.e., statewide). Analogous service patterns pertained
to post-release programming. Some sites restricted the
program to individuals who were to be released to a single
community (e.g., a single Zip code in Baltimore, Maryland);
others provided the program to individuals who were to be
released to multiple communities; and a few sites provided the
program regardless of post-release community (i.e., with no
post-release geographic constraint). The configurations for the
SVORI programs along these two dimensions are shown in
Exhibit 4. As can be seen, the most common model was one in
which there was targeting of institutions and/or communities.
Exhibit 4. Geographic targeting for SVORI participation by SVORI programs

Pre-Release
Single prison
Multiple prisons
All prisons
Total

Single Community
5
15
15
35

Post-Release
Multiple Communities
1
25
20
46

Statewide
1
2
4
7

Total
7
42
39
88

Note: One program did not provide information.

The programs were categorized based on pre- and post-release
geographic targeting because the various geographic
configurations had implications both for the nature and
potential effect of the program and for the conduct of the Multisite Evaluation of SVORI in terms of the feasibility and cost of
data collection activities. For the programs, the different preand post-geographic configurations reflected

different resource allocation decisions (i.e., lots of
participants in many facilities implies fewer resources
per participant than if programs are provided to fewer
participants, all else being equal);

different training requirements (training multiple staff at
multiple facilities is a greater challenge than training a
few individuals at one facility); and

different communication strategies (systems integration
between one prison and multiple resources in a single
community poses substantial communication and
coordination issues that are compounded if the program
includes many facilities and communities statewide).

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Each of these factors has implications for program
implementation and impact. For example, programs that were
providing programming to inmates in a single facility who were
to be released to a single community could concentrate their
planning, boundary spanning activities, and coordination among
relatively few individuals. Training and communication among
staff and community resources were directed to those in one
facility and one community rather than to multiple groups.
Under these circumstances, one might expect better
communication and quicker implementation of new services. On
the other hand, reentry programs provide the greatest benefit—
assuming they are effective—if they can be implemented to
serve large proportions of prison populations, which implies
including most prisons and all communities statewide. Finally,
of course, the costs implications for the evaluation of collecting
data statewide versus in a few jurisdictions had to be
accommodated in making site selections.

The primary factors used
in developing the list of
sites to be visited were
anticipated program
enrollment (greater than
100, unless the program
was in the same site as
another program with
enrollment greater than
100) and status of
program implementation.

After the initial review of program type, enrollment, and
geographic targeting, research staff began to narrow down the
69 grantees for site visits. The primary factors used in
developing the list of sites to be visited were anticipated
program enrollment (greater than 100, unless the program was
in the same site as another program with enrollment greater
than 100) and status of program implementation, although the
other site selection criteria shown in Exhibit 2 were considered.

Site Selection and Enlistment

Site Visits

Twenty-nine grantees providing 39 separate programs in 21
states were selected for site visits to collect additional
information from SVORI program directors and staff. 7 These
site visits were conducted by SVORI site teams during the last
quarter of 2003. A copy of the site visit protocol, including
topics covered by the 2-day visits, is included in Appendix A.
The primary purpose of the site visits was to update
information from the work plans, gather information about the
availability and quality of administrative data, confirm program
implementation progress, assess site willingness to participate,
and explore opportunities for identifying comparison subjects (if
the site was not randomly assigning SVORI participation). The
typical 1.5- or 2-day site visit included interviews
(approximately 1 hour) conducted in groups with the following
stakeholders

program director (and program staff they wanted to
include),

local evaluation staff (if the site is doing a local
evaluation),

DOC institutional program staff,

DOC research/MIS staff,

community supervision staff (parole/probation), and

community service providers.

Site visit reports were prepared by the site teams and selected
information was entered into the SVORIMIS in preparation for
7

Programs operated by the following grantees were visited: Colorado
Department of Corrections, Delaware Health and Social Services,
Florida Department of Corrections, Florida Department of Juvenile
Justice, Georgia Criminal Justice Coordinating Council, Iowa
Department of Corrections, Indiana Department of Corrections,
Kansas Department of Corrections, Kansas Juvenile Justice
Authority, Maryland Department of Public Safety and Correctional
Services, Maine Department of Corrections, Michigan Department of
Corrections, Michigan Family Independence Agency, Minnesota
Department of Corrections, Missouri Department of Corrections,
Missouri Department of Social Services, North Carolina Department
of Corrections, North Carolina Department of Juvenile Justice and
Delinquency Prevention, New Mexico Department of Corrections,
Nevada Department of Corrections, Nevada Department of Human
Resources, Ohio Department of Rehabilitation and Corrections,
Oklahoma Department of Corrections, Oklahoma Office of Juvenile
Affairs, Pennsylvania Department of Corrections, Virginia
Department of Corrections, and Washington State Department of
Corrections.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

review and final impact program selection, which is described
next.
Review and Final Impact Program Selection

Exhibit 5 summarizes the reasons that programs were excluded
from the impact evaluation. Most programs were excluded
because of the small number of participants expected to be
released from institutions during the projected Wave 1
interview period (July 2004–October 2005 at the time of site
selection). Specifically, 32 grantees operating 49 programs
were excluded because of expected case flow. In addition, 11
grantees operating 11 programs were excluded because of
program characteristics (e.g., program was highly
decentralized). The four programs operated by the four
grantees in Alaska, Hawaii, and the Virgin Islands were
eliminated from consideration because their locations would
have made them costly to include and also because they were
viewed as not representative of typical states. Three programs
operated by two grantees were eliminated because of concerns
about the quality of their administrative data. Finally, the
program in Texas was excluded because it was targeting
prisoners in administrative segregation, and the program in
Connecticut was excluded because it was focused on prisoners
with serious mental illness.
Exhibit 5. Selection of impact evaluation programs: reasons for program exclusion

Special populations
Admin data
Geography
Program/other
Expected case flow (N)

Total Excluded
Grantees
Programs
2
2
2
3
4
4
11
11
32
49

Remaining After Exclusions
Grantees
Programs
67
84
65
81
61
77
50
66
18
19

The final list of sites proposed for the impact evaluation
incorporated a diversity of program types, geographic regions,
and corrections philosophies. This list was presented to and
discussed with NIJ in December 2003. As noted earlier, the
initial goal was to interview about 4,000 subjects. Based on
flow analyses conducted at the time of final site selection, this
target still appeared feasible, although it had become less
obvious that the targets for adult female and juvenile subjects

Site Selection and Enlistment

could be met because of limited case flow of these populations
in the selected sites. Thus, although 4,000 was still the goal, it
seemed likely that a more realistic goal was 3,000 subjects
(1,500 SVORI participants and 1,500 non-SVORI comparison
subjects). This total was to be distributed as about 2,000 adult
males, 500 adult females, and 500 juveniles. The anticipated
contribution of each program to these totals was included in the
final program-selection calculus.
Exhibit 6 identifies the sites that were initially identified for
inclusion in the impact evaluation. One adult site (Virginia) was
dropped shortly after data collection began because of logistical
problems associated with the identification and interviewing of
subjects. 8 The juvenile program in Maine, juvenile females in all
juvenile impact sites, and adult participants in the northern
Nevada site were included in the original list of sites, but were
dropped from the impact evaluation because of insufficient case
flow.
In the end, 16 programs in 14 states were included in the
impact evaluation. The distribution of these sites over pre- and
post-release geographic targeting of the SVORI programs is
shown in Exhibit 7. Descriptions of the programs are included in
Appendix B.
The impact sites represented a set of programs that were
diverse in approach and geographically distributed. Although
the resulting programs were not randomly selected, the adult
programs are in states that, at year-end 2003, incarcerated
about 20% of all adult state prisoners and supervised about
23% of all adult state parolees in the United States. 9

The Virginia program was a reentry program that was offered to
state prisoners who returned home through the Fairfax County jail,
where the program that included employment and other transition
services was offered. Potential program participants flowed to this
program from all Virginia prisons and would have required
considerable travel costs to interview comparison subjects.
Estimates are based on data from the Bureau of Justice Statistics’
Adults on Parole in the United States (Glaze & Palla, 2005) and
Prisoners under the Jurisdiction of State or Federal Correctional
Authorities (Harrison & Beck, 2005). The twelve states had an
estimated prison population of 259,971 in mid-year 2004 (19.8% of
all state prisoners) and 154,532 individuals on parole at year-end
2004 (22.9% of all individuals under state parole supervision).

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 6. Original programs selected for the impact evaluation
State
Grantee Agency
CO
Colorado Department of Corrections

Program
Colorado Affirms Reentry
Efforts (CARE)
Going Home

Florida Department of Juvenile Justice

IA
IN
KS

Iowa Department of Corrections
Indiana Department of Corrections
Kansas Department of Corrections

Kansas Juvenile Justice Authority

ME
MEa
MD
MO
NV
OH

Maine Department of Corrections
Maine Department of Corrections
Maryland Department of Public Safety and
Correctional Services
Missouri Department of Corrections
Nevada Department of Corrections
Ohio Department of Rehabilitation and Corrections

OK
PA

Oklahoma Department of Corrections
Pennsylvania Department of Corrections

SC
SC
VAa
WA

South Carolina Department of Corrections
South Carolina Department of Juvenile Justice
Virginia Department of Corrections
Washington State Department of Corrections

Focus of Impact
Evaluation
Juveniles

Iowa SVORI
Allen County SVORI
Shawnee County Reentry
Program (SCRP)
Kansas JJA Going Home
Initiative (GHI)
Maine Reentry Network
Maine Reentry Network
Re-Entry Partnership (REP)
Going Home-SVORI
Going Home Prepared
Community-Oriented Reentry
Program
PROTECT Oklahoma County
Erie, PA, Reentry Project
(EPRP)
SC Department of Corrections
SC DJJ Reentry Initiative
Going Home to Stay-VASAVOR
Going Home

Juveniles
(Dade County)
Adults
Adults
Adults
Juveniles
Adults
Juveniles
Adults
Adults
Adults
Adults
Adults
Adults
Adults
Juveniles
Adults
Adults

Subsequently dropped from the evaluation.

Exhibit 7. Configuration of selected SVORI Multi-site Evaluation impact sites

PreRelease
Single
prison
Multiple
prisons

All prisons

Post-Release
Single Community
MD (adult)
FL (juvenile; Dade County only)
IA (adult; participants transferred to 1
of 3 facilities for programming)
KS (adult; male participants
transferred to 1 facility)
PA (adult; participants transferred to 1
male and 1 female facility)
CO (juvenile)
IN (adult)
OK (adult)

Multiple Communities
None

Statewide
None

MO (adult)
NV (adult; male participants
transferred to 1 facility)
OH (adult)
WA (adult; male participants
transferred to 1 of 3 facilities
and female participants
transferred to 1 female facility)
ME (adult)
KS (juvenile)
SC (juvenile)

SC (adult)

Site Selection and Enlistment

The impact sites were representative of all sites along many
dimensions, although they were purposively selected. As
expected, the impact sites did vary from the non-impact sites
with regard to the selection criteria discussed above. In
particular, the impact sites generally planned to have larger
enrollments (Exhibit 8); this was true for both adult and
juvenile sites. Discrepancies between expected and actual
enrollment were similar for adult impact and non-impact sites.
However, discrepancies between expected and actual
enrollments were not as similar in juvenile impact and nonimpact sites: one third of juvenile non-impact sites experienced
enrollments that exceeded expectations, while the four juvenile
impact sites experienced enrollments that either failed to meet
or met expectations. As of March 2006, the adult impact sites
had enrolled an average of 326 program participants in
comparison with an average enrollment of 290 participants by
the non-impact sites. The juvenile impact sites had enrolled an
average of 153 SVORI program participants in comparison with
an average of 204 participants by the juvenile non-impact sites.
Program directors for the impact sites also were more likely
than program directors in non-impact sites to report being
further along in terms of full implementation in both the 2005
and 2006 program director surveys (Exhibit 9). Again, this
discrepancy between impact and non-impact sites was
expected, because likelihood of full program implementation
was one of the selection criteria for inclusion in the impact
evaluation.
Additional details are provided in Appendix C, which provides
additional comparisons of the characteristics of the impact
sites, non-impact sites, and all sites that were derived from the
surveys of SVORI program directors. Overall, these tables
reveal relatively few differences in distributions for adult or
juvenile programs with regard to program director turnover
(Exhibit C-1), basic program characteristics (Exhibits C-2a and
C-2b), targeted outcomes (Exhibits C-3a and C-3b), pre-release
and post-release service provision (Exhibits C-4a, C-4b, C-4c,
and C-4d), agency involvement and contributions (Exhibits
C-5a and C-5b), stakeholder support and resistance (Exhibits
C-6a and C-6b), and pre-release and post-release geographic
targeting (Exhibits C-7a and C-7b). There were differences in

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 8. Program sizes among impact and non–impact sites (as reported by program
directors)

Program size
Adult planneda
Fewer than 100
101–150
More than 151
Adult in 2006b
Fewer than 100
101–150
More than 151
Adult compared with plannedc
Fewer
About the same
More
Juvenile plannedd
Fewer than 100 planned
101–150
More than 151 planned
Juvenile in 2006e
Fewer than 100 enrolled
101–150
More than 151 enrolled
Juvenile compared with plannedf
Fewer than originally projected
About the same as projected
More than originally projected

Impact Sites
%
N

All Sites
%
N

—
25.0
75.0

0
3
9

59.0
17.9
23.1

23
7
9

45.1
19.6
35.3

23
10
18

—
33.3
66.6

0
4
8

51.3
12.8
35.9

20
5
14

39.2
17.7
43.1

20
9
22

50.0
25.0
25.0

6
3
3

50.0
23.7
26.3

19
9
10

50.0
24.0
26.0

25
12
13

25.0
50.0
25.0

1
2
1

67.7
12.9
19.4

21
4
6

62.9
17.1
20.0

22
6
7

—
50.0
50.0

0
2
2

54.8
25.8
19.4

17
8
6

48.6
28.6
22.9

17
10
8

50.0
50.0
0.0

2
2
0

33.3
36.7
30.0

10
11
9

35.3
38.2
26.5

12
13
9

Fifty-one programs reporting; source: 2003 program work plan review.

Fifty-one programs reporting; source: 2006 program director survey.

Fifty programs reporting; source: 2006 program director survey.

Thirty-five programs reporting; source: 2003 program work plan review.

Thirty-five programs reporting; source: 2006 program director survey.

Thirty-four programs reporting; source: 2006 program director survey.

Non–impact
Sites
%
N

Site Selection and Enlistment

Exhibit 9. Implementation status among adult impact and non–impact sites

Impact Sites
%
N

Program and status
Adult, fully operational in 2005a
No
16.7
2
Yes
83.3
10
b
Adult, time to full implementation in 2005
Less than 3 months
—
0
3–5 months
27.3
3
6–8 months
27.3
3
9–11 months
9.1
1
12 months or more
36.4
4
Adult, planned elements fully operational in 2006c
No
16.7
2
Yes
83.3
10
d
Juvenile, fully operational in 2005
No
25.0
1
Yes
75.0
3
Juvenile, time to full implementation in 2005e
Less than 3 months
25.0
1
3–5 months
—
0
6–8 months
25.0
1
9–11 months
25.0
1
12 months or more
25.0
1
Juvenile, planned elements fully operational in 2006f
No
—
0
Yes
100.0
4
a

Fifty-two programs reporting; source: 2005 program director survey.

Forty-one programs reporting; source: 2005 program director survey.

Fifty programs reporting; source: 2006 program director survey.

Thirty-seven programs reporting; source: 2005 program director survey.

Thirty-three programs reporting; 2005 program director survey.

Thirty-four programs reporting; source: 2006 program director survey.

Non–impact
Sites
%
N

All Sites
%
N

37.5
62.5

15
25

32.7
67.3

17
35

20.0
3.3
23.3
13.3
40.0

6
1
7
4
12

14.6
9.8
24.4
12.2
39.0

6
4
10
5
16

7.9
92.1

3
35

10.0
90.0

5
45

15.2
84.8

5
28

16.2
83.8

6
31

27.6
10.3
27.6
13.8
20.7

8
3
8
4
6

27.3
9.1
27.3
15.1
21.2

9
3
9
5
7

3.3
96.7

1
29

2.9
97.1

1
33

expected pre-release and post-release service enhancements,
which may have been associated with anticipated strength of
implementation (Exhibits C-8a, C-8b, C-8c, and C-8d).
Once sites were selected, the next task was to comply with
agency requirements for conducting research in their facilities.
This undertaking is described in the following section.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Site Enlistment

In February 2004, letters were sent to the agency heads and to
the SVORI program directors explaining their selection for
inclusion in the Multi-site Evaluation of SVORI. In addition, an
administrator at NIJ also sent letters to the sites thanking them
for their past and future cooperation with the evaluation
activities. These letters were followed by telephone contacts
between the site teams, the SVORI program directors, and
individuals in the agencies who could provide information on
the research agreement protocols and approval processes that
needed to be navigated to proceed with research in the site.
In preparation for this process, a Memorandum of
Understanding (MOU) was drafted that served as a template for
the development of site-specific MOUs (see Appendix D). This
template was used either in lieu of or in conjunction with any
state-required research agreements. The primary topics
covered by the MOU were

responsibilities of agency staff, which included providing
assistance with identifying, accessing, and tracking
study participants;

responsibilities of RTI field interviewers, including
complying with all institutional rules, regulations, and
requirements and reporting emotional distress of
respondents;

parameters of the interview process for interviews
conducted in prison (or juvenile detention) facilities,
including the requirement for a private space for
administration of the interview; and

participant confidentiality and data security, including
the requirement that participation in the interview was
voluntary and that all individual data would be held
confidential and would not be disclosed to the agency.10

In addition, most sites required the completion of a research
application. These applications ranged in complexity from
relatively short, straightforward forms to extensive
requirements documenting purposes and procedures. In some
cases, a copy of RTI’s Institutional Review Board (IRB) approval

An exception noted in the agreement was to disclose any statement
of intent by the respondent to hurt himself or others. Notification of
this exception to confidentiality was included in the informed
consent that was reviewed and signed by the participants.

Site Selection and Enlistment

was sufficient; in other cases, the agency requested a full copy
of RTI’s IRB application and amendments.
Finally, MOUs were executed with each facility in which
interviews were conducted (including local jails for Waves 2
through 4 interviews). A copy of the template for this MOU is
also included in Appendix D.
Once negotiated research agreements were in place, the next
step was to develop evaluation plans for each site (or each
program, if a site had multiple programs). To accomplish this
task, site leads and site liaisons talked with project directors
and agency management information system personnel. This
process is described in the next section.

Identification of
Comparison
Population Pools
and Evaluation
Participants
The Multi-site Evaluation
of SVORI took an “intent
to treat” approach.

Two pathways to inclusion in SVORI programs were identified:
(1) random assignment to SVORI programming or standard
programming after a decision to participate in SVORI was made
by the offender and (2) program and offender determination of
SVORI program eligibility and participation. For the first
pathway, those randomly assigned to standard programming
constituted the pool of potential comparison subjects. For the
second, evaluation team members worked with local personnel
to identify the site-specific SVORI eligibility criteria and to
establish procedures for selecting a comparison group. In most
cases, the comparison subjects were offenders who would have
been eligible for (i.e., offered) SVORI if they had been in a
facility that offered the SVORI program or if they had planned
to return to a community with a post-release SVORI program.
The Multi-site Evaluation of SVORI took an “intent to treat”
approach with respect to the classification of subjects as SVORI
participants or non-SVORI comparison subjects. Practically, this
meant that an individual was classified as SVORI or non-SVORI
depending upon whether he/she was enrolled in a SVORI
program at any time during the period between when the site
first provided the case information to the evaluation team and
when the case was fielded. Appendix B provides specific
information on the SVORI program eligibility criteria and the

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

criteria that were used to identify the comparison population for
each site.
The following sections summarize the processes for identifying
comparison population pools and the criteria applied to identify
eligible evaluation participants from those pools. The first
section describes the process in the two random-assignment
sites. Subsequently, the quasi-experimental process for the
remaining sites is described. The final section describes the
criteria for the identification of evaluation respondents, which
were consistent across all sites.

RANDOM-ASSIGNMENT SITES
The adult program sites in Iowa and Ohio randomly assigned
eligible cases to participate in the SVORI program or to receive
standard programming. (Note that the assignment probability
sometimes differed from 0.5.) This design is shown in
Exhibit 10. As can be seen, these two sites present the simplest
case, in which random assignment to SVORI or non-SVORI
follows the decision to participate by the offender. In this case,
eligible individuals who were not offered SVORI and those who
refused SVORI were ineligible for inclusion in the evaluation.
The remainder of randomly assigned SVORI and Non-SVORI
composes the potential evaluation respondent pool.

QUASI-EXPERIMENTAL SITE-LEVEL
DESIGNS
In most sites, a quasi-experimental design was developed in
close collaboration with site personnel to identify a pool of
individuals comparable to those who were offered the SVORI
program but who were not asked to participate in SVORI. These
were individuals who met the individual-level program eligibility
criteria (e.g., offense type, expected release date, age), but
who were not included in SVORI. In many cases, these were
individuals who met all local SVORI program eligibility criteria
except for those related to where they were housed (i.e., in a
facility offering SVORI) or where they were returning at release
(i.e., to a community with a post-release SVORI program).

Identification of Comparison Population Pools and Evaluation Participants

Exhibit 10. Random assignment to SVORI program and identification of potential evaluation
participants

Exhibit 11 shows the process for the quasi-experimental
identification of the potential evaluation respondents.
The identification of “Non-SVORI” participants was critical to
the individual design decisions that were made with each site.
Site teams worked closely with the local sites to identify the
criteria that were used to determine SVORI program eligibility
and to develop procedures with local program staff (usually in
conjunction with agency management information system
personnel) for obtaining the names of SVORI participants and
for identifying those who were SVORI-eligible but who were not
offered the SVORI program. The geographic targeting (Which
prisons? Which communities?) of the SVORI programs was used
to frame the discussion with the sites.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 11. Quasi-experimental identification of potential evaluation participants

The evaluation site teams received guidance with respect to the
following factors that (1) were to remain paramount in
negotiating the design with the program and agency personnel
and (2) guided the development of alternative strategies for
comparison pool identification:

Comparison subjects should be “similar” to SVORI
participants to reduce selection bias. Although the
determination of how “similar” SVORI and non-SVORI
offenders were on most eligibility criteria (e.g., LSI
scores, instant offense, county of post-release
residence) was straightforward, many of the programs
were designed to be voluntary, thereby potentially
complicating the identification of a comparison group
that was truly comparable to the treatment group in

Identification of Comparison Population Pools and Evaluation Participants

terms of motivation. 11 This was a particular concern for
programs that claimed they were enrolling all eligible
participants.

Comparison subjects should be returning to the same
(or “similar”) communities to minimize potential
differences in effects that could be attributed to
unobserved/unmeasured environmental factors (e.g.,
differences in employment opportunities, treatment
resources).

Comparison subjects should have been incarcerated in
the same (or “similar” prisons) to minimize potential
differences in effects that could be attributed to
unobserved/unmeasured prison environmental factors
(e.g., availability of programming). This factor was
addressed by attempting to ensure that if “comparison
prisons” were to be used as a source of comparison
subjects, then those comparison prisons should have
similar prison environments (e.g., particularly, custody
level).

Exhibit 7 in the previous chapter classified the selected SVORI
programs by their pre- and post-release geographic targeting.
Options for identifying potential comparison subjects were
developed on the basis of this programmatic categorization. In
summary, once random assignment was eliminated, the best
comparison for any configuration was the same: prisoners from
a SVORI facility who were returning to the SVORI post-release
community and who were very similar to SVORI participants
(but didn’t participate in SVORI for reasons not expected to be
related to future success). This is the Similar Subject-SVORI
Prison-SVORI Community option. If this wasn’t possible (e.g., if
all individuals fitting these criteria either were enrolled in or had
rejected the SVORI program), variations were explored on the
three components: Different Subjects (and/or) Different Prisons
(and/or) Different Communities. Not all options were possible
for all program configurations (e.g., Different Prisons is not an
option if the program is implemented system wide). The least
desirable option was to select subjects from another system
(i.e., state), presumably Similar Subjects/Different
Prisons/Different Communities, but this option was not
11

Most sites with post-release target areas used pre-prison county of
residence to identify potential SVORI participants and evaluation
comparison subjects. This was a complicating factor for the
identification of both SVORI and non-SVORI participants and was
mentioned by at least a few site program directors as a factor that
contributed to low enrollment and excessive drop-outs.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

needed. 12 Exhibit 12 summarizes the options that were
explored for identification of comparison subjects by program
configuration.
The results of the negotiating process for each site are shown in
Exhibit 13, which also describes the SVORI program
participants who were eligible for the evaluation.
One site emerged as potentially problematic in terms of the
evaluation design. Specifically, Maryland was one of the two
adult programs that had a post-release focus (i.e., most of the
program was to be delivered in the community). In Maryland,
individuals were identified as potentially eligible for SVORI
before release and were identified as such for the evaluation if
they attended a meeting describing the Reentry Partnership
Program while they were incarcerated and if they indicated
interest in the post-release program components. Program
participation was not confirmed until after they were released
and voluntarily went to the program that was providing SVORI
services. The Multi-site Evaluation took an intent-to-treat
approach with respect to all programs and did the same with
Maryland: Individuals who were designated as SVORI
participants at any time between when the case information
was provided by the site and when RTI fielded the case for the
pre-release (Wave 1) interview were considered SVORI
participants regardless of whether they attended the program.
As a result, not all individuals classified as “SVORI” for
evaluation purposes actually received SVORI programming or
services, and some individuals who ultimately enrolled in the
SVORI program were treated as comparisons for evaluation
purposes. The Maryland program focused on individuals
returning to a specific neighborhood in Baltimore. The nonSVORI comparison subjects were identified as individuals
otherwise eligible for SVORI except that they were returning to
neighborhoods in Baltimore where SVORI was not provided.
12

As is described in more detail in Lindquist et al. (2009), the
distribution of female subjects between SVORI and Non-SVORI
varied across the 11 sites that included women in their SVORI
programming. Approximately half of the 204 Non-SVORI evaluation
participants were from Indiana compared with only 12 of the SVORI
evaluation participants. In contrast, there were no Non-SVORI
comparisons in two states (Missouri and Pennsylvania). The
decision was made early in the enrollment of subjects to include
these “extra” Non-SVORI comparisons in Indiana in the hopes that
they would be appropriate comparisons for SVORI participants in
other states.

Identification of Comparison Population Pools and Evaluation Participants

Exhibit 12. Options for the identification of comparison population pools by program
geographic composition
PreRelease
Single
prison

Single Community
Individuals at the participating
prison returning to the same
community who are (very)
similar to SVORI participants
but do not participate—
optimally for reasons other than
motivation or other individual
characteristics that would be
expected to be related to future
success

1. Individuals from different
prisons in the same system
who would have been
SVORI-eligible who are
returning to the SVORI
target community
2. Individuals from SVORI
prison who are different in
measurable ways from
SVORI participants who are
returning to the SVORI
target community
3. Individuals from SVORI
prison returning to
communities similar to the
SVORI community
4. Individuals from non-SVORI
prisons in the same system
returning to different
communities
5. Individuals similar to SVORI
participants from another
site
Multiple
prisons

Post-Release
Multiple Communities
Best Comparison
Individuals at the participating
prisons returning to the same
communities who are (very)
similar to SVORI participants
but do not participate—
optimally for reasons other than
motivation or other individual
characteristics that would be
expected to be related to future
success
Other Options
1. Individuals from different
prisons in the same system
who would have been
SVORI-eligible who are
returning to the SVORI
target communities
2. Individuals from SVORI
prison who are different in
measurable ways from
SVORI participants who are
returning to the SVORI
target communities
3. Individuals from SVORI
prison returning to
communities similar to the
SVORI community
4. Individuals from non-SVORI
prisons in the same system
returning to different
communities
5. Individuals similar to SVORI
participants from another
site
Best Comparison
Individuals at the participating
prisons returning to the same
communities who are (very)
similar to SVORI participants
but do not participate—
optimally for reasons other than
motivation or other individual
characteristics that would be
expected to be related to future
success

Statewide
Individuals at the participating
prison returning to the same
communities (i.e., anywhere in
the state) who are (very) similar
to SVORI participants but do
not participate—optimally for
reasons other than motivation
or other individual
characteristics that would be
expected to be related to future
success
1. Individuals from different
prisons in the same system
who would have been
SVORI-eligible who are
returning to the SVORI
target communities (i.e.,
anywhere in the state)
2. Individuals from SVORI
prison who are different in
measurable ways from
SVORI participants who are
returning to the SVORI
target communities (i.e.,
anywhere in the state)
3. NA
4. NA
5. Individuals similar to SVORI
participants from another
site

Individuals at the participating
prisons returning to the same
community (i.e., anywhere in
the state) who are (very) similar
to SVORI participants but do
not participate—optimally for
reasons other than motivation
or other individual
characteristics that would be
expected to be related to future
success
(continued)

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 12. Options for the identification of comparison population pools by program
geographic composition (continued)
PreRelease

Single Community

Multiple
prisons
(cont.)

1. Individuals from different
prisons in the same system
who would have been
SVORI-eligible who are
returning to the SVORI
target community
2. Individuals from SVORI
prisons who are different in
measurable ways from
SVORI participants who are
returning to the SVORI
target community
3. Individuals from SVORI
prisons returning to
communities similar to the
SVORI community
4. Individuals from non-SVORI
prisons in the same system
returning to different
communities
5. Individuals similar to SVORI
participants from another
site

All
prisons

Individuals at the participating
prisons returning to the same
community who are (very)
similar to SVORI participants
but do not participate—
optimally for reasons other than
motivation or other individual
characteristics that would be
expected to be related to future
success
1. NA
2. Individuals from SVORI
prisons who are different in
measurable ways from
SVORI participants who are
returning to the SVORI
target community
3. Individuals from SVORI
prisons returning to
communities similar to the
SVORI community
4. NA
5. Individuals similar to SVORI
participants from another
site

Post-Release
Multiple Communities
Other Options
1. Individuals from different
prisons in the same system
who would have been
SVORI-eligible who are
returning to the SVORI
target communities
2. Individuals from SVORI
prisons who are different in
measurable ways from
SVORI participants who are
returning to the SVORI
target communities
3. Individuals from SVORI
prisons returning to
communities similar to the
SVORI community
4. Individuals from non-SVORI
prisons in the same system
returning to different
communities
5. Individuals similar to SVORI
participants from another
site
Best Comparison
Individuals at the participating
prisons returning to the same
community who are (very)
similar to SVORI participants
but do not participate—
optimally for reasons other than
motivation or other individual
characteristics that would be
expected to be related to future
success
Other Options
1. NA
2. Individuals from SVORI
prisons who are different in
measurable ways from
SVORI participants who are
returning to the SVORI
target communities
3. Individuals from SVORI
prisons returning to
communities similar to the
SVORI community
4. NA
5. Individuals similar to SVORI
participants from another
site

Statewide
1. Individuals from different
prisons in the same system
who would have been
SVORI-eligible who are
returning to the SVORI
target communities (i.e.,
anywhere in the state)
2. Individuals from SVORI
prisons who are different in
measurable ways from
SVORI participants who are
returning to the SVORI
target communities (i.e.,
anywhere in the state)
3. NA
4. NA
5. Individuals similar to SVORI
participants from another
site

Individuals at the participating
prisons returning to the same
community who are (very)
similar to SVORI participants
but do not participate—
optimally for reasons other than
motivation or other individual
characteristics that would be
expected to be related to future
success
1. NA
2. Individuals from SVORI
prisons who are different in
measurable ways from
SVORI participants who are
returning to the SVORI
target community
3. NA
4. NA
5. Individuals similar to SVORI
participants from another
site

Identification of Comparison Population Pools and Evaluation Participants

Exhibit 13. Evaluation selection criteria for SVORI and non-SVORI comparison groups by
site
Site

SVORI Population Selected for Evaluation

Comparison Population Selected for
Evaluation

Adult Programs
Iowa

Individuals randomly assigned to the KEYS
group and projected to be (and actually were)
released during the baseline enrollment
period. (For the individuals who were
released to work-release facilities, “release”
was defined as release from the workrelease facilities, not from the original
institution. Therefore, although the KEYS
curriculum was delivered in only three
prisons, SVORI interviews took place in five
facilities—the three prisons of interest and
the two work-release facilities: the Fort Des
Moines Community Corrections Center and
the Women’s Residential Correctional
Facility—where some KEYS participants
were sent before being released to the
community.)
Random assignment entailed the following:
once the eligible prisoners from each facility
were identified, the names were sent to the
parole board for a prescreening process (to
verify likelihood of release). Cases receiving
prescreen approval were then sent to CJJP
for random assignment, which involved an
SPSS program to randomly split the sample
of eligible and prescreened offenders into
two groups and then identify one group as
the KEYS group and the other as the control
group (Note: the groups were not equally
distributed—KEYS slots were filled first and
then the remaining individuals were allocated
to the control group). An intent-to-treat
design was employed (drop-outs could not
be considered control group members).

Individuals randomly assigned to the control
group and projected to be (and actually were)
released during the baseline enrollment
period. Random assignment procedures are
described in the “SVORI population selected
for Evaluation” column.

Indiana

Individuals identified as eligible for the
Community Transitions Program (CTP) who
had not declined the program, who were
released from one of seven designated
facilities for the evaluation (Indiana Women’s
Prison, Westville, Chain O’Lakes, Rockville,
Plainfield, Putnamville, Miami), and who were
projected to be (and actually were) released
to Allen County Community Corrections
during the baseline enrollment period.

Individuals incarcerated in the same seven
facilities from which CTP participants could
come but who were returning to Marion
(rather than Allen) County, who matched the
selection criteria used for the CTP program,
and who were projected to be (and actually
were) released during the baseline
enrollment period. If it became known that
comparison group members were enrolled in
the CTP program or the Community Chaplain
Program in Marion County, the cases were
dropped from further follow-up.
(continued)

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 13. Evaluation selection criteria for SVORI and non-SVORI comparison groups by
site (continued)
Site

SVORI Population Selected for Evaluation

Comparison Population Selected for
Evaluation

Adult Programs (cont.)
Kansas

All Shawnee County Reentry Program
(SCRP) participants who were projected to
be (and actually were) released during the
baseline enrollment period.

Individuals in the SCRP facilities (Lansing
and Topeka prisons) who met all of the
SVORI eligibility criteria, were projected to be
(and actually were) released during the
baseline enrollment period, but who were
returning to Sedgwick County (using preprison county of residence as a proxy, since
the DOC database does not have a field for
projected post-incarceration residence).

Maine

All SVORI participants from all facilities who
were projected to be (and actually were)
released during the baseline enrollment
period and who were from Washington,
Androscoggin, Penobscot, and Knox
counties. (Note: on 3/31/2005, it was decided
to stop recruiting participants from the
juvenile system [i.e., the two juvenile
facilities] for interviews.)

Four comparison counties selected to identify
male comparisons: Piscataquis County (to
compare with SVORI Washington County),
York County (to compare with SVORI
Androscoggin County), Kennebec County (to
compare with SVORI Penobscot County),
Lincoln County (to compare with SVORI
Knox County). There is no comparison group
of women in Maine because women
returning to comparison counties receive
SVORI-like services through another
contract.

Maryland

Individuals flagged for Maryland Reentry
Partnership (REP) participation who were
projected to be (and actually were) released
during the baseline enrollment period.

Missouri

Male SVORI participants at Crossroads
Correctional Center (Cameron), Western
Missouri Correctional Center (Cameron), or
Western Reception Diagnostic and
Correctional Center (St. Joseph) who were
returning to specific zip codes in Kansas City.
Female SVORI participants in Chillicothe
Correctional Center (Chillicothe) and
Women's Eastern Reception Diagnostic and
Correctional Center (Vandalia) who were
returning to specific zip codes in Kansas City.

Individuals who were housed at MTC, were
not enrolled in any specific reentry
programming, met all other program eligibility
criteria except zip code (comparison
individuals were those returning to Baltimore
zip codes other than those targeted by the
REP program), and were projected to be
(and actually were) released during the
baseline enrollment period.
Males and females returning to specific zip
codes in Kansas City.

(continued)

Identification of Comparison Population Pools and Evaluation Participants

Exhibit 13. Evaluation selection criteria for SVORI and non-SVORI comparison groups by
site (continued)
Site

SVORI Population Selected for Evaluation

Comparison Population Selected for
Evaluation

Adult Programs (cont.)
Nevada

All individuals 18+ who were enrolled in the
Going Home Prepared (GHP) program (and
therefore transferred to one of the three
facilities in which programming was
delivered), who were projected to be (and
actually were) released during the baseline
enrollment period, and who were returning to
the Southern Nevada county of Clark,
Lincoln, Nye, or Esmeralda.

Individuals who met all GHP eligibility criteria
except for the “subjective” exclusion criteria
(e.g., inability to transfer to program facility,
insufficient time left on sentence) and postrelease geographic parameters.a
Comparison subjects came from facilities
located in Southern Nevada (High Desert
State Prison, Indian Springs Conservation
Camp, Jean Conservation Camp (female),
Southern Desert Correctional Center,
Southern Nevada Women’s Correctional
Center), and (until 9/23/2004) Northwest
Nevada. Comparison group members were
projected to be (and actually were) released
during the baseline enrollment period.

Ohio

Individuals randomly assigned to the Ohio
Community-Oriented Reentry Program
(CORE) program and who were projected to
be (and actually were) released during the
baseline enrollment period. Random
assignment entailed the following: offenders
who meet the eligibility requirements met
with program staff to explain the program,
ascertain interest, and make a final eligibility
determination; the remaining inmates were
randomly assigned to the experimental or
control group.

Individuals randomly assigned to the control
group and projected to be (and actually were)
released during the baseline enrollment
period. Random assignment procedures
described in the “SVORI Population Selected
for Evaluation” column.

Oklahoma

Pennsylvania

All PROTECT participants from the 19 state
prisons (and, beginning 11/23/04, any of the
four private prisons and six community
corrections centers) who were projected to be
(and actually were) released during the
baseline enrollment period.

All EPRP participants who were projected to
be (and actually were) released from Erie
CCC (males) and Gaudenzia (females) to the
community during the baseline enrollment
period.

Individuals who met the PROTECT age

and LSI criteria, were projected to be
(and actually were) released during the
baseline enrollment period, but who
were projected to return to Tulsa county
(based on pre-incarceration county of
residence).
Male state parolees, state re-parolees,
technical parole violators (TPVs) with
community parole center (CPC) placements
("halfway-backs"), and pre-release cases
who were between 18 and 35 years old,
returning to Erie, Crawford, or Warren
County, and were projected to be (and
actually were) released from Erie CCC or
Gateway Erie (another treatment facility
contracted by DOC) during the baseline
enrollment period. There was no female
comparison group.
(continued)

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 13. Evaluation selection criteria for SVORI and non-SVORI comparison groups by
site (continued)
Site

SVORI Population Selected for Evaluation

Comparison Population Selected for
Evaluation

Adult Programs (cont.)
South
Carolina

All individuals enrolled in SVORI and who
were projected to be (and actually were)
released during the baseline enrollment
period.

Three comparison groups were identified as
follows: (a) individuals from the seven SVORI
facilities who met all SVORI eligibility criteria
(including post-release supervision), (b) nonsex offenders from Tyger River who met all
SVORI eligibility criteria (including postrelease supervision), and (c) individuals who
met all SVORI eligibility criteria (including
post-release supervision) but who were
incarcerated in three non-SVORI medium- or
maximum-security facilities (i.e., Camille
Griffin Graham [women’s], Allendale, and
Evans). For all three comparison groups,
individuals must have been projected to be
(and actually were) released during the
baseline enrollment period.

Washington

Going Home participants returning to King
and Pierce Counties (Spokane was
excluded) who were projected to be (and
actually were) released during the baseline
enrollment period.

Individuals who met program criteria, were
incarcerated in six correctional facilities and
11 work-release facilities (near Seattle and in
and around Walla Walla) in which Going
Home programming was not offered, were
returning to Pierce and King Counties, and
were projected to be (and actually were)
released during the baseline enrollment
period.

Juvenile Programs
Colorado

Male Colorado Affirms Reentry Efforts
(CARE) participants incarcerated in all
facilities served by the program (Lookout
Mountain, Ridgeview, and later, Everest) who
were projected to be (and actually were)
released during the baseline enrollment
period.

Male youth within the CARE facilities who
were supervised by case managers other
than the ones from which CARE participants
were recruited, who met all other CARE
criteria (e.g., released to the Denver/Metro
area, scored 28+ on the CLSI), and who
were projected to be (and actually were)
released during the baseline enrollment
period.

Florida

All SVORI participants at the Miami-Dade
site who were projected to be (and actually
were) released during the baseline
enrollment period. Enrollment was
discontinued for females.

Youth who met the same program eligibility
criteria, were committed to the same
facilities, and were projected to be (and
actually were) released during the baseline
enrollment period, but who were committed
from Broward and West Palm (rather than
Dade) counties. Note that YLS/CMI scores
are not available on youth who are not
participating in SVORI so this criterion could
not be applied to them.
(continued)

Identification of Comparison Population Pools and Evaluation Participants

Exhibit 13. Evaluation selection criteria for SVORI and non-SVORI comparison groups by
site (continued)
Site

SVORI Population Selected for Evaluation

Comparison Population Selected for
Evaluation

Juvenile Programs (cont.)

Kansas

All GHI participants in the Topeka Juvenile
Correctional Facility (JCF) who were
projected to be (and actually were) released
during the baseline enrollment period.

Individuals at Topeka JCF with a conditional
release period of at least six months who
were projected to be (and actually were)
released during the baseline enrollment
period, but who were returning to Shawnee
County, Johnson County*, or Wyandotte
County*. [*Note: The GHI program in
Johnson and Wyandotte counties was not
operational during much of the baseline
enrollment period, so individuals who
otherwise would have been GHI participants
were enrolled in the evaluation as
comparison subjects. During the brief period
of time in the baseline enrollment period
when the GHI program in those counties was
operational, only individuals returning to
Shawnee County were selected as
comparisons for the evaluation.]

South
Carolina

All individuals enrolled in SVORI and who
were projected to be (and actually were)
released during the baseline enrollment
period.

The comparison group primarily consisted of
youth incarcerated in the same facilities
served by SVORI but who were committed
from different counties (i.e., Greenville,
Aiken, and Sumter) but also included a small
number of youth incarcerated in the same
facilities as served by SVORI and committed
from the same counties. (These comparison
group members were likely not enrolled in
SVORI because of case-flow caps
established for the reentry coordinators.)

Interviewing was discontinued in northern Nevada due to small numbers of eligible participants in late September
2004; however, before that time, comparison subjects could have released to a county in northwest Nevada,
including Washoe, Churchill, Carson City, Douglas, and Lyon counties.

EVALUATION ELIGIBILITY
Exhibit 14 shows the path from the potential respondent pool to
the respondent pool. Three criteria were used to identify the
individuals in the potential respondent pool to be approached to
participate in the evaluation: (1) expected release within the
next 3 months between July 2004 and November 2005 (Wave 1
data collection period); (2) housed in a facility where Wave 1
interviews were being conducted; and (3) access to the
individual was allowed (e.g., the individual was not in
segregation or away from the prison for court appearances or
medical treatment).

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 14. Identification of evaluation respondents

Individuals who were potential respondents and had expected
release dates within 3 months were included on lists provided
periodically by the program to the evaluation site liaisons. 13
These lists were reviewed by the site liaison and then provided
to the field data collection team task leader who reviewed the
lists and uploaded cases to a computerized case management
13

Electronic files containing name, prison, state identification number,
date of birth, and expected release date were posted by state
agency personnel to a secure FTP site, triggering an e-mail to a
data manager who immediately moved the file to a server behind
RTI’s firewall and notified the site liaison that the file was available
for processing. Each site liaison maintained the master list of names
for his/her site and passed the processed lists on to the evaluation’s
data collection task leader. Because names of individuals were
acquired on a rolling basis (typically monthly), it was necessary to
double-check monthly lists against prior lists to ensure that
duplicate names were removed. Duplicates occurred because
eligibility was based on an individual’s expected release date
(expected to be released between 60 and 90 days in the future)
and an individual who wasn’t released when expected could end up
on subsequent lists.

Identification of Comparison Population Pools and Evaluation Participants

system that was designed for the project. Uploaded cases were
assigned to field supervisors who subsequently assigned the
cases to field interviewers. The field interviewers, in turn,
contacted the facilities to verify the locations of potential
respondents. If, for example, an individual was not at the
indicated facility, sometimes the facility contact could identify
where the individual had been transferred; if not, the site
liaison contacted his/her agency contact to obtain updated
information. If the individual had been transferred to a facility
in which interviews were being conducted, the records were
updated and the field interviewer approached the new facility
with a request to interview the individual.
On some occasions, individuals were released before an
interview could be conducted with him or her. This was a
particular problem during the start-up of the interviewing
process, when individuals with expected release dates within 30
days were initially targeted. This 30-day time period turned out
to be too short and resulted in substantial “leakage” from the
potential respondent pool. The protocol was adjusted to obtain
names 3 months before expected release; this adjustment
greatly reduced the number of individuals who were released
before interviews could be scheduled.
The third step was to gain access to the individual in the
facility. This was generally not a problem, but on occasion
access was denied. Also occasionally, the individual was not at
the facility when the interviewer arrived for the interview (e.g.,
the individual had been transported to court or for medical
treatment). Finally, on a few occasions, a facility was in
“lockdown,” or a private space was not available in which the
interview could be conducted. In all of these cases, the field
interviewers continued to try to obtain the interview until either
(1) the individual was released or transferred to a facility that
was not included in the evaluation or (2) a change in the
expected release date (e.g., parole was denied) made the
individual ineligible for the evaluation because the individual
would not be released during the data collection period.
Once access to the individual was obtained, the field
interviewer explained the study and went through the consent
process. Individuals who agreed to participate became
respondents. Nonrespondents included those who were moved
to a facility not participating in the study, those to whom the

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

interviewers were not provided access, those who refused, and
those who were not released during the data collection period.
The data collection procedures are detailed in the following
section.

Data Collection
Procedures
Data were collected from four sources for the implementation
assessment and the impact evaluation. These sources,
purposes, modes, and dates of data collection are shown in
Exhibit 15.
The planning and design data collection activities were
described in the previous section. The following sections
provide details on the implementation and impact data
collection activities. The next section describes data collection
from SVORI grantees and program directors. Subsequently,
data collection from impact evaluation participants is described.
The final section in this chapter describes the collection and
processing of data from administrative sources.

IMPLEMENTATION: SVORI GRANTEE AND
PROGRAM DIRECTOR SURVEYS
The primary source of data for the implementation assessment
was four rounds of data collection from the SVORI program
directors.
1. The initial data collection from the program directors
provided basic information on the nature of the local
SVORI program(s), including information on program
focus and components, as well as the anticipated
enrollment; the target population(s), including inclusion
and exclusion criteria; whether the program(s) was
(were) targeting one, a few, or all institutions prerelease and one, a few, or all communities statewide
post-release (geographic criteria); and program goals.
As discussed earlier, data collection included a review by
the SVORI site liaisons of the 69 telephone interviews

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 15. Data sources for the Multi-site Evaluation of SVORI

Source
Purpose
Planning and Designa
Grantee
Extract information of
applications and
program characteristics,
work plans
target population(s)
SVORI program
directors—
selected sites
Implementation
SVORI program
directors

SVORI
programs—impact
sites only
Impact Evaluation
Treatment and
comparison
subjects

State agency data

National Crime
Information Center
(NCIC) data

Assess program(s) for
inclusion in the SVORI multisite impact evaluation
Collect information on
program characteristics and
status

Collect information on plans
for sustaining SVORI
program elements and other
reentry activities
Review program status;
discuss administrative data
requirements
Wave 1 interview
(approximately 30 days
before expected release)
Wave 2 interview
(approximately 3 months
after release)
Wave 3 interview
(approximately 9 months
after release)
Wave 4 interview
(approximately 15 months
after release)
Criminal history and
recidivism information
(incarceration and
probation/parole)
Criminal history and
recidivism information
(arrest)

These activities were described in the Site Selection section.

Mode

Dates

SVORI site team
review; telephone
follow-up to clarify and
complete information
Site visits

August–October 2003

1.Telephone follow-up
after work plan
extraction
2 & 3. Paper
interviews mailed to
SVORI program
directors; telephone
follow-up to ensure
response and clarify
answers
4. E-mail survey;
telephone follow-up to
encourage response

1. August–October
2003
2. March 2005
3. March 2006

Site visits

2005
2006

Computer-assisted
personal interviewing
(CAPI)
CAPI
Oral swab drug test

July 2004–November
2005

CAPI

April 2005–October
2006

CAPI
Oral swab drug test

October 2005–April
2007

Electronic files
provided by state
agencies

February 2007–March
2009

PDF and hardcopy
arrest records

August 2008; March
2009: second request
to obtain records not
returned in response to
the initial request

October–December
2003

4. July 2007

October 2004–April
2006

Data Collection Procedures

with the program directors. This work began in August
2003 and concluded in October 2003. This effort
identified a total of 88 separate SVORI programs (an
additional program was later discovered, resulting in a
total of 89 separate SVORI programs that were
proposed by the 69 SVORI grantees). Site liaisons
entered data into a Web-based data entry site that was
established on the SVORIMIS. (A copy of the data
collection elements is included in Appendix A.) All 69
grantees responded. Results, including descriptions of all
programs, were stored in a project management
information system (SVORIMIS) and are summarized in
Lattimore et al. (2004).
2. The program directors were mailed hard copy surveys in
March 2005. This survey collected additional information
on the planned structure of the SVORI program,
enrollment to date, and information on barriers and
challenges to implementation. Information on the types
of programming that would have been available for
SVORI participants in the absence of the SVORI program
was also collected. The site liaisons entered the data on
the returned questionnaires into the SVORIMIS, followed
up with delinquent respondents, and contacted program
directors by telephone to verify and clarify responses. (A
copy of the survey is included in Appendix A.) Responses
were received from 89 of the 89 program directors,
although not every director responded to every
question.
3. A second survey was mailed to the program directors in
March 2006. This survey collected updated information
on enrollment, as well as services provided,
implementation, and sustainability. The site liaisons
entered the data on the returned questionnaires into the
SVORIMIS, followed up with delinquent respondents,
and contacted program directors by telephone to verify
and clarify responses. (A copy of the survey is included
in Appendix A.) Responses were received from 86 of the
89 program directors.
4. A final survey was e-mailed to the 89 program directors
in July 2007 to obtain information on ongoing reentry
efforts in their states, after the conclusion of the SVORI
grants. (A copy of the questions is included in
Appendix A.) Data were keyed by project staff, who also
made follow-up telephone and e-mail inquiries to
increase response rates. Responses were obtained from
52 of the 89 programs.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

In addition to the program director surveys, which generated
descriptive data (albeit self-reported) for all 89 SVORI
programs, two rounds of site visits were conducted with the
subset of programs included in the impact evaluation. The site
visits generated detailed information from a variety of key
stakeholders involved in SVORI (including line staff,
supervisors, and top administrators from the pre- and postrelease supervision agencies, service provider agencies, and
other key partners) and enabled the evaluation team to more
fully characterize program implementation, interagency
collaboration, and sustainability in the sites selected for the
impact evaluation. Copies of the site visit protocols are included
in Appendix A.

IMPACT: INTERVIEWS WITH SVORI
PROGRAM PARTICIPANTS AND
COMPARISON SUBJECTS
The data collection consisted of four in-person interviews with
offenders (approximately 1 month before release and 3, 9, and
15 months after release). Drug tests (oral swabs) were
conducted at the 3- and 15-month interviews with individuals
who were in the community (i.e., not in a correctional or
treatment facility) at the time of the interview and provided a
separate consent for the tests.
As described previously, eligible respondents (both SVORI and
comparison) were identified on a monthly (or more frequent)
basis during a 16-month Wave 1 (pre-release) interviewing
period (July 31, 2004, through November 30, 2005). A
computerized case management system was used to assign
cases to field interviewers and to track the status of fielded
cases. Reasons that interviews were not conducted were
tracked carefully and field interviewers were provided
assistance if their data suggested that they were having
difficulty (e.g., with conversions).
All interviews were conducted in private settings, using
computer-assisted personal interviewing (CAPI) by experienced
RTI field interviewers who had completed extensive training on
interviewing in correctional settings and locating difficult-to-find
respondents. The training also included modules on human
subjects’ regulations and administering consents (assents for
juvenile subjects, passive consent for parents/guardians of

Data Collection Procedures

juvenile subjects), as well as information on identifying and
responding to distressed respondents. Copies of consent forms
are included in Appendix E.
Wave 1 (Pre-release) Interviews

Wave 1 interviews were conducted in about 150 prisons and
juvenile detention facilities across the country. 14 Field
interviewers contacted the facility where the potential
respondent was housed and requested an appointment through
the evaluation’s facility contact. The MOUs with the facilities
included a specification that the potential respondent was to be
told only that he or she had a visitor and was not to be
provided any information about the evaluation or the potential
interview. All information about the research was provided by
the field interviewers. The reason for this was (1) to ensure
that facility staff did not coerce participation and (2) to ensure
that information about the study was provided in a consistent
manner (as dictated by the interviewer training).
Each potential respondent was shown a brochure describing the
research study (see Appendix E) and the field interviewer
described the project. Individuals who indicated that they were
willing to participate were read the consent (or assent) form.
The consent was witnessed, not signed, to minimize the chance
of revealing the identity of a study participant (e.g., if the
consent form was lost by the shipping company when it was
returned to RTI). For each juvenile subject, a letter describing
the research project was mailed several weeks before the
anticipated interview date to the parent or guardian whose
name had been provided by the juvenile justice agency. The
letter described the study and provided a toll-free number to
call if the parent/guardian wanted additional information or did
not want the juvenile to participate in the research. The
juvenile subjects were also queried to ensure that they
understood the materials in the assent forms. The juvenile
subject was asked eight questions about the content and
meaning of the consent form before the field interviewer began
the survey. A copy of the questionnaire is included in
Appendix E. The remaining procedures for contacting facilities
14

Interviews were arranged through communication between site
team staff, facility staff, and field interviewers. All descriptions of
and explanations concerning the study purpose were provided to
potential respondents by the field interviewers, who also
administered the consent procedures.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

and potential juvenile respondents were similar to those for the
adults.
Wave 1 interviews lasted approximately 1.5 hours and were
conducted approximately 1 month before release. The interview
was designed to obtain data on respondents’ experiences and
receipt of services during incarceration, as well as document
respondents’ immediate post-release plans. 15 No incentive was
provided for the Wave 1 interviews.
Waves 2, 3, and 4 (Follow-up) Interviews

Follow-up interviews were conducted at 3, 9, and 15 months
after release. The follow-up interviews lasted approximately 1.5
hours and covered topics such as housing, employment,
education, family, peer relationships, community involvement,
physical and mental health, substance use, crime and
delinquency, supervision, service needs, and service receipt.
The follow-up interviews were conducted in the community or,
for those reincarcerated, in prison or jail (if possible). 16 For
interviews conducted in the community, respondents were paid
$35 for the 3-month interview, $40 for the 9-month interview,
and $50 for the 15-month interview. At the final interview,
respondents were paid an additional $50 if they completed all
four interviews. In addition, respondents were paid an extra $5
at each follow-up wave if they called a toll-free number to
schedule their interview. As the 15-month interviews began,
the original protocol with respect to compensation was
adjusted. Specifically, where agreements could be negotiated
with corrections departments and local jails, participants who
were incarcerated were provided compensation. The reason for
this change was to boost response rates, but also out of
fairness—the additional $50 for completing all four interviews
that was promised at the time of earlier interviews would not

In most sites, SVORI programming began several months before
release, although there was wide variability both within and across
sites.
Follow-up interviews were also conducted in treatment facilities,
when possible. Facility MOUs were negotiated with all prisons, jails,
and treatment facilities to protect the confidentiality of the
participants and the data collected from them. Site liaisons were
responsible for making contact with facilities that were not included
in the Wave 1 interviews and for negotiating MOUs. Information on
all facilities (i.e., prisons, jails, juvenile detention facilities, or
treatment facilities) was maintained in the SVORIMIS.

Data Collection Procedures

have been available to these participants. 17 Compensation was
provided either (1) to the inmate’s canteen or personal account
at the prison (or jail) or (2) to an individual in the community
whose name and address were provided by the inmate. In the
second instance, the offender had to acknowledge that the
study team took no responsibility other than to mail the
compensation to the person the offender had identified;
offenders were specifically told that there was no guarantee
that the money would be held for them.
Oral swab drug tests were conducted in conjunction with the 3and 15-month interviews conducted in the community (i.e., not
in prisons, jails, or treatment facilities). Respondents were
provided an additional $15 if they consented to provide an oral
swab. The field interviewers were trained to collect, package,
and mail the oral swabs to a drug testing laboratory. 18 The
chosen test was a six-panel oral fluid screen for amphetamines,
cannabinoids, cocaine, methamphetamines, opiates, and
phencyclidine. All positive findings were confirmed by gas
chromatography/mass spectrometry.

IMPACT: ADMINISTRATIVE DATA 19
The evaluation requested official criminal records data to
supplement the self-reported interview data, particularly with
respect to measures of criminal history and recidivism. The two
sources of data were (1) state DOC/DJJ/Probation and Parole
(P&P) agencies and (2) the National Crime Information Center
(NCIC). State DOC/DJJ agencies provided data on return to
prison after being released, as well as information on
performance during post-release parole or probation. The NCIC
provided data on arrests, including prior arrests and rearrests,
as well as information on convictions and reincarcerations for
some states. This section first describes the acquisition and
processing of state agency data. Subsequently, the procedures
for acquiring and processing NCIC records are described.
17

Some agencies had a firm policy against the payment of
compensation to inmates. In those sites, compensation was not
provided to incarcerated participants.
Scientific Testing Laboratories, Inc. (STL) was contracted for the
drug testing. STL provided intercept collection oral fluid devices,
biohazard bags, packaging materials for shipping, and chain of
custody forms, as well as test and confirmatory test results. STL
was acquired by Kroll Laboratory Specialists, Inc., in November
2005; the contract was continued with Kroll.
Mark Pope and Debbie Dawes of RTI contributed to this section.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Acquisition and Processing of State Agency Data

The administrative data collection and processing task was a
substantial undertaking for several reasons:

It required coordination between the evaluation and 19
state agencies over the 14 impact sites.

Data elements of interest available at each site varied
such that it was necessary to customize the approach
taken for each site.

The format of each site’s recidivism data was different,
requiring extensive programming to make each site’s
data useable.

An administrative data protocol was drafted that was designed
to guide site teams in their negotiations with the impact sites
and to be shared with appropriate individuals in each site to
provide specific information on the data request. The protocol
also provided information on the procedures implemented to
ensure the secure transmission of data between the sites and
RTI. The protocol described the project (including how the
administrative data supplemented the interview data), the data
sources being accessed, the individuals for whom data were
needed, and the data elements being requested. (A copy of the
protocol is included in Appendix F.) This protocol was reviewed
and approved by an RTI IRB.
The following paragraphs describe the procedures that were
followed to acquire and process the agencies’ data. These
procedures included (1) negotiating with sites, (2) establishing
procedures for transferring data from the state agencies to RTI;
(3) identifying subjects for whom data were needed;
(4) processing data; and, importantly, (5) acquiring data from
online criminal history repositories to supplement data provided
by some states.
Negotiating with Sites. After the protocol was drafted and
reviewed, it was submitted to the RTI IRB for review and
approval. Once approval was obtained, site liaisons coordinated
with SVORI program directors and agency personnel to identify
an initial administrative data contact for each relevant state
agency. This individual either served as the point of contact for
the administrative data collection or provided a referral to the
appropriate individual at the agency. After these individuals
were identified, they were sent a notification letter signed by
the evaluation co-PIs informing them that a member of the

Data Collection Procedures

evaluation team would be contacting them to begin
negotiations to obtain administrative data.
The negotiation process with each agency included completing
a data use agreement, identifying the desired data elements
that were available from their administrative data system(s),
describing the individuals for whom data would be needed, and
determining the format in which the data would be provided.
The negotiation process was ongoing, with some sites able to
provide data sooner than others depending on their data
systems. In addition, the amount and quality of data available
varied substantially across the states.
Transferring Data. The project’s IRB requirements
necessitated that the administrative data be transferred in a
secure manner. To meet these requirements, two data transfer
options were available. First, each agency could securely upload
their data to RTI using the evaluation’s Web site to access an
upload process that encrypted the data file(s) using SSL during
transmission to RTI; the encryption protected the data during
transmission. Second, the site could send the data file(s) to RTI
on a password-protected CD using Federal Express. In both
cases, all data files were separated such that identifiers (e.g.,
name, address) and data elements (e.g., incarceration
variables) were in two separate files linked by a common sitegenerated unique identifier. These separated files were
uploaded separately or sent via FedEx in separate shipments to
ensure that if one file was lost or intercepted it did not contain
identifiers and data elements together. As data were received
by the evaluation team, the files were stored on an encrypted
drive. The data transfer protocol is included in Appendix F.
Identifying Subjects. An important step in the negotiation
process entailed identifying the individuals who were to be
included in each site’s data extraction. The evaluation needed
to obtain data for all individuals who participated in pre-release
interviews; however, information on an expanded sample would
provide an opportunity to examine whether the respondents
were comparable to those who refused to be interviewed. Thus,
the sites were asked to provide data for either

all individuals enrolled in the SVORI program and
released between the start of the program and
December 31, 2005, and all individuals comparable to

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

those individuals who were released in 2004 and 2005
(but who did not participate in SVORI), or

all individuals on the lists of eligible respondents
provided by the agency during the Wave 1 (pre-release)
enrollment period for the offender interviews (July
2005–November 2005).

In a few instances, the sites could not identify either of these
two groups in their data systems. In these cases, the
evaluation team provided to the agency a list of identifiers that
was the cumulative list of identifiers received from the site as
potential respondents.
Finally, in some cases, agencies routinely prepare annual
admission and release cohort research files and preferred to
provide those files to the evaluation team rather than conduct a
special data run. In such cases, the evaluation team accepted
those files and matched the list of evaluation subjects against
the cohort files to identify release dates and new admissions, as
well as criminal history information.
Processing Instant Incarceration and Reincarceration
Data. Processing began on each site’s administrative data files
as they were received. Because each site’s files differed in
format (e.g., Excel spreadsheets, relational data tables) and
content, there was considerable variation in common data
elements. The minimum common set of needed variables was
limited to a few key variables related to the instant
incarceration and reincarceration events; these variables were
extracted from each site’s data. Among the data elements of
primary interest to the evaluation were

1. the admission and release dates of the “instant”
incarceration (i.e., the incarceration event that led to
participation in SVORI or, for comparisons, the event
that led to inclusion in the study),
2. the most serious offense associated with the instant
incarceration,
3. prison admissions subsequent to the instant
incarceration release date, and
4. the most serious offense associated with or reason for
(e.g., technical violation, new offense) reincarceration.
Not all sites had sufficient data available to identify the offenses
associated with the instant incarceration or reincarceration for
all sites. If data allowed, other site-specific data items were

Data Collection Procedures

also extracted, such as the number of prison infractions,
participation in programs/services, and prior incarcerations.
Obtaining Recidivism Data from Online Resources. In a
few cases where the data provided by the states were
incomplete (i.e., cases or variables missing), the databases
were supplemented with information obtained from online
criminal record sites. This somewhat tedious process involved
searching for each individual separately on the appropriate Web
site(s), downloading the record to PDF, and manually extracting
needed information. All files were stored on an encrypted drive.
Acquisition and Processing of NCIC Criminal
History/Arrest Data

Originally, arrest data were to be obtained from each state;
however, after discussions with NIJ, the evaluation team
decided to obtain arrest data from the NCIC. The rationale was
twofold: Obtaining the data would entail a single data
use/research agreement as opposed to 14 (one for each of the
impact sites); and arrest data would be available from all states
that participate in the NCIC, ensuring that arrests that occurred
outside the study’s 14 states would be included. The evaluation
team worked with NIJ, which had to request the NCIC data on
behalf of the evaluation. Initial contact with NCIC was made in
2006; work to establish a research agreement began in 2007;
approval was received from the FBI in spring 2008; and NCIC
data were obtained in the summer of 2008.
The evaluation team provided NCIC with a list of identifiers (FBI
fingerprint number, state identifier [if any], name, sex, Social
Security Number, and date of birth) for use in extracting arrest
records. The NCIC provided data in two formats depending
upon the state providing the arrest record: PDF files (three
files, each approximately 36,000 pages) and hardcopy rap
sheets (2 boxes). The PDF files were sent to RTI by the NCIC
on password-protected CDs and were stored on the project’s
encrypted drive at RTI. The hardcopy files were sent to RTI’s
Survey Support Department, where they were stored in a
locked filing cabinet in a secure area.
A second request was made to NCIC in March 2009 in an effort
to obtain records for 328 subjects whose arrest records were
not returned in the original data received from NCIC. This
second request resulted in arrest records for an additional 250

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

subjects, leaving only 78 of 2,174 without a match 20, 21. Only 53
of 1,966 adult subjects were not matched to at least one arrest
record, a match rate of 97.3%.
Separate procedures were required for the two types of
records. The following subsections describe the procedures for
processing the PDF files and the hardcopy rap sheets, creating
arrest summary variables, and implementing quality control
procedures.
Processing PDF Arrest Record Files. The match procedure
used by NCIC is broad and returned multiple potential matching
criminal history records for each study subject. As a first step in
processing both the PDF files and hardcopy records, true record
matches had to be identified among all potential record
matches provided by the NCIC. The PDF files were converted to
text files, and PERL software was used to parse and extract the
evaluation study identification and associated identifying
information from each potential matching record. A refined
electronic matching procedure was employed to identify true
matches. Once true matches were identified, the criminal
history records associated with each true match were extracted
from the file of all potential matching records using the unique
FBI number. The extracted arrest data, which consisted of text
descriptions of each arrest charge for a specific date, were
stored in an Excel spreadsheet as an array variable for each
individual, along with the date of arrest, FBI number, and
evaluation study identification (Exhibit 16).

Of the 78 without an NCIC match, 25 were juvenile subjects from
South Carolina or Kansas. NCIC does not contain juvenile records
from these two states, suggesting that these “non-matches” were
subjects who had not been processed through the adult system at
the time of data acquisition.
There were 103 cases for which no arrest history information (i.e.,
information on arrests that preceded the release date for the period
of instant incarceration) was received. Those 103 cases had at least
one arrest after the date of release of interest (so they generated
NCIC matches). Of the 103 without a documented arrest history, 96
were juveniles and 6 were adults (4 Nevada, 1 Oklahoma, 1
Pennsylvania). Many states do not submit juvenile arrests to the
NCIC.

Data Collection Procedures

Exhibit 16. Structure of parsed arrest records file

Study
ID
1
2

FBI#
A2AB4BC23
MMMNNN1

Arrest
Date
3/15/2002
2/18/1998

1
Asslt
Pos. M/J

Charge
2
3
Prob. Viol
Robbery
Pos. Drug
Paraphernalia

…

Assault

The parsing and extraction process successfully produced
distinct arrest charges from the electronic rap sheets for each
individual; the arrest text literals had to be coded into distinct
offense categories to be useful for recidivism analyses. To
categorize the arrest charges, the Excel spreadsheets of
extracted arrest charges were loaded into an Access database
for further processing. The first processing step entailed
converting the array format of the arrest record charges to a
stacked format; that is, multiple records per individual per
arrest date, each one showing a distinct arrest charge. This
step was completed by searching each array arrest variable for
records that were not null (i.e., that had text in them). For each
search, the returned records were written to a new table along
with the evaluation study identification, FBI number, and arrest
date for the charge in question. (Using the date from
Exhibit 16, the resulting format is shown in Exhibit 17.) This
process resulted in 84,429 distinct arrest records for 4,286
individuals.
Exhibit 17. Transformed arrest records data

Study ID
1
1
2
2
2

FBI#
A2AB4BC23
A2AB4BC23
MMMNNN1
MMMNNN1
MMMNNN1

Arrest Date
3/15/2002
3/15/2002
2/18/1998
2/18/1998
2/18/1998

Charge
Asslt
Prob. Viol
Pos. M/J
Pos. Drug Paraphernalia
Robbery

Once this conversion was complete, each arrest literal was
coded into a specific offense category. The offense categories
used by the National Criminal Reporting Program (NCRP) were
used for classification. To code the arrest offenses, research
staff ran a series of update queries in Access to search for
specific keywords in the arrest literal text string. If the

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

keywords were found, then the record was assigned to the
offense category related to the keywords. The offense
categorization process began by manually reviewing some
records to identify important keywords or groups of keywords
that could be used in the update queries. For example, “ASSLT”
would be coded as “Simple Assault,” while “Assault Strongarm”
would be coded as “Aggravated Assault.” This process was
iterative, as different jurisdictions coded charges differently
such that when identified, additional keywords or groups of
keywords were added to each offense category’s update query.
In addition, some records that could not be assigned using this
approach had to be manually reviewed and a determination
regarding the correct offense category made by the reviewer.
Once all arrest literals were assigned to one of the NCRP
offense categories, a final summarization was made to reduce
these categories to the following broad categories of offenses—
person/violent, property, drug offenses, public order, and
other.
Processing Hardcopy Arrest Records. Initial attempts to
convert the hardcopy rap sheets into PDF files so that the
procedures established for the PDF files could be used were
unsuccessful, primarily because of the lack of uniformity in the
format of the hardcopy forms. As a result, arrest information
had to be extracted manually from these hardcopy records.

As with the PDF files, the first step in processing the hardcopy
records was to identify true record matches from all potential
record matches provided by NCIC. In this case, the identifiers
on each criminal history record were compared to the identifiers
of study subjects. True matching records were culled from the
pool of record, and the data elements of interest (e.g., date of
arrest, charge text, disposition date, convicted offense,
disposition) were abstracted from the hardcopy records and
keyed into an Excel spreadsheet. Once the Excel spreadsheet
was complete, the same offense categorization process using
automated queries in Access was employed as was done for the
PDF files.
Creating Summary Records for Arrest Data. Once all arrest
charges were categorized, the data consisted of multiple arrest
records per individual. These records were summarized to
provide a single record for each individual. For each individual,
the arrest dates were compared to the individual’s date of

Data Collection Procedures

admission to and date of release from prison for the
incarceration that coincided with his/her inclusion in the study.
Pre- and post-release flags were then created to identify
whether or not the arrest was before or after incarceration.
Using these pre- and post-release flags, each person’s charges
were counted by rolling the offense categories into higher-order
levels consisting of person/violent, property, drug, public order,
and other charges. In addition, the date of the person’s first
arrest after release was identified and the time (in days) to first
arrest was calculated.
Quality Control. To verify the accuracy of the data extraction
procedures, quality control was implemented that consisted of
verifying the electronic data against hardcopy or PDF files for
100% of the records.

Case Flow and
Threats to Validity
The three primary threats to validity of concern to the Multi-site
Evaluation of SVORI are selection bias, nonresponse bias and
attrition bias. Each of these is potentially a threat to the ability
to draw correct inferences from evaluation findings.
This section provides a description of the flow of cases for the
interviews with SVORI program participants and comparison
subjects. Subsequently, for each of the threats, the implications
of case flow, analyses directed at attempting to understand the
potential for bias, and efforts undertaken to address the
potential threat are described.

CASE FLOW
A total of 4,354 cases were fielded for inclusion in the Multi-site
Evaluation of SVORI. Wave 1 (30 days pre-release) interviews
were obtained for 2,583 cases (59%) 22. A total of 2,391
individuals comprised the final sample. The remaining cases
included 718 cases released before interviews could be
scheduled, 635 cases that were ineligible for the evaluation,
370 refusals, 192 cases that were dropped because the
respondents were not released during the period when the first
post-release interview was being conducted, and 48 other
noninterviews. Among eligible subjects approached for
22

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

interviews, refusal rates were reasonably low—12% for adult
men, 7% for adult women, and 8% for juvenile males.
Additional detail is provided below for each of the three study
groups.
Exhibit 18 summarizes the case flow for Wave 1 interviews
scheduled with 3,527 eligible respondents between July 2004
and November 2005. (Additional information on “ineligible
respondents” is provided after Exhibit 18.) As can be seen,
2,391 interviews were completed with eligible respondents—
1,697 adult males, 357 adult females, and 337 juvenile males
(defined as those who were at least 14 years of age who were
housed in a juvenile detention facility).
Pre-release interviews were completed with 66.2% of the adult
males, 69.2% of the adult females, and 75.4% of the juvenile
males. The likelihood of completing an interview was greater for
SVORI program participants than non-SVORI comparisons for
the adult males. Specifically, 73.7% of attempted interviews
with SVORI adult male subjects were completed but only
59.9% with non-SVORI comparison males. Almost all of the
difference, however, was in the percentage of subjects who
were released before an interview could be scheduled and
completed. Fully 26.5% of the fielded non-SVORI interviews
with adult males were not completed because the subject was
released before the interview could be scheduled and
completed. In contrast, only 14.4% of the SVORI adult male
cases were missed because the potential respondent was
released before the interview. This difference may be due to
the fact that the expected release dates for SVORI program
participants that were obtained from SVORI programs were
generally more accurate than the expected release dates for
non-SVORI comparison subjects obtained from the DOC MIS.
Such an explanation, however, doesn’t apply to the adult
female or juvenile male samples for whom the likelihood of
completing an interview was similar for the SVORI participants
and the non-SVORI comparison subjects.
The most common reason that an interview was not completed
was that the subject was released before an interview could be
scheduled and completed. Release prevented interviews with
21% of the eligible adult male cases fielded, 22% of the adult
females, and 15% of the juvenile males. There was no reason

Case Flow and Threats to Validity

Exhibit 18. Outcomes of Wave 1 interview attempts with eligible respondents

SVORI
Disposition/Demographic Group
Adult Males
Interview completed-incarcerated
Released before Wave 1 interview
Final refusal by R, guardian, or other
Access to R denied by prison
R absconded
Private setting not available
R deceased
Language barrier Spanish
Language barrier Other
Physically/mentally incapable
Other noninterview
Total Eligible Adult Male Cases
Adult Females
Interview completed-incarcerated
Released before Wave 1 interview
Final refusal by R, guardian, or other
Access to R denied by prison
R absconded
Private setting not available
R deceased
Language barrier-Spanish
Language barrier-Other
Physically/mentally incapable
Other noninterview
Total Eligible Adult Female Cases
Juvenile Males
Interview completed-Incarcerated
Released before Wave 1 interview
Final refusal by R, guardian or other
Access to R denied by prison
R absconded
Private setting not available
R Deceased
Language barrier Spanish
Language barrier Other
Physically/mentally incapable
Other noninterview
Total Eligible Juvenile Male Cases

Non-SVORI
N
%

Total
N

863
169
126
6
2
2
1
1
0
1
0
1,171

73.7
14.4
10.8
0.5
0.2
0.2
0.1
0.1
0.0
0.1
0.0
100.0

834
369
169
8
3
1
0
5
1
2
1
1,393

59.9
26.5
12.1
0.6
0.2
0.1
0.0
0.4
0.1
0.1
0.1
100.0

1697
538
295
14
5
3
1
6
1
3
1
2,564

66.2
21.0
11.5
0.5
0.2
0.1
0.0
0.2
0.0
0.1
0.0
100.0

153
48
12
2
1
0
0
0
0
2
1
219

69.9
21.9
5.5
0.9
0.5
0.0
0.0
0.0
0.0
0.9
0.5
100.0

204
66
26
1
0
0
0
0
0
0
0
297

68.7
22.2
8.8
0.3
0.0
0.0
0.0
0.0
0.0
0.0
0.0
100.0

357
114
38
3
1
0
0
0
0
2
1
516

69.2
22.1
7.4
0.6
0.2
0.0
0.0
0.0
0.0
0.4
0.2
100.0

152
31
17
1
0
0
0
0
0
0
1
202

75.2
15.3
8.4
0.5
0.0
0.0
0.0
0.0
0.0
0.0
0.5
100.0

185
35
20
2
2
0
0
0
0
0
1
245

75.5
14.3
8.2
0.8
0.8
0.0
0.0
0.0
0.0
0.0
0.4
100.0

337
66
37
3
2
0
0
0
0
0
2
447

75.4
14.8
8.3
0.7
0.4
0.0
0.0
0.0
0.0
0.0
0.4
100.0

Note: R= respondent.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

that could be identified that suggested that participating in
SVORI was in any way associated with “early release” and,
therefore, this leakage is assumed to not be a potential source
of bias. 23
Refusal rates were relatively low. About 11.5% of adult males,
7.4% of adult females, and 8.3% of juvenile males refused
participation (or there was refusal by a guardian or another
individual). Exhibit 18 shows that access, absconding,
availability of private settings, and language or impairment
barriers were not substantial factors in terms of interview
completion.
Some cases were erroneously fielded and some cases that were
properly fielded at the time were subsequently determined to
be ineligible because, for example, the subject was not released
during the period in which the 3-month interview was being
fielded, 24 the group to which the respondent belonged was
dropped from the evaluation, or the individual was determined
to be ineligible for the evaluation after an interview was
completed. Exhibit 19 shows the disposition of the ineligible
cases for the Wave 1 data collection.
Exhibit 20 provides information on the attrition of cases
between the Wave 1 (pre-release) and Wave 2 (3-month postrelease) interviews. As can be seen, the primary reason that an
initial follow-up interview was not completed was that the
respondent could not be located (or his/her location was known
but the subject was unavailable). Among adult females, nonSVORI comparisons were more likely than SVORI participants
(13.7% vs. 25.0%) to be unlocatable or unavailable at Wave 2.
Although the same pattern was observed among adult males

“Early” in this case means several weeks at most. Initially, the
protocol specified that potential respondents were those eligible
individuals who were expected to be released within the next 60
days. Although interviews were generally scheduled within 2 to 4
weeks, in a substantial number of cases, the potential respondents
had been released before an interview could be completed. The
protocol was adjusted to address this problem such that cases were
identified who were expected to be released within the next 90
days.
As the purpose of the evaluation was to examine post-release
behavior, release was a criterion for evaluation eligibility.

Case Flow and Threats to Validity

Exhibit 19. Case disposition of ineligible cases (Wave 1 data collection)

SVORI
Disposition/Demographic Group
Adult Males
R transferred to non-study facility
R not released/releasing during field
period
Case fielded incorrectly
R ineligible to participate
Other (non)interview-ineligible
R ineligible—age
Other ineligible (groups dropped)
R being released to non-study
area/state
R release date unknown
Total Ineligible Adult Male Cases
Adult Females
R transferred to non-study facility
R not released/releasing during field
period
Case fielded incorrectly
R ineligible to participate
Other (non)interview-ineligible
R ineligible—age
Other ineligible (groups dropped)
R being released to non-study
area/state
R release date unknown
Total Ineligible Adult Female Cases
Juvenile Males
R transferred to non-study facility
R not released/releasing during field
period
Case fielded incorrectly
R ineligible to participate
Other (non)interview-ineligible
R ineligible—age
Other ineligible (groups dropped)
R being released to non-study
area/state
R release date unknown
Total Ineligible Juvenile Male Cases

Non-SVORI
N
%

21
100

8.9
42.6

56
92

5
86
10
0
4
7

2.1
36.6
4.3
0.0
1.7
3.0

2
235

Total
N

13.9
22.9

77
192

12.6
31.5

158
12
18
0
37
25

39.3
3.0
4.5
0.0
9.2
6.2

163
98
28
0
41
32

26.8
16.1
4.6
0.0
6.7
5.3

0.9
100.0

4
402

1.0
100.0

6
609

1.0
100.0

2
12

4.4
26.7

5
10

17.9
35.7

7
22

9.6
30.1

2
24
3
0
0
1

4.4
53.3
6.7
0.0
0.0
2.2

3
9
1
0
0
0

10.7
32.1
3.6
0.0
0.0
0.0

5
33
4
0
0
1

6.8
45.2
5.5
0.0
0.0
1.4

1
45

2.2
100.0

0
28

0.0
100.0

1
73

1.4
100.0

2
5

5.0
12.5

1
17

2.4
40.5

3
22

3.7
26.8

1
11
8
3
7
1

2.5
27.5
20.0
7.5
17.5
2.5

0
5
3
1
10
0

0.0
11.9
7.1
2.4
23.8
0.0

1
16
11
4
17
1

1.2
19.5
13.4
4.9
20.7
1.2

2
40

5.0
100.0

5
42

11.9
100.0

7
82

8.5
100.0

Note: R= respondent.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 20. Wave 2 (3-month post-release) interview case flow

Disposition/Demographic Group
Adult Males
Eligible Wave 1 cases
Eligible Wave 1 cases fielded**
Refusal (by R or guardian)
Unable to contact
Non-study facility/no access
Moved out of area
Other noninterview
Interview Completed Adult Males
Adult Females
Eligible Wave 1 cases
Eligible Wave 1 cases fielded**
Refusal (by R or guardian)
Unable to contact
Non-study facility/no access
Moved out of area
Other noninterview
Interview Completed Adult Females
Juvenile Males
Eligible Wave 1 cases
Eligible Wave 1 cases fielded**
Refusal (by R or guardian)
Unable to contact
Non-study facility/no access
Moved out of area
Other noninterview
Interview Completed Juvenile Males

SVORI
%*

Non-SVORI
N
%*

Total
N

863
837
41
222
27
12
32
529

100.0
97.0
4.8
25.7
3.1
1.4
3.7
61.3

834
801
19
263
38
20
39
455

100.0
96.0
2.3
31.5
4.6
2.4
4.7
54.6

1,697
1,638
60
485
65
32
71
984

100.0
96.5
3.5
28.6
3.4
1.9
4.2
58.0

153
149
9
21
4
0
9
110

100.0
97.4
5.9
13.7
2.6
0.0
5.9
71.9

204
199
4
51
4
5
6
134

100.0
97.5
2.0
25.0
2.0
2.5
2.9
65.7

357
348
13
72
8
5
15
244

100.0
97.5
3.6
20.2
2.2
1.4
4.2
68.3

152
151
4
33
3
3
1
105

100.0
99.3
2.7
21.8
2.0
1.3
0.7
69.1

185
183
9
32
4
8
1
131

100.0
98.9
4.9
17.5
2.2
4.3
0.5
70.8

337
334
13
65
7
11
2
236

100.0
99.1
3.9
19.3
2.1
3.3
0.6
70.0

Note: Wave 1 = 30 days pre-release. R = respondent.
*Percentage of all eligible cases, rather than percentage of eligible fielded cases (i.e., based on total number of Wave
1 study-eligible completers).
**Percentage of all eligible cases actually fielded.

(i.e., non-SVORI comparisons were more likely to be
unlocatable or unavailable than SVORI participants), the
discrepancy was not nearly as great (25.7% of SVORI vs.
31.5% of non-SVORI) as that observed among adult females.
The pattern did not hold for juvenile males; SVORI participants
were slightly more likely to be unlocatable or unavailable at
Wave 2 than the non-SVORI comparisons (19.7% vs. 18.9%,
respectively).

Case Flow and Threats to Validity

The numbers of refusals were relatively small. Only 60 adult
males (3.5%) refused to participate at Wave 2. Of those,
SVORI participants were more likely than non-SVORI
comparisons to refuse participation (4.8% and 2.3%,
respectively). Similarly, only 3.6% of adult females and 4.5%
of juveniles refused to participate (or had a guardian who
refused participation) at Wave 2, and, for both groups, SVORI
participants were more likely to refuse to participate than the
non-SVORI comparisons.
The “other noninterview” category includes cases for which the
release date was not received/unknown; subjects who had
absconded, were institutionalized, or were deceased; and cases
for which language or physical/mental impairment prevented
interview completion. Similar to Wave 1, absconding,
availability of private settings, and language or impairment
barriers were not substantial factors in terms of interview
completion for any of the groups. Among adult males; 59 of the
71 cases in this category were eligible cases that were not
fielded at Wave 2, usually because release information was not
received in time to field the case for the 3-month follow-up.
Exhibit 21 provides case flow information for the Wave 3 (9month) follow-up interviews. Most of the eligible cases were
fielded at Wave 3, and more interviews were completed at
Wave 3 than at Wave 2 for all three groups.
Exhibit 22 provides the case flow information for the 15-month
follow-up interview. Again, all cases were fielded and more
subjects were found and interviewed at Wave 4 than at
Wave 3.
Overall, response rates for follow-up interviews increased over
time for all groups. Furthermore, 80% of all subjects completed
at least one follow-up interview. Exhibit 23 shows, for each
demographic group, the percentages of subjects who completed
(1) a Wave 2 interview, (2) a Wave 3 interview, (3) a Wave 4
interview, (4) all 3 follow-up interviews, and (5) at least one
follow-up interview.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 21. Wave 3 (9-month post-release) interview case flow

Disposition/Demographic Group
Adult Males
Eligible Wave 1 cases
Eligible Wave 1 cases fieldedb
Refusal (by R or guardian)
Unable to contact
Non-study facility/no access
Moved out of area
Other noninterview
Interview Completed Adult Males
Adult Females
Eligible Wave 1 cases
Eligible Wave 1 cases fieldedb
Refusal (by R or guardian)
Unable to contact
Non-study facility/no access
Moved out of area
Other noninterview
Interview Completed Adult Females
Juvenile Males
Eligible Wave 1 cases
Eligible Wave 1 cases fieldedb
Refusal (by R or guardian)
Unable to contact
Non-study facility/no access
Moved out of area
Other noninterview
Interview Completed Juvenile Males

SVORI
%a

Non-SVORI
N
%a

Total
N

863
843
46
188
22
19
16
565

100.0
97.7
5.3
21.8
2.6
2.2
1.9
65.5

834
811
37
228
41
32
18
470

100.0
97.2
4.4
27.3
4.9
3.8
3.1
56.4

1,697
1,654
83
416
63
51
34
1,035

100.0
97.5
4.9
24.5
3.7
3.0
2.8
61.0

153
149
6
22
0
4
2
119

100.0
97.4
3.9
14.4
0.0
2.6
1.3
77.8

204
199
7
46
1
12
4
134

100.0
97.5
3.4
22.6
0.5
5.9
2.0
65.7

357
348
13
68
1
16
6
253

100.0
97.5
3.6
19.1
0.3
4.5
1.7
70.9

152
151
4
33
3
3
1
108

100.0
99.3
2.6
21.7
2.0
2.0
0.7
71.0

185
183
9
32
4
8
1
131

100.0
98.9
4.9
17.3
2.2
4.3
0.5
70.8

337
334
13
65
7
11
2
239

100.0
99.1
3.9
19.3
2.1
3.3
0.6
70.9

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

Percentage of all eligible cases, rather than percentage of eligible fielded cases (i.e., based on total number of
Wave 1 study-eligible completers).
b
Percentage of all eligible cases actually fielded.

Case Flow and Threats to Validity

Exhibit 22. Wave 4 (15-month post-release) interview case flow

SVORI
Disposition/Demographic Group
Adult Males
Eligible Wave 1 cases
Eligible Wave 1 cases fieldedb
Refusal (by R or guardian)
Unable to contact
Non-study facility/no access
Moved out of area
Other noninterview
Interview Completed Adult Males
Adult Females
Eligible Wave 1 cases
Eligible Wave 1 cases fielded
Refusal (by R or guardian)
Unable to contact
Non-study facility/no access
Moved out of area
Other noninterview
Interview Completed Adult Females
Juvenile Males
Eligible Wave 1 cases
Eligible Wave 1 cases fielded
Refusal (by R or guardian)
Unable to contact
Non-study facility/no access
Moved out of area
Other noninterview
Interview Completed Juvenile Males

Non-SVORI
N
%a

863
825
40
153
33
37
18
582

100.0
95.6
4.6
17.7
3.8
4.3
2.1
67.4

834
789
41
162
34
48
17
531

153
147
7
10
1
6
5
124

100.0
96.1
4.6
6.5
0.7
4.0
3.3
81.0

152
149
5
30
5
4
1
107

100.0
98.0
3.3
19.7
3.3
2.6
0.7
70.4

Total
N

100.0
94.6
4.9
19.5
4.1
5.8
2.0
63.8

1,697
1,614
81
315
67
85
35
1113

100.0
95.1
4.8
18.6
4.0
5.0
2.1
65.6

204
194
2
36
1
9
4
152

100.0
95.1
1.0
17.7
0.5
4.4
2.0
74.5

357
341
9
46
2
15
9
276

100.0
95.5
2.5
12.9
0.6
4.2
2.5
77.3

185
176
5
24
3
10
2
141

100.0
95.1
2.7
13.0
1.6
5.4
1.1
76.3

337
325
10
54
8
14
3
248

100.0
96.4
3.0
16.0
2.4
4.2
0.9
73.6

Note: Wave 1 = 30 days pre-release. R = respondent.
a
Percentage of all eligible cases, rather than percentage of eligible fielded cases (i.e., based on total number of
Wave 1 study-eligible completers).
b
Percentage of all eligible cases actually fielded.
Exhibit 23. Completed interviews by wave, all waves, and any waves by demographic group

Interview(s)
Wave 2
Wave 3
Wave 4
All 3 follow-ups
Any follow-up

Completed Interviews (% of Fielded Interviews)
Adult Males
Adult Females
Juvenile Males
58.0%
68.4%
70.0%
61.0%
70.9%
70.9%
65.6%
77.3%
73.6%
42.3%
54.9%
54.3%
79.3%
87.1%
87.2%

Note: Wave 2 = 3 months post-release; Wave 3 = 9 months post-release; Wave 4 = 15 months post-release.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Before turning to bias-related issues, the distribution of cases
across the study sites is briefly discussed. Exhibit 24 shows the
distribution of cases at each wave for the SVORI and nonSVORI groups. For the adult males, there are substantial
differences in the contribution of each site to the evaluation,
but overall the contribution of each site of SVORI and nonSVORI participants is roughly equal. The most prominent
exceptions (at Wave 1) are Iowa (114 SVORI, 55 non-SVORI),
Kansas (23 SVORI, 48 non-SVORI), and Nevada (107 SVORI,
50 non-SVORI).
The distribution looks quite different for the adult females. 25
There are fewer cases, and the distribution of SVORI to nonSVORI within site is substantially less even. Most notably, fully
101 of the non-SVORI adult female cases are from Indiana—
representing 49.5% of all non-SVORI subjects and 28.3% of all
cases. 26 Three of the 11 sites contributed fewer than 10 total
cases and, in two cases, Missouri and Pennsylvania, contributed
no non-SVORI subjects. The distribution of adult female
subjects across the two groups and 11 sites limited the ability
to address site in any outcome analyses.

25
26

The Maryland program was for adult males only.
The evaluation team was, of course, aware of the large number of
Indiana non-SVORI cases that were fielded and chose to include
them because of overall concern about the numbers of female
subjects that were being identified.

Case Flow and Threats to Validity

Exhibit 24. Completed interviews by wave, by demographic group and site

Wave 1
NonSVORI
SVORI

State
Adult Males
IA
IN
KS
MD
ME
MO
NV
OH
OK
PA
SC
WA
Total
Adult Females
IA
IN
KS
ME
MO
NV
OH
OK
PA
SC
WA
Total
Juvenile Males
CO
FL
KS
SC
Total

Wave 2
NonSVORI
SVORI

Wave 3
NonSVORI
SVORI

Wave 4
NonSVORI
SVORI

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

55
94
48
124
44
50
50
38
51
66
166
48
834

59
49
11
58
20
26
77
25
26
43
123
12
529

29
53
15
63
21
31
31
26
12
50
104
20
455

82
41
14
64
24
27
81
28
29
44
119
12
565

39
56
15
56
26
24
31
27
17
50
95
34
470

87
45
15
65
25
26
82
28
24
46
126
13
582

46
59
24
65
30
35
29
26
27
48
109
33
531

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

3
101
31
2
0
8
12
7
0
31
9
204

19
10
13
4
18
9
12
3
4
16
2
110

2
62
23
1
0
6
5
5
0
24
6
134

27
12
11
5
16
9
12
2
4
19
2
119

2
68
18
2
0
6
4
3
0
24
7
134

30
11
11
6
19
8
11
1
4
21
2
124

3
75
20
2
0
7
4
4
0
30
7
152

23
40
49
40
152

37
89
20
39
185

11
37
27
30
105

14
81
10
26
131

9
32
34
33
108

15
74
13
29
131

11
36
28
32
107

18
75
15
33
141

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

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

The distribution of cases among the four juvenile male sites
shows roughly equal numbers of cases from South Carolina,
more SVORI participants than non-SVORI from Kansas, and
more non-SVORI than SVORI participants from Colorado and
Florida.

SELECTION BIAS
Selection bias occurs when the process that generates
admission to the treatment group differs from that which leads
to inclusion in the control/comparison group. Under these
circumstances, the treatment (T) and comparison (C) groups
may differ on both observed and unobserved measures that
relate both to the likelihood of participating in the treatment
and the outcome(s) of interest. For example, if those entering
treatment are volunteers (i.e., motivated to participate) and
the comparison group is selected from those who did not
volunteer, any observed differences in outcomes may be due to
differences in motivation between the Ts and the Cs and not to
the treatment. Propensity score matching and multivariate
techniques can control for observed differences, while
instrumental variable approaches can control for differences in
unobserved measures (assuming appropriate instruments can
be identified). Random assignment theoretically takes care of
selection bias because assignment to T or C occurs after the
selection has taken place (or, more accurately, random
assignment assures that any observed differences are due to
chance and not to selection).
Two of the 16 programs (Iowa and Ohio adult programs)
included in the impact evaluation used random assignment to
assign eligible participants to the SVORI program or to
treatment as usual. For the remaining programs, the evaluation
team worked with each program to establish procedures for
identifying appropriate comparison groups, as was described
previously in the chapter Identifying Comparison Population
Pools and Evaluation Eligibles.

Although it is impossible to determine whether members of the
SVORI and non-SVORI groups differed on unobserved
variables, the Wave 1 interview contained extensive questions
related to the backgrounds of evaluation participants. These
data were used to assess the extent to which SVORI and nonSVORI differed on observed characteristics. The questions

Case Flow and Threats to Validity

related to (1) immutable characteristics (e.g., race, ethnicity,
date of birth); (2) pre-prison characteristics and behaviors
(e.g., drug and alcohol use); and (3) lifetime experiences (e.g.,
ever treated for substance abuse or mental health problems;
currently have a GED or high school diploma). To examine
issues related to selection bias, current measures that could
reflect differences that were due to—or potentially occurred
subsequent to—program assignment (e.g., have GED or high
school diploma) were distinguished from those that predated
assignment. Practically, this meant examining measures that
either were immutable or reflected values before the instant
incarceration, because there was no date for the non-SVORI
subjects that was comparable to the SVORI program enrollment
date.
SVORI and non-SVORI subjects were compared on a multitude
of variables. Few differences were observed between the two
adult male groups; greater numbers of differences were
observed between the two adult female groups and the two
juvenile male groups. Pre-release Characteristics and Service
Receipt among Adult Male Participants in the SVORI Multi-site
Evaluation (Lattimore et al. 2008; also see Lattimore, Steffey,
& Visher, 2009) presents a thorough comparison of the
characteristics of the adult male SVORI and non-SVORI
respondents; similar comparisons for the adult female and the
juvenile male subjects are presented in Lindquist et al. and
Hawkins et al. , respectively. Interested readers are directed to
those publications for full descriptions. Here, differences on key
variables are examined.

Exhibit 25 shows the t-statistics for comparisons between
SVORI and non-SVORI respondents for each demographic
group. For the adult males, there are statistically significant
differences for several variables, some of which have
traditionally been linked to criminal behavior. In particular,
those in SVORI programs were younger on average at the time
of the instant incarceration (26.1 years versus 27.1 years),
were more likely to be black (57% versus 50% black; 32%
versus 37% white), and less likely to have been employed
either in the 6 months before the current incarceration (64%
versus 68%) or ever (89% versus 92%)—although the latter
differences are relatively small. Although there were no
significant differences in self-reported drug use immediately
before the current incarceration, those in the non-SVORI group

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 25. t-statistics comparing means of SVORI and non-SVORI groups by demographic
group

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

Adult
Males
N = 1691
−2.57*
−2.30*
2.74*
−0.83
−0.12
−0.28
−2.04*
−2.21*
0.42
−0.52
0.61
0.80
0.43
−0.76
−1.92
−1.25
−2.09*
−2.59*
0.92
−1.35
2.36*
−2.58*
−3.18*
−0.47
−0.50
−0.25
0.86
−3.28*
−2.97*

Adult
Females
N = 357
−1.90
1.32
−1.78
0.59
−1.04
−0.07
−0.55
−1.31
0.67
−1.96
−0.03
−0.26
−0.66
−0.84
−0.59
0.30
−0.24
2.04*
−0.63
0.80
1.22
−1.37
1.74
−0.33
−1.18
−0.72
−1.31
−0.43
−0.58

Juvenile
Males
N = 337
−0.85
−2.31*
1.42
0.44
−0.70
0.19
−1.07
−1.47
0.15
−0.82
1.25
1.84
1.22
−0.38
0.94
−1.26
−0.20
−0.24
−0.21
−0.47
−2.24*
−2.64*
0.94
−1.66
−1.77
−1.46
−2.27
−1.36
NA

Note: NA = not applicable.
*p < 0.05, two-tailed test.

were more likely to report ever using cocaine (58% versus
53%) and heroin (23% versus 18%). SVORI respondents were
more likely than non-SVORI respondents to be serving time for
a drug crime (36% versus 31%), while non-SVORI respondents
were more likely to be serving time for a public order crime
(22% versus 17%). This last finding is consistent with nonSVORI respondents’ being more likely than SVORI participants
to report that they were currently incarcerated for a parole
violation (which was coded as a public order crime; 35% versus

Case Flow and Threats to Validity

27%) and for non-SVORI respondents to report more prior
prison incarcerations on average (1.33 versus 1.12). 27
There are fewer significant differences between the adult
female SVORI and non-SVORI participants. The only difference
among the variables included in the table is “ever used heroin,”
where 27% of the SVORI participants in comparison to 18% of
the non-SVORI participants were more likely to report heroin
use.
There also are relatively few differences among SVORI and
non-SVORI juvenile males. SVORI participants are less likely to
be white (14% versus 24%). Also, SVORI program participants
were less likely to be serving time for a drug crime (11%
versus 19%) or a public order crime (20% versus 32%).

NONRESPONSE
Nonresponse bias is potentially an issue when less than 100%
of a sample is interviewed and nonresponse is unlikely to be at
random. In those cases, there may be systematic differences
between those who are interviewed and those who are not. If
these differences are associated with treatment participation or
outcomes, estimates of treatment effect may be biased.
Wave 1 interviews were completed with 68.8% of all eligible
cases fielded (2,391 of 3,527). Refusal rates were relatively low
(370 of 3,527 or 10.5%). 28 The modal reason for no interview
was that the offender was released before he or she could be
interviewed (41% of all noninterviews), which was probably
equally likely to be true for SVORI and non-SVORI potential
respondents.

Note that respondents were asked to indicate all crimes for which
they were currently incarcerated, so an individual could have
reported serving time for, e.g., a violent crime and a parole
violation.
As shown in Exhibit 18, refusals were 295 of 2,564 eligible adult
males (11.5%), 38 of 516 eligible adult females (7.4%), and 37 of
447 eligible juvenile males (8.3%). Although adult male and female
non-SVORI comparisons were slightly more likely than SVORI
program participants to refuse the Wave 1 interview, the
differences were not large.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

ATTRITION
Attrition bias occurs when respondents who complete Wave 1
interviews do not complete follow-up interviews (similar to
nonresponse bias at subsequent waves). The primary validity
threat of attrition is that, if the people who complete follow-up
interviews are different from the people who do not complete
follow-ups, the population to whom findings can be generalized
is no longer known. This problem is particularly serious if there
is evidence that there is differential attrition from the study
groups that could be correlated with outcomes. All eligible
cases were fielded at each wave to maximize the likelihood that
at least one follow-up interview would be conducted with every
respondent.
Throughout the evaluation, various approaches were used to
investigate attrition. Differences between groups at successive
follow-up waves were similar to those observed at Wave 1,
which suggested that attrition was either random or similarly
affecting the SVORI and non-SVORI groups. Additionally,
balance between the SVORI and the non-SVORI subjects (for
each demographic group—i.e., adult males, adult females, and
juvenile males) was observed at each wave (see next chapter).
Once data collection was completed and a propensity model for
assignment to SVORI was estimated (see next chapter), the
issue of attrition bias was addressed from the perspective of
whether nonresponse was an issue when controlling for
selection into SVORI (see Imai, King, & Stuart, 2008; also
Fitzmaurice & Laird, 2000). The following equation was used to
examine the pattern of non-response for the adult male
subjects. The dependent variable was coded to indicate whether
a subject was a completer (completed all follow-up interviews),
an attriter (did not complete successive interviews), or missing
(had prior and after responses). 29
Y = β0 + β1*S + β1*p(R) + β1*S*p(R)

This approach comes from the clinical trials literature, where the
regimen of treatment is set and a subject can miss the last
treatment. Because the three data collection points in this study are
somewhat arbitrary compared with a clinical trial, alternative coding
schemes were applied for the dependent variable in this analysis;
results were similar.

Case Flow and Threats to Validity

where
S equals 1 if the individual is a SVORI participant, 0
otherwise;
p(R) is the estimated likelihood of having a follow-up for
each observation (based on a logistic regression
propensity score model); and
S*p(R) is an interaction term. If the main effect for SVORI
and the interaction term are not significant, the findings
suggest that there is not differential attrition conditional
on SVORI.
The response pattern is shown in Exhibit 26, where 0 =
response and 1 = nonresponse. The patterns show that three
follow-up interviews were obtained from 718 subjects (42%)
and no interviews from 351 subjects (21%). Two interviews
were obtained for 351 subjects (21%) and only one interview
for 277 subjects (16%).
Exhibit 26. Nonresponse pattern for three waves of follow-up interviews, adult males

Pattern
000
001
010
011
100
101
110
111

Classification
Completer
Attriter
Missing
Attriter
Missing
Attriter
Missing
Attriter

Frequency
718
78
88
100
185
55
122
351

Percentage
42.31
4.60
5.19
5.89
10.90
3.24
7.19
20.68

Exhibit 27 shows the results from the estimation of the
equation. 30 Once the main effect of response was controlled,
SVORI program participation was not related to whether a
response was obtained (i.e., none of the parameter estimates
for SVORI or the interaction term was statistically significant at
any usually accepted level). Given that the propensity score
weights generated good balance between the SVORI and nonSVORI groups on data at each wave for all three demographic
groups, the determination was made that it was not necessary

The model was run as a nonordered multinomial logistic regression
using SAS® 9.1.3.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 27. Results of nonordered multinomial logistic regression examining nonresponse,
adult males

Analysis of Maximum Likelihood Estimates
Parameter
Intercept
SVORI
Probability of no follow-up (pNOFU)
pNOFU*SVORI

Function
Number
1
2
1
2
1
2
1
2

Estimate
1.2299
0.4014
0.0412
−0.0453
−4.8335
−3.7349
−0.324
0.7615

Standard
Error
0.1266
0.1446
0.1266
0.1446
0.5368
0.6188
0.5368
0.6188

ChiSquare
94.38
7.7
0.11
0.1
81.09
36.43
0.36
1.51

Pr > Chi
Sq
<.0001
0.0055
0.745
0.7543
<.0001
<.0001
0.5461
0.2185

to control for nonresponse in addition to SVORI program
participation because no differential attrition was identified
between the two groups.

Propensity Score
Models
Propensity score models were used to address potential
selection bias due to the quasi-experimental design (see Rubin,
2006, for a collection of seminal papers in propensity score
modeling; see D’Agostino, 1998, for an accessible tutorial). 31
Propensity score models use observed characteristics to model
the likelihood that an individual with those characteristics will
be selected (or assigned) to the intervention. The approach is
to identify a set of parameters that are then used to estimate
the probability of assignment to the intervention for each
individual in a study. For example, logistic regression can be
31

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

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

used with a dichotomous indicator of intervention participation
as the dependent variable and individual characteristics as the
independent variables. Coefficients from this model are then
applied to the data to produce estimates of the probability of
ˆ ) are then used
assignment. These probabilities (p-hats or p
either (1) to stratify (or “bin”) subjects according to their
probabilities of receiving the intervention, (2) to weight
observations in subsequent analyses, or (3) to match subjects
ˆ in the
in the intervention group to subjects with similar p
comparison group. Initially, the assessment of outcomes was
conducted using the stratification or binning approach; but the
final outcome models were estimated using the weighting
approach, as it greatly simplifies the presentation of findings.32
Using the weighting approach allowed the estimation of one set
of outcome models for each demographic group. Presenting
findings by strata would have multiplied the number of models
and results to be presented by the number of strata. For
example, if adult male subjects were assigned to one of five
strata, differences in outcomes would have to be assessed
within each stratum, increasing the number of models by a
factor of five.
The success of the propensity score model estimation is judged
by the effectiveness of the strata or weights to reduce
differences between the treatment and control groups on
observed characteristics or, in the common terminology, to
achieve balance between the two groups. Two ways of checking
for balance are (1) to examine t-statistics comparing group
means or (2) to examine standardized differences between the
two groups (see, e.g., Rosenbaum & Rubin, 1985). Both
approaches were used in this study.
The propensity score approach is useful only if it produces
ˆ between groups. The goal is to
adequate overlap in the p
develop scores that, for example, can be used to sort
individuals into strata where the probability of assignment to
the intervention is similar. Once individuals are assigned to
ˆ , the strata should contain individuals
strata based on their p

Preliminary results showed that population average treatment
effects estimated by combining results from the analyses based on
strata for the adult male groups were nearly identical to those
derived from the weighted models—as would be expected. Results
were also similar for the adult female and juvenile males groups;
those results are not presented here.

Propensity Score Models

from both groups—otherwise there is no comparison between
groups.
Item missingness was relatively rare in the data, but
imputation procedures were employed so that no observations
had to be dropped from the outcome analyses because of
missing p-hats. Logit models to generate the probability of
assignment to SVORI [p(SVORI) or p(S)] were estimated within
the framework of SAS 9.1.3 PROC MI and PROC MIANALYZE for
each of the three demographic groups (adult males, adult
females, and juvenile males). These SAS procedures
accommodated item missingness by imputing values for
missing data. A two-step imputation procedure was used within
PROC MI in which (1) a Monte Carlo procedure (MCMC) was
employed to impute values until the data set reached a pattern
of monotone missingness and then (2) regression was
employed to impute the remaining values (Allison, 2001; SAS
Institute, 2004). As noted earlier, the independent variables for
the propensity score models included only variables that
reflected the values of measures before program assignment
(effectively preincarceration).
Exhibit 28 lists the variables that were included in the
propensity score models. Also shown are the numbers of
missing values for each of the included variables. Two
variables—arrest rate and conviction rate—had the highest
missing rate because respondents failed to report the numbers
of prior arrests and convictions.
The adult male sample included 1,697 observations, 1,500
(88.4%) of which had no missing values on any of the
variables. A MCMC procedure in SAS 9.1.3 was used to impute
values until monotone missingness was achieved in the data
set. For the adult male sample, monotone missingness was
achieved once imputations were generated for all variables
except three—Juvie, Convict_rate, and Arrest_rate. Regression
was then used to generate values for arrest rate only (4.5% of

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 28. Variables included in the SVORI propensity score models and numbers of
observations with missing values

Number of Observations with
Missing Values
Variable
Name
SVORI
Age_inc
Race_white

Variable Description
= 1 if SVORI program participant; 0 otherwise
Age (years) at incarceration
= 1 if self-report race/ethnicity is white only; 0
otherwise (black only is reference category)
Race_other
= 1 if self-report race/ethnicity is other than white
only or black only; 0 otherwise (black only is
reference category)
Nohome
= 1 if homeless/shelter/no set place to live before
incarceration; 0 otherwise
Employed
= 1 if employed during the 6 months before
incarceration; 0 otherwise
Steady_rel
= 1 if report in steady relationship in the 6 months
before incarceration; 0 otherwise
Prior_school
= 1 if regularly attending school before
incarceration
Parent_relation Quality of relationship with parents scale
Fam_support
Family instrumental support scale
Antisoc_fam
Family deviance = 1 if anyone in family ever
convicted & anyone in family ever incarcerated &
anyone in family ever had alcohol and other drug
(substance use) problems (all three questions = 1)
Antisoc_peer
Peer deviance = 1 if any friends ever convicted &
any friends ever incarcerated & any friends ever
had substance use problems (all three questions =
1)
AOD_tx_prior
= 1 if received substance use treatment before
incarceration; 0 otherwise
MH_tx_prior
= 1 if reported receiving mental health treatment
before current incarceration; 0 otherwise
Victim_prior
= 1 if experienced victimization before
incarceration; 0 otherwise
Victim_score
Preincarceration victimization scale
Perpetration
= 1 if any perpetration of violence 6 months before
incarceration; 0 otherwise
Perp_score
Preincarceration perpetration scale
ALC_30
= 1 if self-report drank alcohol in the 30 days
before incarceration; 0 otherwise
MJ_30
= 1 if self-report used marijuana in the 30 days
before incarceration; 0 otherwise
Otherdrug_30
= 1 if used drugs other than marijuana 30 days
before incarceration; 0 otherwise
Person
= 1 if report incarcerated for a person/violent
crime; 0 otherwise
Property
= 1 if report incarcerated for a property crime; 0
otherwise

Adult
Males
N = 1697
0
0
3

Adult
Females
N = 357
0
0
0

Juvenile
Males
N = 337
0
NA
0

NA
NA
NA

9
4
4

NA
0

0
NA

NA
4

NA
2

1
2

(continued)

Propensity Score Models

Exhibit 28. Variables included in the SVORI propensity score models and numbers of
observations with missing values (continued)

Number of Observations with
Missing Values
Variable
Name
Drug
Public_Other
VO_Parole
Age_first
Arrests
Arrest_rate
Convictions
Convict_rate
Inc_rate
Juvie

Variable Description
= 1 if report incarcerated for a drug crime; 0
otherwise
= 1 if report incarcerated for a public order crime or
crime other than person/property/drug; 0 otherwise
= 1 if currently serving for parole violation; 0
otherwise
Age (years) at first arrest
Times arrested
Arrest rate (number of arrests/age at time of
incarceration)
Times convicted
Conviction rate (number of convictions/age at time
of incarceration)
Incarceration rate (number of incarcerations/age at
time of incarceration)
Times locked up in juvenile detention

Adult
Males
N = 1697
9

Adult
Females
N = 357
1

Juvenile
Males
N = 337
2

12
NA
113

10
NA
28

NA
19
NA

NA
39

NA
24

10
NA

Note: NA = the variable was not included in the respective propensity model.

the observations), arrest and conviction rates (1.53% of
observations), and arrest and conviction rates plus number of
times in juvenile detention (0.41% of observations).33
The adult female sample included 357 observations, 293
(82.1%) of which had no missing values on any of the
variables. For the adult female sample, monotone missingness
was achieved once imputations were generated for all variables

Site indicators were not included because these variables can only
capture the likelihood of SVORI in each site compared to the
reference site. For the adult male samples, this was roughly 50:50
SVORI:non-SVORI (52% versus 48%). The variation from 50:50 was
much greater in the adult female and juvenile male samples. In fact,
because this was not an experiment with 50:50 random assignment,
the actual (and relative) numbers of SVORI and non-SVORI cases in
each site are meaningless, so site is of no use in this analysis of
likelihood of assignment to SVORI. Additionally, because some sites
diverged from the 50:50 assignment, indicators for site were quite
influential. Thus, for the adult male model, the coefficients on site in
preliminary models were quite large—i.e., they had a substantial
impact on the values of the p-hats. One result of this was that the
distribution of sites across the bins was quite skewed. To the extent
that site is related to outcomes, site was acting as a confounder and,
thus, the site indicator variables were not included in the SVORI
propensity score model.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

except four—Arrest_rate, Convict_rate, Incar_rate, and Juvie.
Regression was then used to generate values for arrest rate
only (3.64% of the observations), arrest and conviction rates
(3.36% of observations), and arrest and conviction rates plus
number of times in juvenile detention (0.56% of observations),
and arrest, conviction, number of times in juvenile detention,
and incarceration rate (0.28% of observations).
The juvenile male sample included 337 observations, 292
(86.6%%) of which had no missing values on any of the
variables. For the juvenile male sample, monotone missingness
was achieved once imputations were generated for all variables
except three—number of times in a juvenile facility (Juvie),
number of convictions (Convict), and the number of arrests
(Arrests). Regression was then used to generate values for
number of arrests only (5.04% of the observations), number of
arrests and juvenile detentions (0.30% of observations), and
number of arrests, convictions, and times in juvenile detention
(0.30% of observations).
The MI and MI ANALYZE procedures generated five data sets
for each of the three samples—each data set included different
estimates for the missing values—then used logistic regression
to generate parameter estimates for each data set. Generally,
the five sets of parameter estimates are used to produce a
single set of parameters that are then reported to reflect the
impact of the independent variables on the dependent, outcome
variable. For the propensity score model, however, the
parameter values themselves are not of interest—what is of
interest are the ^
p ’s that are generated by applying the
parameters to the data. The imputation procedures generated
five p-hat values for each individual; the five values were then
averaged to generate the final p-hats that were used in the
outcome analyses. 34
As noted, separate models were estimated for each of the three
demographic groups. The results for the adult male sample are
presented below. The results for the adult female and juvenile

Note that a review of the literature identified no applications in
which missing value imputation was used in the production of
propensity scores. After a thorough discussion among the analysis
team and the expert panel, it was determined that the average of
ˆ estimates was an appropriate value to use.
the five p

Propensity Score Models

male samples follow. Diagnostics (e.g., variation for individual
p-hat values) are presented in the discussion below.

PROPENSITY SCORE MODEL DEVELOPMENT
FOR ADULT MALE SUBJECTS
The imputation procedures described above generated five sets
of parameters, based on five sets of imputed missing values.
The five sets of coefficient values for the adult male propensity
score model are shown in Exhibit 29. Because the range of
parameter estimates is small for all of the variables included in
the model, the difference between the minimum and maximum
p-hat values is also small for most observations. The difference
between the minimum and maximum p-hat values ranged from
0.0008 to 0.1271, with a mean of 0.0120 (standard deviation =
0.0120) and median of 0.0080.35
The final model results derived from the five sets of estimates
are shown in Exhibit 30. These estimates are the traditional
output that in an outcome analysis would be examined to
determine the relationship between the independent and
dependent variables. They are of less interest to a propensity
score analysis because they are simply a means to an end—the
^
p.
The validity of the model is assessed by examining the extent
to which the use of the propensity scores achieves balance
between the SVORI and the non-SVORI groups. These analyses
are discussed below. But first, the distributions of the
propensity scores are described. Then, the extent to which the
propensity scores improved balance is addressed.

Extreme values in the maximum difference between estimated were
relatively rare. The 99th, 95th, 90th, and 75th percentiles were
0.063, 0.032, 0.023 and 0.012, respectively.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 29. Parameter estimates for the SVORI propensity score models for the adult male
sample generated using PROC MI

Variable

Imputed
Value 1

Imputed
Value 2

Imputed
Value 3

Imputed
Value 4

Imputed
Value 5

Intercept

0.7250

0.7137

0.6902

0.6905

0.6859

Age_inc

−0.0163

−0.0167

−0.0165

−0.0160

−0.0155

Race_white

−0.2406

−0.2446

−0.2321

−0.2421

−0.2419

Race_other

−0.2769

−0.2722

−0.2733

−0.2666

−0.2747

Nohome

−0.0065

−0.0177

−0.0127

−0.0176

−0.0133

Employed

−0.1562

−0.1425

−0.1614

−0.1537

−0.1556

Steady_rel

−0.0461

−0.0514

−0.0383

−0.0483

−0.0297

0.2086

0.2101

0.2121

0.2099

0.2124

MH_tx_prior

AOD_tx_prior

−0.0008

−0.0055

−0.0080

−0.0042

−0.0111

Victim_prior

0.0314

0.0413

0.0337

0.0272

0.0370

Perpetration

0.0014

−0.0075

0.0053

0.0041

−0.0049

Alc_30

0.1412

0.1486

0.1463

0.1445

0.1337

MJ_30

−0.1776

−0.1783

−0.1692

−0.1736

−0.1676

Otherdrug_30

−0.1324

−0.1373

−0.1428

−0.1322

−0.1414

0.1011

0.0902

0.0955

0.1209

0.1039

Person
Property

0.0618

0.0419

0.0580

0.0774

0.0667

Drug

0.3030

0.3094

0.3182

0.3256

0.3015

Public_other

−0.1996

−0.2126

−0.1873

−0.1978

−0.2139

VO_Parole

−0.2962

−0.2975

−0.3027

−0.2963

−0.2940

0.0005

0.0011

0.0020

0.0008

0.0010

−0.1079

−0.1430

−0.1589

−0.1781

−0.1389

Convict_rate

0.1179

0.2594

0.1711

0.2077

0.2038

Juvie

0.0134

0.0139

0.0189

0.0202

0.0180

Age_first
Arrest_rate

Inc_rate
–2 Log Likelihood

−2.1535

−2.1106

−2.0227

−2.1101

−2.2783

−1150.7462

−1150.2227

−1150.4433

−1150.0810

−1150.0189

Propensity Score Models

Exhibit 30. Final propensity model for adult male sample
Parameter

Estimate

95%
LCL

95%
UCL

Min

Max

P−value

Intercept

0.7011

0.3424

0.0299

1.3722

431194.1

0.6859

0.7250

2.0474

0.0406

Age_inc

−0.0162

0.0081

−0.0322

−0.0002

309459.5

−0.0167

−0.0155

−1.9884

0.0468

Race_white

−0.2403

0.1243

−0.4839

0.0034

1284434.7

−0.2446

−0.2321

−1.9329

0.0533

Race_other

−0.2727

0.1610

−0.5883

0.0428

8660040.5

−0.2769

−0.2666

−1.6941

0.0902

Nohome

−0.0136

0.1541

−0.3156

0.2884

3526365.0

−0.0177

−0.0065

−0.0880

0.9299

Employed

−0.1539

0.1104

−0.3702

0.0625

173700.3

−0.1614

−0.1425

−1.3938

0.1634

Steady_rel

−0.0427

0.1089

−0.2562

0.1707

66904.5

−0.0514

−0.0297

−0.3926

0.6946

0.2106

0.1093

−0.0036

0.4248

58084057.9

0.2086

0.2124

1.9272

0.0540

MH_tx_prior

−0.0059

0.1224

−0.2458

0.2340

2722605.6

−0.0111

−0.0008

−0.0483

0.9615

Victim_prior

0.0341

0.1228

−0.2066

0.2749

760027.8

0.0272

0.0413

0.2777

0.7812

Perpetration

−0.0003

0.1310

−0.2570

0.2564

805724.8

−0.0075

0.0053

−0.0023

0.9981

Alc_30

0.1429

0.1149

−0.0823

0.3680

426821.8

0.1337

0.1486

1.2439

0.2135

MJ_30

−0.1733

0.1111

−0.3909

0.0444

789717.5

−0.1783

−0.1676

−1.5602

0.1187

Otherdrug_30

−0.1372

0.1133

−0.3594

0.0849

781385.5

−0.1428

−0.1322

−1.2111

0.2259

Person

0.1023

0.1473

−0.1864

0.3910

70942.1

0.0902

0.1209

0.6946

0.4873

Property

0.0612

0.1470

−0.2270

0.3493

45301.2

0.0419

0.0774

0.4159

0.6775

Drug

0.3115

0.1513

0.0149

0.6082

131339.2

0.3015

0.3256

2.0583

0.0396

Public_other

−0.2022

0.1400

−0.4767

0.0722

70017.6

−0.2139

−0.1873

−1.4444

0.1486

VO_Parole

−0.2973

0.1170

−0.5267

−0.0680

4628965.5

−0.3027

−0.2940

−2.5408

0.0111

AOD_tx_prior

Age_first

0.0011

0.0124

−0.0232

0.0254

586496.0

0.0005

0.0020

0.0883

0.9297

−0.1453

0.1486

−0.4368

0.1461

2962.6

−0.1781

−0.1079

−0.9778

0.3283

Convict_rate

0.1920

0.2763

−0.3498

0.7338

2201.0

0.1179

0.2594

0.6949

0.4872

Juvie

0.0169

0.0194

−0.0212

0.0549

4555.8

0.0134

0.0202

0.8683

0.3853

−2.1350

0.9302

−3.9582

−0.3119

27621.7

−2.2783

−2.0227

−2.2953

0.0217

Arrest_rate

Inc_rate

Note: DF = degrees of freedom. LCL = lower confidence limit. Max = maximum value. Min = minimum value. UCL =
upper confidence limit.

ˆ are shown in Exhibit 31. As can
The distributional findings for p
ˆ
be seen, p ranges from a low of 0.1806 to a high of 0.7412.
The means of the distributions of the SVORI and non-SVORI
groups are similar—0.5232 for SVORI and 0.4934 for nonSVORI. Not surprisingly, there is considerable overlap, as
demonstrated by the box plots shown in Exhibit 32.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

ˆ distributions for all adult males, adult male SVORI
Exhibit 31. Characteristics of p
participants, and non-SVORI comparisons

Characteristic
N
^ mean
p
^ standard deviation
p
^ minimum
p
^ maximum
p

Exhibit 32. Boxplot of
p-hat distributions for
SVORI (SVORI = 1) and
non- SVORI (SVORI = 0)
adult males

All

SVORI

Non-SVORI

1,697
0.5085
0.0862
0.1806
0.7412

863
0.5232
0.0823
0.1933
0.7412

834
0.4934
0.0876
0.1806
0.7020

Propensity Score Models

ˆ to assign subjects to five
The initial investigations used the p
ˆ
equal probability strata or bins across the observed p
distribution [0.1806, 0.7412]. These investigations included
assessing the extent to which the propensity score weights
resulted in balance between the SVORI and non-SVORI groups
across waves of data collection and initial investigations of
outcomes. Those analyses are not discussed here because,
subsequently, a more parsimonious approach to the data was
taken in which weights generated from the propensity scores
were applied to the data. The discussion here focuses on those
methods.
Balance

The purpose of the propensity score matching is to achieve
greater comparability between treatment and comparison
groups. Two ways of checking for balance are to examine tstatistics comparing SVORI and non-SVORI means or to
examine standardized differences (see, e.g., Rosenbaum &
Rubin, 1985). Results for both of these approaches are shown
below.
Exhibit 33 shows the t-statistics for the comparison of
unweighted means between the SVORI and Non-SVORI for the
variables included in the propensity model. The exhibit shows
that there were significant differences among means (α = 0.05;
two-tailed t-test) for 8 of the 24 variables for the adult male
sample. Specifically, SVORI participants were younger at
incarceration than the Non-SVORI, less likely to be white (more
likely to be black), less likely to have been employed during the
6 months before incarceration, more likely to have a current
conviction for a drug offense and less likely for a public order
offense, less likely to be serving time for a parole violation, and
have a lower prior incarceration rate.
Rosenbaum and Rubin (1985) described checking for balance
using standardized differences as well as t-statistics (p. 34).
The standardized difference between two means is defined as
( x1 − x 2 )
( s12 + s22 ) 2

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 33. t-statistics and standardized differences from the comparison of mean values
(unweighted) for SVORI to Non-SVORI for selected variables from the Wave 1 interview

Variable
Age_inc
Race_white
Race_black (Reference category)
Race_other
Nohome
Employed
Steady_rel
AOD_tx_prior
MH_tx_prior
Victim_prior
Perpetration
Alc_30
MJ_30
Otherdrug_30
Person
Property
Drug
Public_other
VO_Parole
Age_first
Arrest_rate
Conv_rate
Juvie
Inc_rate

T statistic
−2.57*
−2.30*
2.74*
−0.83
−0.12
−2.04*
−0.28
0.42
−0.52
0.61
0.80
0.43
−0.76
−1.92
0.92
−1.35
2.36*
−2.58*
−3.18*
−0.47
−0.50
−0.25
0.86
−3.28*

Standardized Difference
−12.49
−11.18
13.31
−4.01
−0.60
−9.93
−1.37
2.04
−2.52
2.98
3.86
2.10
−3.72
−9.33
4.47
−6.56
11.48
−12.58
−15.48
−2.27
−2.51
−1.23
4.22
−15.98

*p < 0.05.

Exhibit 33 also shows the standardized differences for the
variables included in the propensity score model. Interestingly,
0 of 24 values are greater than 0.20. This suggests good
balance between the two groups even before correcting for
differences. Exhibit 34 summarizes the results of the balance
checks using t-statistics and standardized differences. The
results use the same breakpoints for the t-statistic that
Rosenbaum and Rubin used—although more logical breakpoints
that correspond to p-value breakpoints (e.g., 1.64, 1.96)
perhaps would be more meaningful, since the samples are large
enough that these values are meaningful.

Propensity Score Models

Exhibit 34. Summary of
balance checks using
two-sample t-statistics
and percentage of
standardized differences

Number of Values
(of 24 total)
14
2
6
2
14
3
6
1
0
0

Variable
0 ≤ |t| < 1
1 ≤ |t| < 2
2 ≤ |t| < 3
|t| ≥ 3
0% ≤ |std. diff.| < 5%
5% ≤ |std. diff.| < 10%
10% ≤ |std. diff.| < 15%
15% ≤ |std. diff.| < 20%
20% ≤ |std. diff.| < 25%
|std. diff.| ≥ 25%

The propensity scores were used to develop weights to examine
the population average treatment effect (PATE) for the outcome
models. The PATE is the average treatment effect one would
expect if the population were treated. 36 The PATE weights were
calculated as follows:
If subject i was a SVORI participant,
wi =

1
ˆ
pi

or else
wi =

1
.
ˆi
1 − p

The goal of the balance check is to determine whether
knowledge of the value of an independent variable results in
better prediction of participation in SVORI. PROC Survey
Logistic in SAS® 9.1 was used to regress the SVORI indicator
on each of the variables that were included in the model.
Exhibit 35 shows the results for the Wave 1 data. As can be
seen, the Wald chi square test statistics are effectively zero for
all of the significance tests, suggesting that the propensity
score weights generated good balance for the data. Balance
results for Waves 2, 3, and 4 data sets are shown in Exhibit 36.

In contrast, the average treatment effect on the treated (ATET)
provides an estimate of the effect of treatment on the treated.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 35. Balance checks for Wave 1 data based on propensity score weighted regression
of the SVORI indicator on each of the variables (adult males)

Variable
Age_inc
Race_white
Race_black
Race_other
Nohome
Employed
Steady_rel
AOD_tx_prior
MH_tx_prior
Victim_prior
Perpetration
Alc_30
MJ_30
Otherdrug_30
Person
Property
Drug
Public_other
VO_parole
Age_first
Arrest_rate
Convict_rate
Juvie
Inc_rate

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

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

Wald
Chi Sq
0.000
0.001
0.000
0.000
0.000
0.000
0.000
0.005
0.001
0.004
0.000
0.004
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.005
0.005
0.000
0.008

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

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

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

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

Note: CL = confidence limit.

The following sections address propensity model development
for the adult female and the juvenile male samples.

MODEL DEVELOPMENT FOR ADULT FEMALE
SUBJECTS
The data included 357 adult female subjects distributed across
11 sites and the two evaluation groups. As noted earlier, the
distribution of adult female subjects across site and group was
not proportional—some sites had no non-SVORI subjects and
one site contributed nearly 30% of the non-SVORI comparison
cases.

Propensity Score Models

Exhibit 36. Balance checks for Waves 2, 3, and 4 data based on propensity score weighted
regression of the variable on a SVORI indicator (adult males)

Wave 2
Variable
Age_inc
Race_white
Race_black
Race_other
Nohome
Employed
Steady_rel
AOD_tx_prior
MH_tx_prior
Victim_prior
Perpetration
Alc_30
MJ_30
Otherdrug_30
Person
Property
Drug
Public_other
VO_parole
Age_first
Arrest_rate
Convict_rate
Juvie
Inc_rate

Estimate
−0.0077
−0.0216
−0.0674
0.2032
0.2444
0.1230
0.0551
−0.1035
−0.2263
−0.0834
−0.1371
0.1614
0.1636
0.0519
−0.0868
0.0240
0.0075
0.1715
−0.1356
−0.0095
0.0110
0.2251
0.0000
0.1353

Prob
Chi Sq
0.3774
0.8742
0.6047
0.3011
0.2186
0.3750
0.6976
0.4295
0.1291
0.5292
0.3277
0.2443
0.2085
0.6911
0.5076
0.8740
0.9576
0.2910
0.3824
0.4373
0.9394
0.4043
1.0000
0.9073

Wave 3
Estimate
−0.0098
−0.1396
0.0830
0.1043
0.0389
−0.0486
−0.0814
−0.0983
−0.1220
−0.1305
−0.0641
−0.0491
0.0370
0.0516
−0.1843
0.1162
0.1310
0.0567
0.1351
−0.0142
0.0442
0.0042
−0.0146
−0.3591

Prob
Chi Sq
0.2564
0.2907
0.5139
0.5826
0.8417
0.7228
0.5567
0.4437
0.4051
0.3140
0.6379
0.7159
0.7710
0.6852
0.1497
0.4317
0.3429
0.7170
0.3614
0.2373
0.7470
0.9864
0.4580
0.7458

Wave 4
Estimate
−0.0084
0.0177
−0.0373
0.0486
−0.1163
−0.0195
−0.0044
0.0155
0.0066
−0.0332
0.0561
−0.0459
−0.0005
−0.0360
−0.1030
0.0772
0.0490
0.0588
−0.0771
−0.0015
−0.0643
−0.0091
−0.0095
−1.8569

Prob
Chi Sq
0.3200
0.8883
0.7601
0.7928
0.5261
0.8816
0.9732
0.8999
0.9623
0.7894
0.6688
0.7232
0.9964
0.7680
0.4053
0.5775
0.7109
0.6952
0.5834
0.8983
0.6287
0.9706
0.6268
0.0688

The propensity score model for the adult females included the
same variables as the model that was estimated for adult male
ˆ values
subjects. The same procedures were employed. The p
ˆ values estimated
were calculated as the average of the five p
37
from the five sets of parameter estimates. The final
parameter estimates generated by the propensity score models
are shown in Exhibit 37. The Min and Max columns show the
minimum and maximum parameter estimates from the five
models.

ˆ values;
The parameter estimates provided relatively consistent p
although, with a smaller sample of adult female subjects, the
variation around the estimates was somewhat greater than was
observed with the adult male sample. Extreme values in the
maximum difference between estimated were relatively rare. The
99th, 95th, 90th, and 75th percentiles were 0.075, 0.054, 0.034
and 0.022, respectively.
89

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 37. Final propensity model for adult female sample
Parameter

Estimate
Intercept
0.9669
−0.0344
Age_inc
Race_white
0.5749
Race_other
0.4677
Nohome
−0.3825
Employed
−0.2432
Steady_rel
0.0219
AOD_tx_prior
0.2533
MH_tx_prior
−0.5821
Victim_prior
0.0740
Perpetration
−0.0753
Alc_30
0.0030
MJ_30
−0.2244
Otherdrug_30 −0.1526
Person
−0.1307
Property
0.2204
Drug
0.1743
Public_other
−0.1953
VO_Parole
0.6231
Age_first
−0.0049
Arrest_rate
−0.0355
Convict_rate
−0.5161
Juvie
−0.0370
Inc_rate
0.6707

SE
0.8385
0.0196
0.2714
0.3456
0.2990
0.2390
0.2584
0.2508
0.2433
0.2738
0.2827
0.2419
0.2595
0.2673
0.3564
0.3231
0.3388
0.3183
0.2912
0.0242
0.4977
0.7145
0.0448
1.8228

95%
LCL
−0.6766
−0.0727
0.0430
−0.2096
−0.9685
−0.7116
−0.4847
−0.2382
−1.0589
−0.4627
−0.6295
−0.4712
−0.7330
−0.6764
−0.8292
−0.4128
−0.4896
−0.8191
0.0523
−0.0523
−1.0111
−1.9165
−0.1248
−2.9026

95%
UCL
DF
2.6103
444489.69
0.0040
80554.24
1.1068 1132924340.99
1.1450
1947009.29
0.2035
3255633.17
0.2252
66223368.27
0.5284
44963961.71
0.7449
7173216.44
−0.1053
7483436.49
0.6107
6264585.26
0.4788 822117962.48
0.4772
38685.11
0.2842
33003.91
0.3713
3129608.15
0.5677
647426.90
0.8537
2057745.60
0.8383
180730.25
0.4285
707334.73
1.1939
36479351.68
0.0426
10781.52
0.9400
124896.52
0.8843
23406.51
0.0508
12116.40
4.2440
5678.60

Min
0.9042
−0.0364
0.5716
0.4532
−0.3950
−0.2472
0.0183
0.2479
−0.5883
0.0657
−0.0785
−0.0339
−0.2634
−0.1582
−0.1524
0.2095
0.1411
−0.2133
0.6176
−0.0085
−0.0850
−0.5967
−0.0456
0.3090

Max
1.0037
−0.0325
0.5763
0.4801
−0.3717
−0.2397
0.0282
0.2609
−0.5735
0.0822
−0.0729
0.0207
−0.2043
−0.1384
−0.1071
0.2354
0.1965
−0.1790
0.6286
−0.0002
0.0105
−0.4368
−0.0324
1.0549

t
P value
1.1531
0.2489
−1.7549
0.0793
2.1183
0.0341
1.3534
0.1759
−1.2794
0.2007
−1.0176
0.3089
0.0846
0.9326
1.0100
0.3125
−2.3928
0.0167
0.2701
0.7871
−0.2665
0.7899
0.0125
0.9900
−0.8648
0.3872
−0.5709
0.5681
−0.3668
0.7138
0.6823
0.4950
0.5146
0.6068
−0.6137
0.5394
2.1397
0.0324
−0.2008
0.8409
−0.0714
0.9431
−0.7223
0.4701
−0.8254
0.4092
0.3680
0.7129

Note: LCL = lower confidence limit; UCL = upper confidence interval.

ˆ are shown in Exhibit 38. As can
The distributional findings for p
ˆ
be seen, p ranges from a low of 0.0816 to a high of 0.8206.
The means of the distributions of the SVORI and non-SVORI
groups are 0.4715 for SVORI and 0.3964 for non-SVORI adult
females. There is considerable overlap between the two
distributions, as demonstrated by the box plots shown in
Exhibit 39.

The results of the model with respect to balance are shown in
Exhibit 40. The results suggest that the weights generated by
the propensity scores model generated balance across all four
waves of interview data.

Propensity Score Models

Exhibit 38.
ˆ
Characteristics of p
distributions for all adult
females, adult female
SVORI participants, and
non-SVORI comparisons

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

All

SVORI

Non-SVORI

357
0.4286
0.1358
0.0816
0.8206

153
0.4715
0.1316
0.1789
0.7627

204
0.3964
0.1303
0.0816
0.8206

Exhibit 39. Boxplot of
p-hat distributions for
SVORI (SVORI = 1) and
non- SVORI (SVORI = 0)
adult females

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit 40. Balance checks for Waves 1 through 4 adult female data, based on PATEweighted regressions
Wave 1
N = 357
Variable

Age_inc
Race_white
Race_black
Race_other
Nohome
Employed
Steady_rel
AOD_tx_prior
MH_tx_prior
Victim_prior
Perpetration
Alc_30
MJ_30
Otherdrug_30
Person
Property
Drug
Public_other
VO_Parole
Age_first
Arrest_rate
Convict_rate
Juvie
Inc_rate

Wave 2
N = 244

Wave 3
N = 253

Wave 4
N = 276

Estimate

Prob Chi
Sqa

Estimate

Prob Chi
Sqa

Estimate

Prob Chi
Sqa

Estimate

Prob Chi
Sqa

−0.0064
−0.0522
0.0257
0.0483
−0.0703
−0.0566
0.0327
−0.0088
0.0029
0.0021
0.0049
0.0454
0.0535
−0.0109
0.0810
−0.0464
−0.0632
−0.0088
−0.0066
−0.0015
−0.0326
0.1016
−0.0031
−0.0480

0.6847
0.8170
0.9108
0.8749
0.7990
0.8015
0.8950
0.9689
0.9897
0.9928
0.9835
0.8406
0.8147
0.9621
0.7439
0.8390
0.7976
0.9741
0.9813
0.9364
0.9259
0.8575
0.9365
0.9761

−0.0231
−0.2887
0.2538
0.0575
0.0752
0.0614
0.1006
−0.0544
0.1380
0.2664
0.2332
0.0221
−0.0034
−0.0201
0.1394
−0.2306
−0.0675
−0.0276
0.0137
−0.0052
−0.0354
0.2004
0.0290
4.7014

0.2435
0.2893
0.3516
0.8771
0.8230
0.8211
0.7541
0.8415
0.6095
0.3434
0.4195
0.9355
0.9903
0.9418
0.6261
0.4001
0.8263
0.9364
0.9697
0.8211
0.9398
0.8101
0.6366
0.0404

−0.0154
−0.0295
0.0080
0.0407
0.4219
−0.0542
0.0359
−0.0858
0.1515
0.3153
0.1983
0.0416
0.0271
−0.0949
0.1608
−0.3551
0.1430
−0.0222
−0.0100
0.0030
−0.1426
−0.1422
0.0083
0.5751

0.4144
0.9119
0.9765
0.9127
0.2074
0.8383
0.9067
0.7472
0.5667
0.2507
0.4808
0.8761
0.9201
0.7289
0.5677
0.1844
0.6407
0.9477
0.9776
0.8946
0.7236
0.8253
0.8684
0.7986

−0.0101
0.0171
0.0771
−0.1782
0.2004
0.0519
−0.0392
−0.0129
0.0114
0.1395
0.1214
0.1665
0.0579
−0.1129
0.1496
−0.1394
0.1456
−0.0778
0.1818
0.0030
0.0044
0.0982
0.0125
−0.0156

0.5745
0.9465
0.7649
0.6219
0.5215
0.8382
0.8908
0.9597
0.9642
0.5917
0.6487
0.5129
0.8219
0.6651
0.5845
0.5862
0.6111
0.8071
0.6093
0.8917
0.9911
0.8734
0.7827
0.9944

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.
a
p value of Wald Chi-square statistic tests that the parameter is equal to zero.

MODEL DEVELOPMENT FOR JUVENILE MALE
SUBJECTS
The set of variables used to estimate the propensity score
model for juvenile male subjects was similar to that used for
the adult samples. Exhibit 41 shows the final logistic regression
model results for the probability that a juvenile male was
assigned to a SVORI program in one of the four juvenile SVORI
sites. Again, these values reflect the average of the five
estimates from the PROC MI and MI ANALYZE procedures. As
can be seen from the Min and Max columns, there was
relatively little variability in the estimates produced by the five
imputation models.

Propensity Score Models

Exhibit 41. Final propensity model for juvenile male sample
Parameter
Intercept
Age_inc
Race_white
Race_other
Prior_school
Parent_relation
Fam_support
Antisoc_fam
Antisoc_peer
AOD_tx_prior
MH_tx_prior
Victim_score
Perp_score
Alc_30
MJ_30
Otherdrug_30
Person
Property
Drug
Public_other
VO_Parole
Arrests
Convictions
Juvie

Estimate
−0.3933
0.0159
−1.4279
−0.1922
0.5327
0.0214
−0.0012
0.1356
0.2634
0.5193
−0.1983
−0.0419
0.0106
0.5276
−0.4643
0.9188
−0.4157
−0.3289
−0.8970
−0.5894
0.3996
−0.0023
−0.0233
−0.0777

SE
95% LCL
1.8503 −4.0198
0.1005 −0.1812
0.3863 −2.1851
0.3053 −0.7906
0.2648
0.0137
0.0350 −0.0471
0.0363 −0.0724
0.2617 −0.3774
0.2805 −0.2864
0.3092 −0.0867
0.2937 −0.7740
0.0283 −0.0973
0.0237 −0.0358
0.3086 −0.0772
0.3149 −1.0815
0.3704
0.1928
0.2869 −0.9779
0.2735 −0.8650
0.3675 −1.6173
0.3104 −1.1978
0.2629 −0.1157
0.0352 −0.0715
0.0554 −0.1320
0.0531 −0.1818

95%
UCL
DF
3.2332 30218941.15
0.2129 3204594.27
−0.6707
192290.80
0.4063 1700262.98
1.0517
79119.70
0.0899
106715.03
0.0700
43137.80
0.6486
12178.01
0.8132
211525.56
1.1253 1055328.35
0.3773
474985.65
0.0134
477306.91
0.0569
257171.67
1.1324
93705.13
0.1529
13225.05
1.6448
894213.70
0.1465
477122.23
0.2072
101329.07
−0.1767
292780.31
0.0190
53395.45
0.9149 4560968.15
0.0668
1704.04
0.0854
2771.65
0.0263
260374.43

Min
−0.4362
0.0121
−1.4548
−0.2069
0.5025
0.0178
−0.0043
0.0841
0.2491
0.5099
−0.2221
−0.0438
0.0089
0.5008
−0.5194
0.9011
−0.4315
−0.3505
−0.9171
−0.6189
0.3901
−0.0147
−0.0372
−0.0826

Max
−0.3471
0.0206
−1.3956
−0.1805
0.5589
0.0240
0.0042
0.1686
0.2868
0.5399
−0.1855
−0.0402
0.0120
0.5563
−0.4175
0.9324
−0.3930
−0.2974
−0.8689
−0.5554
0.4079
0.0029
−0.0140
−0.0742

t
P−value
−0.2126
0.8317
0.1578
0.8746
−3.6960
0.0002
−0.6294
0.5291
2.0119
0.0442
0.6137
0.5394
−0.0324
0.9741
0.5183
0.6043
0.9390
0.3478
1.6795
0.0931
−0.6753
0.4995
−1.4844
0.1377
0.4463
0.6554
1.7097
0.0873
−1.4744
0.1404
2.4806
0.0131
−1.4492
0.1473
−1.2024
0.2292
−2.4408
0.0147
−1.8987
0.0576
1.5200
0.1285
−0.0664
0.9471
−0.4210
0.6738
−1.4639
0.1432

Note: LCL = lower confidence limit; UCL = upper confidence limit.

ˆ values were calculated as the average of the
As before, the p
ˆ
five p values estimated from the five sets of parameter
ˆ values for all
estimates. 38 The resulting distributions of p
subjects and for each group are shown in Exhibit 42. Exhibit 43
shows the boxplots comparing SVORI and non-SVORI
distributions, where it can be seen that there is reasonable
overlap between the two distributions.

ˆ values;
The parameter estimates provided relatively consistent p
although, with a smaller sample of juvenile male subjects, the
variation around the estimates was somewhat greater than was
observed with the adult male sample. Extreme values in the
maximum difference between estimated were relatively rare. The
99th, 95th, 90th, and 75th percentiles were 0.17, 0.07, 0.05 and
0.03, respectively.
93

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

ˆ distributions for all juvenile males, juvenile male SVORI
Exhibit 42. Characteristics of p
participants, and juvenile male non-SVORI comparisons

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

All

SVORI

Non-SVORI

337
0.4510
0.1810
0.0701
0.8578

152
0.5236
0.1733
0.1005
0.8578

185
0.3914
0.1651
0.0701
0.7975

Exhibit 43. Boxplot of
p-hat distributions for
SVORI (SVORI = 1) and
non-SVORI (SVORI = 0)
juvenile males

The balance check results based on the weighted regressions of
these variables on the SVORI indicator are shown in Exhibit 44.
The results suggest that the propensity score model provided
balance across all four waves of interview data.

Propensity Score Models

Exhibit 44. Balance checks for Waves 1 through 4 juvenile male data, based on PATEweighted regressions

Variable
Age_inc
Race_white
Race_black
Race_other
Prior_school
Parent_relation
Fam_support
Antisoc_fam
Antisoc_peer
AOD_tx_prior
MH_tx_prior
Victim_score
Perp_score
Alc_30
MJ_30
Otherdrug_30
Person
Property
Drug
Public_other
VO_Parole
Arrests
Convictions
Juvie

Wave 1
N = 337
Prob Chi
Estimate
Sqa
−0.0369
0.1139
−0.1145
0.7065
−0.0362
0.8814
0.1365
0.6276
0.0212
0.9306
0.0031
0.9111
−0.0118
0.6864
−0.0617
0.7992
−0.0087
0.9712
−0.0308
0.9073
−0.0921
0.7301
−0.0047
0.8066
−0.0019
0.9082
−0.1143
0.6348
0.0318
0.8964
0.0417
0.8876
−0.0490
0.8402
−0.0285
0.9062
0.0913
0.8051
0.0619
0.8274
0.0546
0.8311
−0.0132
0.6092
−0.0115
0.8016
0.0038
0.9363

Wave 2
N = 236
Prob Chi
Estimate
Sqa
−0.0525
0.6374
−0.4218
0.2661
−0.0476
0.8701
0.4515
0.1969
−0.1921
0.5055
−0.0011
0.9755
−0.0270
0.4231
−0.2749
0.3423
0.2198
0.4471
−0.1310
0.6870
−0.1987
0.5242
0.0144
0.5364
0.0226
0.2576
−0.1806
0.5272
0.0326
0.9109
0.3306
0.3503
−0.2332
0.4215
0.2342
0.4159
−0.2029
0.6347
0.1091
0.7450
−0.3594
0.2256
−0.0238
0.4179
−0.0301
0.5696
−0.0073
0.8892

Wave 3
N = 239
Prob Chi
Estimate
Sqa
0.0395
0.7092
−0.1212
0.7327
−0.2975
0.3083
0.5356
0.1382
−0.2882
0.3205
0.0088
0.7957
−0.0281
0.4132
−0.0209
0.9427
−0.0332
0.9087
0.0245
0.9384
−0.1714
0.5793
−0.0072
0.7674
−0.0024
0.9004
−0.0601
0.8342
0.0524
0.8568
0.2862
0.4240
−0.2049
0.4807
0.0451
0.8763
0.0204
0.9620
−0.1726
0.6447
−0.1443
0.6332
−0.0469
0.1247
−0.0529
0.3354
0.0139
0.8052

Wave 4
N = 248
Prob Chi
Estimate
Sqa
0.0158
0.8842
−0.2996
0.4332
−0.0841
0.7724
0.3581
0.3014
−0.2387
0.3998
−0.0088
0.7859
0.0161
0.6411
−0.4210
0.1353
−0.1337
0.6374
−0.2150
0.5007
−0.0581
0.8493
−0.0202
0.3975
−0.0123
0.4989
0.0279
0.9216
−0.1712
0.5496
0.0275
0.9357
−0.1281
0.6521
−0.1356
0.6338
−0.1643
0.7236
−0.0652
0.8477
−0.1908
0.5347
−0.0257
0.4000
−0.0407
0.4390
0.0379
0.4915

Conclusions and
Next Steps
This report has documented the procedures used to identify and
recruit evaluation sites and subjects for the Multi-site
Evaluation of SVORI. In addition, the investigation of potential
sources of bias and adjustments to the data using propensity
score models were discussed.
ˆ generated by the model for each of the demographic
The p
groups resulted in good balance between the two study groups
and good overlap among the two study groups’ propensity
score distribution.

Evaluation findings are presented in the following separate
reports:

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

Lattimore, P. K., Brumbaugh, S., Visher, C. A.,
Lindquist, C., Winterfield, L., Salas, M., et al. (2004).
National portrait of SVORI. Research Triangle Park, NC:
RTI International.

Lattimore, P. K., Visher, C. A., & Steffey, D. M. (2008).
Pre-release characteristics and service receipt among
adult male participants in the SVORI Multi-site
Evaluation. Research Triangle Park, NC: RTI
International.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

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

References
Allison, P. D. (2001). Missing data. Sage university papers
series on quantitative applications in the social sciences,
No. 07-136. Thousand Oaks, CA: Sage Publications.
Cowell, A., Roman, J., & Lattimore, P. K. (2009). An economic
evaluation of the Serious and Violent Offender Reentry
Initiative. Research Triangle Park, NC: RTI International.
D'Agostino, R. B., Jr. (1998). Tutorial in biostatistics:
Propensity score methods for bias reduction in the
comparison of a treatment to a non-randomized control
group. Statistics in Medicine, 17, 2265-2281.
Fitzmaurice, G. M., & Laird, N. M. (2000). Generalized linear
mixture models for handling nonignorable dropouts in
longitudinal studies. Biostatistics, 1, 141-156.
Glaze, L. E., & Palla, S. (2005). Probation and parole in the
United States, 2004. Washington, DC: U.S. Bureau of
Justice Statistics.
Glynn, R. J., Schneeweiss, S., & Sturmer, T. (2006). Indications
for propensity scores and review of their use in
pharmacoepidemiology. Basic Clinical Pharmacology and
Toxicology, 98(3), 253-259.
Harrison, P. M., & Beck, A. J. (2005). Prison and jail inmates at
midyear 2004 (NCJ 195189). Washington, DC: U.S.
Bureau of Justice Statistics.
Hawkins, S., Dawes, D., Lattimore, P. K., & Visher, C. A.
(2009). Detainee reentry experiences of juvenile males:
Characteristics, service receipt, and outcomes of
participants in the SVORI Multi-site Evaluation. Research
Triangle Park, NC: RTI International.
Imai, K., King, G., & Stuart, E. A. (2008). Misunderstandings
between experimentalists and observationalists about
causal inference. Journal of the Royal Statistical Society,
A (Part 2), 171, 481-502.
Lattimore, P. K. (2007). The challenges of reentry. Corrections
Today, 69(2), 88-91.
Lattimore, P. K., Brumbaugh, S., Visher, C., Lindquist, C. H.,
Winterfield, L., Salas, M., et al. (2004). National portrait
of the Serious and Violent Offender Reentry Initiative.
Research Triangle Park, NC: RTI International.

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Lattimore, P. K., Steffey, D. M., & Visher, C. A. (2009). Prisoner
reentry experiences of adult males: Characteristics,
service receipt, and outcomes of participants in the
SVORI Multi-site Evaluation. Research Triangle Park, NC:
RTI International.
Lattimore, P. K., & Visher, C. A. (2009). The Multi-site
Evaluation of SVORI: Summary and synthesis. Research
Triangle Park: RTI International.
Lattimore, P. K., Visher, C. A., & Steffey, D. M. (2008). Prerelease characteristics and service receipt among adult
male participants in the SVORI Multi-site Evaluation.
Research Triangle Park, NC: RTI International.
Lattimore, P. K., Visher, C. A., Winterfield, L., Lindquist, C., &
Brumbaugh, S. (2005). Implementation of prisoner
reentry programs: Findings from the Serious and Violent
Offender Reentry Initiative Multi-site Evaluation. Justice
Research and Policy, 7(2), 87-109.
Lindquist, C. (2005). Reentry research in action:
Implementation of SVORI programs. Research Triangle
Park, NC: RTI International.
Lindquist, C., & Winterfield, L. (2005). Reentry research in
action: Characteristics of prisoner reentry programs.
Research Triangle Park, NC: RTI International.
Lindquist, C. H., Barrick, K., Lattimore, P. K., & Visher, C. A.
(2009). Prisoner reentry experiences of adult females:
Characteristics, service receipt, and outcomes of
participants in the SVORI Multi-site Evaluation. Research
Triangle Park, NC: RTI International.
National Research Council (2007). Parole, desistance from
crime, and community integration. Washington, DC:
National Academy Press.
National Research Council and Institute of Medicine (2001).
Juvenile crime, juvenile justice. Washington, DC:
National Academy Press.
Re-entry Policy Council (2005). Report of the Re-entry Policy
Council: Charting the safe and successful return of
prisoners to the community. New York: Council of State
Governments.
Rosenbaum, P., & Rubin, D. B. (1983). The central role of the
propensity score in observational studies for causal
effects. Biometrika, 70, 41-55.
Rosenbaum, P., & Rubin, D. B. (1985). The bias due to
incomplete matching. Biometrics, 41, 103-116.
Rubin, D. B. (1997). Estimating causal effects from large data
sets using propensity scores. Annals of Internal
Medicine, 127, 757-763.
Rubin, D. B. (2006). Matched sampling for causal effects. New
York: Cambridge University Press.
SAS Institute (2004). SAS/STAT 9.1 user’s guide. Cary, NC:
Author.
Visher, C. A. (2007). Returning home: Emerging findings and
policy lessons about prisoner reentry. Federal
Sentencing Reporter, 20(2), 93-102.

100

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.

101

Appendix A. Implementation
Evaluation Protocols and
Data Collection Instruments

A-1

SERIOUS AND VIOLENT OFFENDER REENTRY INITIATIVE
WORKPLAN REVIEW FORM
1.

Grantee Information

State:

Grant Number:

Grant Agency:
Authorized Signatory (AS):

AS Title:

AS E-mail:

AS Phone No.:

AS Fax No.:

AS City:

AS State:

AS Zip:

Project Director (PD):

PD Title:

PD E-mail:

PD Phone No.:

PD Fax No.:

PD City:

PD State:

PD Zip:

Point of Contact (PC):

PC Title.:

PC E-mail:

AS Phone No.:

AS Fax No.:

PC City:

PC State:

PC Zip:

Total Award Request:

Total Funded Amount:

Requested Substance Training Funds:

Actual Substance Training Funds:

Requested Mental Health Funds:

Actual Mental Health Funds:

Requested Juvenile Justice Funds:

Actual Juvenile Justice Funds:

Requested Adult Justice Funds:

Actual Adult Justice Funds:

Requested Labor Funds:

Actual Labor Funds:

AS Address:

PD Address:

PC Address:

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 1

Federal Grant Manager Information
Federal Grant Manager (FGM):

FGM Phone No.:

FGM E-mail:

FGM Fax No.:

FGM Address:
FGM City:

FGM State:

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

FGM Zip:

February 15, 2003; page 2

Problem to Be Addressed/ Target Population

Serving Post Adjudicatory Serious and Violent Offenders: (circle response)
Target Population: (circle response)

Juveniles

Yes
Adults

No
Both

No. of Eligible Offenders (annually):
Description of High Risk Characteristics of Population:

Risk Assessment Being Used with Current Population: (Prior to reentry initiative)
Services Received by Non-participating Eligible Offenders:
Inducements to Participate in Reentry Initiative: (circle response)

Yes

Description of Inducements:

Age Ranges Being Served:
No. of Offenders to Be Served Year 1:

No. of Offenders to Be Served Year 2:

No. of Offenders to Be Served Year 3:

No. of Offenders to Be Served Overall:

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 3

Organizational Capacity/Decision Makers

Leveraging Resources:
Plan for Leveraging State, Local, and Tribal Resources to Ensure Sustainability:

Memorandum of Understandings (MOUs):
Areas for Collaboration

Existing MOUs (Y/N)

No. of MOUs

No. of State Partners

No. of Local Gov’t. Partners

No. of Other Partners

Labor
Substance Abuse
Housing
Education
Mental Health
Faith-based
Medical Services
Case Management
Offender Supervision
Support Services

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 4

Phase One (Institutional)

Comprehensive Needs Assessment Tool:
Incentives for Participation: (circle response)

Yes

Description of Incentives:

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 5

Programs in the Institution:
Program Areas

Existing
Programs
(Y/N)

List of Agencies

Federal
Programs
(Y/N)

State
Programs
(Y/N)

Local
Programs
(Y/N)

Other
Programs
(Y/N)

Comments

Employment

Substance Abuse

Housing

Education

Mental Health

Faith-based

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 6

Program Areas

Existing
Programs
(Y/N)

List of Agencies

Federal
Programs
(Y/N)

State
Programs
(Y/N)

Local
Programs
(Y/N)

Other
Programs
(Y/N)

Comments

Medical Services

Case
Management

Support Services

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 7

Phase Two (Community Reentry)

Name of Reentry Authority:
Needs Assessment Tool(s):
Existing Continuity of Services Plan from Phase One to Phase Two: (circle response)

Yes

Description of Continuity of Services Plan:

Existing Continuity of Supervision Plan from Phase One to Phase Two: (circle response)
Description of Continuity of Supervision Plan:

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 8

Transition Programs:
Program Areas

Existing
Programs
(Y/N)

List of Agencies

Federal
Programs
(Y/N)

State
Programs
(Y/N)

Local
Programs
(Y/N)

Other
Programs
(Y/N)

Comments

Employment

Substance Abuse

Housing

Education

Mental Health

Faith-based

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 9

Program Areas

Existing
Programs
(Y/N)

List of Agencies

Federal
Programs
(Y/N)

State
Programs
(Y/N)

Local
Programs
(Y/N)

Other
Programs
(Y/N)

Comments

Medical Services

Case
Management

Support Services

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 10

Phase Three (Stabilization)

Existing Continuity of Services Plan: (circle response)

Yes

Description of Continuity of Services Plan:

Existing Oversight and Case Management Plan: (circle response)

Yes

Description of Oversight and Case Management Plan:

Exit Assessment Tool(s):

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 11

Community Programs (for Stabilization):
Program Areas

Existing
Programs
(Y/N)

List of Agencies

Federal
Programs
(Y/N)

State
Programs
(Y/N)

Local
Programs
(Y/N)

Other
Programs
(Y/N)

Comments

Employment

Substance Abuse

Housing

Education

Mental Health

Faith-based

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 12

Program Areas

Existing
Programs
(Y/N)

List of Agencies

Federal
Programs
(Y/N)

State
Programs
(Y/N)

Local
Programs
(Y/N)

Other
Programs
(Y/N)

Comments

Medical Services

Case
Management

Support Services

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 13

Project Management

Existing Management Information System (MIS):

Yes

Description of MIS:

Plan for Communication with Partnering Agencies:
Description of Communication Plan:

K:\CIMS\NTTAC\RE FUNDS REENTRY\REENTRY FORMS AND TEMPLATES\WORKPLAN FORMS & TEMPLATES\WORKPLAN REVIEW FORM.DOC

February 15, 2003; page 14

1
PROTOCOL FOR SVORI WORKPLAN REVIEWS AND
INITIAL TELEPHONE INTERVIEWS

Background
This protocol provides guidelines for the SVORI workplan reviews (which will be conducted in
two stages: a “quick scan” and a “comprehensive review”) and the initial telephone interviews
with the site’s SVORI project director. The purpose of the workplan reviews and telephone
interviews is to collect information from the SVORI sites that will enable us to produce a
“portrait” of all 68 sites and to select the 15-20 sites in which we will conduct our intensive
evaluation (i.e., the offender longitudinal survey and detailed implementation assessment).
Prior to beginning the workplan reviews and telephone interviews, staff should participate in the
SVORI database training and review the following background materials/websites:
• The SVORI site solicitation
• Information about the SVORI initiative (on OJP’s Reentry website at
http://www.ojp.usdoj.gov/reentry/ )
• The SVORI national evaluation solicitation:
http://www.ncjrs.org/pdffiles1/nij/sl000578.pdf
• RTI/UI’s proposal
• The RTI/UI SVORI national evaluation internal database (http://svori.rti.org/secure/)
• The “Site information” in the database for the sites for which you will be gathering
information.
• Other background materials about issues in reentry (Travis, Solomon, Waul. From
Prison to Home: the Dimensions and Consequences of Prisoner Reentry; Lynch and
Sabol. Prisoner Reentry in Perspective; Travis. But They All Come Back: Rethinking
Prisoner Reentry.)

Defining “SVORI Programs” within a site
The remainder of this protocol describes the programmatic information we need for the SVORI
sites. Because some of the 68 sites appear to be operating more than one program (i.e., with
different project directors or program coordinators, sets of services or programs, or special
populations), we would like to have the programmatic information collected for each program
(we are using the term “SVORI program”). Keep in mind that our use of the term “SVORI
program” does not refer to individual stand-alone programs such as an in-prison therapeutic
community or a program such as Alcoholics Anonymous (each site will have multiple individual
programs), but rather a set of programs delivered to a specific population and coordinated by a
program coordinator or project director. For example, a site may be running separate programs
for juveniles and adults (or other special populations), each of which would be classified as a
separate “SVORI program” (for purposes of the national evaluation). Geographic boundaries
should only be used to define “SVORI programs” if the site is truly operating distinct,
independent programs (i.e., with different program coordinators, sets of services or programs,
geographical parameters or special populations) in multiple geographic regions. We expect
many sites to cover several counties or cities in their program, but these geographical regions
will not be considered as distinct “SVORI programs” unless they are clearly operated

2
independently (with different program coordinators or project directors and unique sets of
services or programs). Further guidelines determining whether a site operates single or multiple
programs based on their workplan descriptions include:
• Single Program: The workplan contains only one description of what is being done in
phase 1 (the institutional phase), phase 2 (the initial community phase) and phase 3 (the
final reintegration). The site MAY include multiple age/gender groups (including
juveniles and adults; males and females) and therefore involve multiple pre-release
institutions within which the 'program' is working. The site MAY also include multiple
counties to which offenders are returning; but there is only a single 'unidimensional'
approach that is presented.
• Multiple Programs: The site provides either
o 1) one workplan describing different 'programs' for each phase, based on a
specific population (e.g., adults vs. juveniles). So, the programmatic approach
varies, based on the population, for phase 1, phase 2, and phase 3. This instance
should be set up as two different programs.
o 2) distinct workplans for different programs (for example, the Michigan Family
Independence Agency submitted 5 different workplans, 1 each for a 'juvenile'
intervention in a single county, resulting in a total of 5 different programs). This
instance should be set up as five different programs.
Our default position is to recognize only one “SVORI program” per site unless there is
compelling evidence otherwise. As described in more detail subsequently, you will use the “site
information” section of the database to identify the SVORI program for each site (even if the site
only has one “SVORI program”).
General Guidelines
This section provides general guidelines regarding completing the workplan reviews (both the
“quick scan” and “comprehensive review”) and telephone interviews. It is extremely important
that the information obtained during these activities be gathered in a consistent manner across
sites and entered into the database in a standardized manner. The information to be gathered
from the workplan reviews and telephone interviews is the same (the telephone interviews
should be used to confirm that information included in the workplan is still accurate and obtain
missing information). In situations in which you encounter inconsistent information (i.e., when
confirming something in the telephone interview, you learn that a particular piece of information
provided in the workplan is outdated), please enter the most recent information in the database.
If substantial changes between what was described in the workplan and what is described during
the telephone interview regarding the site’s SVORI population, enrollment numbers, or services
provided are evident, please use the database field called “Programmatic Changes” (an openended text field) to succinctly describe such modifications.
As you enter information into the database, be sure to save the information frequently. The
screen will “time out” after 40 minutes and you will lose any unsaved information. Also, for any
topics for which you were not able to obtain information (either in the workplan reviews or
telephone interviews), please check the boxes for “unknown” so we know that you attempted to
obtain information for that topic but could not find it. For items that are not clearly outlined in
the workplan but for which you think you can infer that the correct answer is “no”, please use

3
“unknown” instead. Assume that information not explicitly provided is unknown until verified
through the telephone interview.
Workplan Reviews
Each site was required to submit a workplan outlining the key aspects of its reentry initiative (in
essence, a revised application). Staff from Caliber Associates (the technical assistance provider)
have reviewed and scored these workplans, creating a brief “workplan summary on SVORI
website” for each site. The “site assignment and workplan status” excel file posted on the project
website indicates the status of both the workplan and workplan summary for each site. The
“workplan summary available” column lists whether the workplan summary is available (if
available, open the document called “workplan summaries” from the national evaluation website
and print up only the summary for your site – all of the workplans are compiled and saved as one
large file so be sure not to print up the whole thing). The “workplan location” column lists the
location of the workplan for each site (the SVORI national evaluation website, the OJP SVORI
website, or hard copy only). If the workplan is available on the national evaluation website,
simply download it. If the workplan is available on the OJP website, you will go to the “State
Activities and Resources” section of this website (http://www.ojp.usdoj.gov/reentry/), click on
the state, then click on the green hyperlink for “workplan” to download it. If the workplan is
only available in hard copy, RTI staff should get the copy from Chris Lindquist and UI staff
should get the copy from Meghan Salas. For some sites, the workplan is either “on the way” or
not yet available in any form. In these cases, please begin with the other sites assigned to you
(and we will continue following up with Caliber to get the workplans as soon as possible).
“Quick Scan”. The first step in reviewing workplans will involve conducting a “quick scan” of
the sites’ workplans and workplan summaries (if available). During the “quick scan”, staff
should only focus on obtaining the “SVORI Population” and “Enrollment Information” items
described in detail below (under “Topics to be Covered”) from the workplans and entering this
information in the project database. This step should begin immediately after the debriefing on
7/07 and be completed by COB Tuesday, July 15th (all tasks and timelines are posted on the
website in a document called “Site Liaison Calendar”). Please only spend a short amount of time
on this task (around 30 minutes for the review and 1 hour for the database entry for each site)
and do not worry if you cannot find every piece of information in the “SVORI Population” and
“Enrollment Information” forms. On 7/16, the management team will order the sites in terms of
priority and share this information with you so that you can then concentrate on completing the
comprehensive reviews for the priority sites first. Please note that if you finish the “quick scan”
early and want to proceed with the comprehensive review, please do so. We will be giving some
priority to the sites that are likely to have the largest number of participants so if you begin the
comprehensive review before receiving our priority order, you may want to start with the larger
sites that you have identified.
Comprehensive Review. Based on the priority order the management team establishes, please
conduct the comprehensive review (i.e., cover all items in all 4 forms described below) of the
workplans. This step will involve going through the workplans in detail and trying to get as
much of the relevant information as possible and entering this in the database. In this stage, you
might also find that you will need to update some of the information you entered as a result of
the “quick scan”. This step should be completed by COB Wednesday, July 30th. When you are

4
finished entering the workplan information into the database, please enter a completion date for
the workplan review in the “site information” form in the database.
Telephone Interviews
Once the comprehensive workplan reviews are well underway, we would like you to begin
conducting telephone interviews with the sites’ SVORI project directors. Once again, please
complete this activity based on the priority order we have established. You should not wait until
you have completed all comprehensive workplan reviews before moving onto this step (it might
make sense to do, say, three workplan reviews and then conduct the telephone interviews for
these three sites) but you do need to wait until July 21st to begin this step (because we will send a
generic “lead letter” on 7/17 from Pam and Christy to the PD’s to inform them of the upcoming
calls1). We would like to have the telephone interviews completed for all sites by COB Friday,
August 15th.
The telephone interviews will be used to confirm the accuracy of any information obtained from
the workplan reviews and to obtain any missing information (i.e., information that was not
available in the workplan reviews). Staff should contact the individual flagged as “site contact”
in the project database and set up an interview time. For sites that have multiple “SVORI
programs” with distinct project directors or program coordinators, you will need to talk with
each contact about his/her “SVORI program”.
When setting up the interview, staff should cover the following points (most of which will be
covered in the lead letter from Pam and Christy):
• You are calling from [RTI/UI]. We are conducting the national evaluation of the “Going
Home: Serious and Violent Offender Reentry Initiative”. This evaluation is being
supported by the federal agencies that have funded their sites.
• This evaluation is a separate activity from the technical assistance provided by Caliber
Associates
• We would like to collect some additional information regarding SVORI activities from
the key contact at each site. This information will enable us to produce a description of
all 68 sites that will be used in a document to be submitted to the federal government next
spring.
• In addition, we are beginning the process of selecting a subset of 15-20 sites in which we
will conduct an intensive evaluation. This intensive impact evaluation will include a
longitudinal study of offenders and a detailed implementation assessment.
• We have already reviewed their revised work plan and (if applicable), the information
they provided to Caliber in their phone interview, in order not to duplicate efforts.
• Confirm that the person is the most knowledgeable person about the site’s SVORI
program. If not, please obtain the name and contact information for the appropriate site
person and enter this in the database. This new contact should then be called.
• The telephone interview will take approximately 30 minutes (be sure to schedule a
specific date and time that you can call the site contact for the interview).

By this time, the site should also have received a lead letter from NIJ introducing RTI and UI as the national
evaluators for SVORI.

5
During this initial conversation, please go ahead and inquire about two topics:
• Whether they have received their SVORI funds yet and if they have received the funds,
the date they received them)
o Please enter this information in the “site information” section of the database, in
the field for “date funding received”. If the site has not received their funds,
please check the “not yet received” box. If the site has received funds, please
enter in the date that the funds were received.
• Which of the implementation activites they have completed
o For this item, please follow the protocol for item 25
After you set up the interview, please:
• Fax or e-mail the site contact a memo summarizing the background information provided
in the bulleted list above, providing your contact information, and listing the date/time of
the scheduled interview (this memo is posted on the project website in the “documents”
section, but you’ll need to customize it for your site contact)
• Accompanying the memo, please also send the site contact a “site information report”
generated from the database. This report will list the topics to be covered in the interview
and any information already gathered for the site.
• Try to go to the state’s Department of Corrections website and obtain a list of all
correctional institutions in the state (this will be useful information to have when
discussing item 8).
When you call back to actually conduct the interview, please go through each of the topics in the
“Topics to Cover” section below, making sure to obtain all of the information we need. The
information gathered during the telephone interview should be entered in the database as soon as
possible following the interview. When you are finished entering the telephone interview
information into the database, please enter a completion date for the telephone interview in the
“site information” form in the database.

Topics to Cover
The topics below should be covered during both the workplan review (and review of any other
available materials from Caliber) and telephone interviews. The topics are listed in the order in
which they appear in the database, but it is fine to deviate from this order (if it helps to keep the
flow going) during the telephone interviews. In the interviews, you may also encounter
situations in which you ask a particular question but receive an answer to a different question
(which may or may not be a topic that we need), so please try to become as familiar with the
topics as possible. This way, you will be able to record their answer for the appropriate topic and
make sure that we have the “right” answers for all topics.
Another point related to the ordering of the items is that in the telephone interviews it may be
helpful to begin with item 25 (regarding the current stage of implementation) so you will know
whether to phrase your questions to ask about what is already happening (i.e., “Is your site doing
…”) or what they plan on happening (i.e., “Does your site plan to do…” or “what is your
anticipated monthly enrollment”). The “potential questions/probes” listed below use generic
wording to cover as many scenarios as possible.

In addition, the wording of “potential questions/probes” assumes that we did not get any
information about the topic during the workplan review (and review of Caliber materials).
Obviously, if we do have some existing information, you will only want to confirm what we
already have and will modify the wording accordingly (i.e., “According to your workplan, your
site will be focusing on offenders between the ages of 18-25. Is this still accurate?”)
Please also note that the sites may be using SVORI funds to simply fill existing service gaps or
expand existing programs, rather than implementing a new program. Because it would be
extremely difficult for sites to isolate SVORI-funded activities and because even an isolated
component funded by SVORI funds is likely to ultimately affect a site’s entire approach to
reentry, we are interested in knowing about each site’s overall reentry approach since SVORI
funds were awarded, not just components specifically funded by SVORI.
“SVORI Programs” Within the Site
The first thing you will need to find out is whether the site is operating more than one “SVORI
program” (see detailed discussion on “Defining ‘SVORI Programs’ Within a Site” above).
1.
•

•

“SVORI Programs”
Description: We would like to know whether the site appears to be operating more than
one “SVORI program” (see description on pages 1-2), because the remaining topics to be
covered in the interview will need to be covered separately for each “SVORI program”
(you will probably want to focus on one program at a time, going through each question
with that program in mind, and then go through the questions for the next program).
Potential questions/probes: “Is your site operating one SVORI program, or are you using
your grant money to operate distinct SVORI programs? By “SVORI program” I mean
distinct sets of programs serving different target populations, having distinct sets of
services, and managed by distinct project directors or program coordinators”. (As you ask
this question, keep in mind that if the state has more than one SVORI site, the project
director may assume that that is what you are referring to by “program”, so be sure to
clarify this by saying “I realize that your state has X grantees, but I am asking
specifically about the grant awarded to X agency.”)
Database field format/special instructions: In the “site information” section of the
database, select the box that says “add new SVORI program for this site” to set up a
program (or multiple programs) and enter a succinct name (50 characters or less) for each
“SVORI program”. Most sites should have a specific name for their initiative/program
that is separate from their site name but if they do not, then just use the site name. You
must enter a “SVORI program” even if the site only has one program. If the site has
more than one “SVORI program”, please obtain the remaining information separately for
each program.

SVORI Population and Eligibility Criteria
This set of items refers to the population on which the site will focus in their SVORI program.
2.

Population Type

7
•

•
•

Description: We would like to know if the site is focusing on adults, juveniles, or
youthful offenders (or a combination of these three population types). “Youthful
offenders” only refers to individuals sentenced under the Youthful Offender statutes
(which can cover those up to age 25 in some states). Be sure not to confuse the term
“young adults” (which some sites may have used in their workplans because it is a
separate category specified in the SVORI solicitation) with “youthful offenders”. We
do not need to separate young adults from adults.
Potential questions/probes: “Is your site including adult offenders, juvenile offenders,
or youthful offenders?”
Database field format/special instructions: check-box (adults, juveniles, youthful
offenders, unknown); select each category that applies

Age Range
• Description: We would like to know the age range (in years) of the site’s SVORI
population. Keep in mind that in some cases the concordance between the population
type and age range may not be evident (for example, a site may say they are focusing
on “adults” but their age range may include those aged 14 and up; this is accurate if
they are including juveniles sentenced as adults).
• Potential questions/probes: “What age range of offenders are you including in your
site?”
• Database field format/special instructions: use the two drop-down boxes (“low
range” and “high range”) to denote the lower age limit and upper age limit; if the site
does not have a lower and/or upper range, select “no lower/upper limit”; if you cannot
obtain the age range, select “unknown”

Gender
• Description: We would like to know whether the site includes males, females, or both
• Potential questions/probes: “Does your SVORI population include male offenders,
female offenders, or both?”
• Database field format/special instructions: check-box (males, females, unknown);
select each category that applies

Offense Type
• Description: We would like to know if the site focuses on offenders who committed
specific offenses.
• Potential questions/probes: “Is your site focusing on offenders with specific offense
types?”
• Database field format/special instructions: check-box (offense type not used as an
identifying factor, sex offenses, other violent offenses, drug offenses, unknown);
select all categories that apply; if necessary, use the “other” open-ended text field to
enter a succinct description of an offense type that is not included in the pre-existing
categories
o Note: Only use the check-boxes when a site gives an affirmative statement in
its workplan regarding targeting particular offense types. If there are no such
statements but the site says that offenders will be assessed with a risk
assessment that will include crime type, then check “none” for “offense type”,

8
“other” for “other inclusion criteria” and write “to be assessed with
assessment instrument (under the “other inclusion criteria” item). If there are
no such statement made and no mention made of any type of risk assessment,
check “unknown” for “offense type”.
6.

Criminal History
• Description: We would like to know if the site focuses on offenders based on
criminal history (i.e., first time serious violent offenders, repeat offenders, etc.).
Criminal history covers previous arrests, previous incarcerations (either jail or
prison), previous probation terms, etc.
• Potential questions/probes: “Is your site focusing on offenders based on any criteria
related to criminal history” (probe: “for example, are you specifically trying to reach
populations such as repeat offenders or first-time prisoners?”)
• Database field format/special instructions: check-box (criminal history not used as an
identifying factor, first time offenders, repeat offenders, unknown); select all
categories that apply; if necessary, use the “other” open-ended text field to enter a
succinct description for criminal history that is not included in the pre-existing
categories
o Note: Only use the check-boxes when a site gives an affirmative statement in
its workplan regarding targeting offenders with particular criminal histories. If
there are no such statements but the site says that offenders will be assessed
with a risk assessment that will include criminal history, then check “none”
for “criminal history”, “other” for “other inclusion criteria” and write “to be
assessed with assessment instrument (under the “other inclusion criteria”
item). If there are no such statement made and no mention made of any type
of risk assessment, check “unknown” for “criminal history”.

Special Populations
• Description: In addition to using age, gender, and offense type to identify the SVORI
population, sites may also specifically focus on “special populations” such as
substance abusers, mentally ill offenders, offenders with co-occurring disorders (or
“dual diagnosis” – i.e., mentally ill and substance abusing offenders), or offenders
with developmental or physical disabilities. Sites may focus on other special
populations not listed above.
• Potential questions/probes: “Are you focusing on any other special population
characteristics?” (probe: For example, are you focusing on substance abusers,
mentally ill offenders, offenders with co-occurring disorders, developmentally
disabled offenders, or physically disabled offenders?”)
• Database field format/special instructions: check-box (none, substance abusers,
mentally ill, co-occurring/dual diagnosis, developmentally disabled, physically
disabled, unknown); select all categories that apply; if necessary, use the “other”
open-ended text field to enter a succinct description for a special population that is
not included in the pre-existing categories
o Note: Only use the check-boxes when a site gives an affirmative statement in
its workplan regarding targeting offenders with particular problems such as
drug use. If there are no such statements but the site says that offenders will be

9
assessed with a risk assessment that will include special problems such as
drug use, then check “none” for “special populations”, “other” for “other
inclusion criteria” and write “to be assessed with assessment instrument
(under the “other inclusion criteria” item). If there are no such statement
made and no mention made of any type of risk assessment, check “unknown”
for “special offenders”.
8.

Geographical Parameters – pre-release
• Description: It is very important for us to know what the geographical parameters of
the site/target are. Some sites may use pre-release facility as a “geographical”
parameter for their SVORI participants (i.e., the site may include offenders
incarcerated in all state prisons or a subset of prisons). We would like to know
whether the site/target includes offenders in all institutions or offenders in selected
institutions. Before contacting the project director, please try to go to the state’s
Department of Corrections website and obtain a list of all correctional institutions in
the state. After you have identified the institutions from which SVORI participants
may come (based on the interview), please enter the relevant institutions in the
“organization table” of the database (you will need addresses and a contact name for
each institution).
• Potential questions/probes: “Do your SVORI participants come from all institutions
in the state or from targeted institutions?” “Can you give me the facility names and a
contact person for each facility?” (note: use your list of state correctional institutions
to help with this question)
• Database field format/special instructions: check-box (all state prisons, selected state
prisons, unknown); if a site/SVORI program targets the majority of institutions
(excluding only highly specialized institutions such as correctional hospitals), please
check “all state prisons”; please also enter each institution in the “organization table”
of the database (be sure to flag them as pre-release facilities and select the type of
facility) and a contact person’s name in the “client table” of the database

Geographical Parameters – post-release
• Description: Some sites may use post-release residence as a geographical parameter
for their SVORI participants. We would like to know whether the site/target includes
offenders who reside (upon release) in a specific city or cities, a specific county or
counties, or the entire state.
• Potential questions/probes: “Do your program include participants whose postrelease residence is a specific city or cities, a specific county or counties, or
throughout the entire state?”
• Database field format/special instructions: check-box (single city or county, multiple
cities or counties, statewide, unknown) and open-ended text field; use the check-box
to check all that apply and use the “specify” field to list each city or county included
in the site (if the entire state is targeted, you do not need to list each county)

10.

Other inclusion criteria
• Description: Sites may use other criteria to select their program participants. For
example, their SVORI program may only include offenders who participated in a

•
•

11.

specific pre-release program (such as a particular treatment program) or the site may
say it is targeting “serious and violent” offenders or offenders at “high risk of
reoffending” (based on some type of assessment instrument). Make sure you any
information that you obtain for this item is not already covered in any of the previous
topics.
Potential questions/probes: “Are there any other inclusion criteria for SVORI
participants?” or “Do you use any other criteria to select your program participants?”
Database field format/special instructions: drop-down box (yes, no, unknown) and
open-ended text field; in the drop-down box, select “yes” or “no” depending on
whether they have any other inclusion criteria; if you select “yes”, use the open-ended
text field (“specify”) to enter a succinct description of other inclusion criteria used by
the site (for example, if the site targets “high risk” offenders based on an assessment
instrument, you could enter “ser/viol offenders; high on assmnt tool”)

Exclusion criteria
• Description: Sites may use all the age, gender, offense type, geographic, special
populations, and any other inclusion criteria above to select their population but then
apply exclusion criteria to potential participants. For example, some programs may
exclude sex offenders or severely mentally ill offenders (even if they meet the other
parameters discussed in items 2-9).
• Potential questions/probes: “Does your program have any criteria that would exclude
offenders from participating?”
• Database field format/special instructions: check-box (none, sex offenders, severely
mentally ill, unknown); select all categories that apply; if necessary, use the “other”
open-ended text field to enter a succinct description of exclusion criteria that are not
included in the check-box categories

Enrollment Information
This set of items pertains to the stage at which SVORI participants are identified for the program
and projected (or actual) enrollment information for the site/program. Most of the items
distinguish between the “pre-release” (i.e., institutional) phase and the “post-release” (i.e.,
community) phase of the program.
12.

Stage of enrollment
• Description: For this item we are interested in knowing the point at which offenders
officially begin participation in the SVORI program. We are defining “beginning”
participation as the point at which they are initially assessed for SVORI participation,
not the point at which they actually begins receiving services. The site’s solicitation
recommended that programming begin 12 months prior to release, which suggests
that assessment/enrollment should begin at least 12 months prior to release.
• Potential questions/probes: “At what point do SVORI participants officially begin
the program? Do you enroll in the institution?” (probe: “Does participation begin
prior to release?” [if so, “how many months prior to release, on average?”])

11
•

13.
•

•

14.
•

•

Database field format/special instructions: check-box (more than 1 year pre-release,
10-12 months pre-release, 7-9 months pre-release, 4-6 months pre-release, 1-3
months pre-release, less than 1 month pre-release, within 1 month after release, 1-3
months post-release, 4 or more months post-release, unknown); select the category
that applies; if necessary, use the “other” open-ended text field to enter information
that is not included in the drop-down categories
o Note: if a site’s stage of enrollment is a range that overlaps with two
categories (i.e., a program enrolls people 8-10 months prior to release), please
select the category that is closest to the release date (in the example above,
you would select the box for “7-9 months prior to release)
o Note: if a single program serves more than one population types with different
stages of enrollment (for example, Georgia includes juvenile and adult
offenders and the juvenile offenders are enrolled 2 months pre-release and the
adults are enrolled 6-9 months pre-release), please check multiple boxes (i.e.,
1-3 months pre-release and 4-6 months pre-release) and make a note in the
“other” box.
o If the site only discusses enrollment in terms of time since admission, then use
the “other” category and enter a succinct description.
o If the site states that its SVORI program will target offenders at diagnostic
using a variety of assessment instruments, please use the “other” category and
enter “at diagnostic”.
Date of enrollment of first participant – pre-release phase
Description: We would like to know both the month and year of the actual first
participant placement into the pre-release programming component of their SVORI
program. If the program is not yet operational, we would like to get the anticipated
month and year of placement of their first participant.
Potential questions/probes: “Have you placed your first client into the pre-release
programming component of your reentry program yet?”; “(if yes) What was the day
and month of the first participant enrollment?”; “(if no) When do you anticipate that
the first client will be enrolled?”
Database field format/special instructions: “forced date format” (mm/dd/yy) and
drop-down box (actual, anticipated); enter the date (mm/dd/yy) in the “forced date”
box and then use the drop-down box to select whether the date is the “actual” date of
enrollment (i.e., for programs that have already begun enrolling participants) or
“anticipated” date of enrollment (i.e., for programs that have not yet begun enrolling
participants)
Date of enrollment of first participant – post-release phase
Description: We would like to know both the month and year of the actual first
participant placement into the post-release programming component of their SVORI
program. If the program is not yet operational, we would like to get the anticipated
month and year of placement of their first participant.
Potential questions/probes: “Have you placed your first client into the post-release
programming component of your reentry program yet?”; “(if yes) What was the day

•

15.

and month of the first participant enrollment?”; “(if no) When do you anticipate that
the first client will be enrolled?”
Database field format/special instructions: “forced date format” (mm/dd/yy) and
drop-down box (actual, anticipated); enter the date (mm/dd/yy) in the “forced date”
box and then use the drop-down box to select whether the date is the “actual” date of
enrollment (i.e., for programs that have already begun enrolling participants) or
“anticipated” date of enrollment (i.e., for programs that have not yet begun enrolling
participants)

Monthly Case Intake – pre-release
Description: We would like to know the site’s actual or anticipated average monthly
intake (i.e., the # of newly enrolled individuals per month) during the pre-release
phase of the program. This number should only include those actually placed into the
pre-release programming component (exclude those assessed and/or accepted but
waiting placement). If a site only provides total monthly intake (i.e., not separated by
pre- or post-release stages), please enter the information in the pre-release field and
do not repeat it in the post-release field. Also, if a site provides annual case intake or
program capacity but not monthly breakdowns, do not use this number to calculate
monthly case intake (just select “unknown” for this field).
• Potential questions/probes: “What is your actual case intake of persons into the prerelease programming component of your reentry program?”; (If client placement has
not taken place long enough to provide a typical case intake) “What is your
anticipated average monthly case intake into the pre-release component of your
reentry program?”
• Database field format/special instructions: check-box (0-10, 11-20, 21-30, 31-40,
41-50, more than 50, unknown) and drop-down box (actual, anticipated); use the
check-box to select the category that applies and the drop-down box to indicate
whether the intake estimate is the “actual” or “anticipated” monthly intake.
•

16.

Monthly Case Intake – post-release
• Description: We would like to know the site’s actual or anticipated average monthly
intake (i.e., the # of newly enrolled individuals per month) during the post-release
phase of the program. This number should include those actually placed into the
post-release programming component (exclude those assessed and/or accepted but
waiting placement). If a site only provides total monthly intake (i.e., not separated by
pre- or post-release stages), please enter the information in the pre-release field and
do not repeat it in the post-release field. Also, if a site provides annual case intake or
program capacity but not monthly breakdowns, do not use this number to calculate
monthly case intake (just select “unknown” for this field).
• Potential questions/probes: “What is your actual case intake of persons into the postrelease programming component of your reentry program?”; (If client placement has
not taken place long enough to provide a typical case intake) “What is your
anticipated average monthly case intake into the post-release component of your
reentry program?”
• Database field format/special instructions: check-box (0-10, 11-20, 21-30, 31-40,
41-50, more than 50, unknown) and drop-down box (actual, anticipated); use the

13
check-box to select the category that applies and the drop-down box to indicate
whether the intake estimate is the “actual” or “anticipated” monthly intake.
17.

Year 1 program capacity – pre-release
• Description: We would like to know the total number of offenders the site plans on
accommodating in the pre-release phase through their SVORI funds during the first
year of the site’s project (i.e., the first year of SVORI funds). If a site only provides
total Year 1 capacity (i.e., not separated by pre- or post-release stages), please enter
the information in the pre-release field and do not repeat it in the post-release field.
• Potential questions/probes: “What is the total number of offenders your site plans on
enrolling in the pre-release phase during the first year of your SVORI program?
Database field format/special instructions: check-box (less than 50, 51-100, 101150, 151-200, more than 200, overall capacity/annual provided, unknown); select the
appropriate category
o Note: if the site only provides overall or annual capacity (i.e., not broken
down by year and phase), please select the “overall capacity/annual provided”
response option and enter the enrollment numbers in the appropriate field.
Only select one of the enrollment ranges if the site actually provides capacity
information for year 1.

18.

Year 2 program capacity – pre-release
• Description: We would like to know the total number of offenders the site plans on
accommodating in the pre-release phase through their SVORI funds during the
second year of their SVORI program (i.e., the second year of SVORI funds). If a site
only provides total Year 2 capacity (i.e., not separated by pre- or post-release stages),
please enter the information in the pre-release field and do not repeat it in the postrelease field
• Potential questions/probes: “What is the total number of offenders your site plans on
enrolling in the pre-release phase during the second year of your SVORI program?
• Database field format/special instructions: check-box (less than 50, 51-100, 101150, 151-200, more than 200, overall capacity/annual provided, unknown); select the
appropriate category
o Note: if the site only provides overall or annual capacity (i.e., not broken
down by year and phase), please select the “overall capacity/annual provided”
response option and enter the enrollment numbers in the appropriate field.
Only select one of the enrollment ranges if the site actually provides capacity
information for year 2.

19.

Year 1 program capacity – post-release
• Description: We would like to know the total number of offenders the site plans on
accommodating in the post-release phase through their SVORI funds during the first
year of the site’s project (i.e., the first year of SVORI funds). If a site only provides
total Year 1 capacity (i.e., not separated by pre- or post-release stages), please enter
the information in the pre-release field and do not repeat it in the post-release field
• Potential questions/probes: “What is the total number of offenders your site plans on
enrolling in the post-release phase during the first year of your SVORI program?

14
•

Database field format/special instructions: check-box (less than 50, 51-100, 101150, 151-200, more than 200, overall capacity/annual provided , unknown); select the
appropriate category
o Note: if the site only provides overall or annual capacity (i.e., not broken
down by year and phase), please select the “overall capacity/annual provided”
response option and enter the enrollment numbers in the appropriate field.
Only select one of the enrollment ranges if the site actually provides capacity
information for year 1.

20.

Year 2 program capacity – post-release
• Description: We would like to know the total number of offenders the site plans on
accommodating in the post-release phase through their SVORI funds during the
second year of their SVORI program (i.e., the second year of SVORI funds). If a site
only provides total Year 2 capacity (i.e., not separated by pre- or post-release stages),
please enter the information in the pre-release field and do not repeat it in the postrelease field.
• Potential questions/probes: “What is the total number of offenders your site plans on
enrolling in the pre-release phase during the second year of your SVORI program?
• Database field format/special instructions: check-box (less than 50, 51-100, 101150, 151-200, more than 200, overall capacity/annual provided ,unknown); select the
appropriate category
o Note: if the site only provides overall or annual capacity (i.e., not broken
down by year and phase), please select the “overall capacity/annual provided”
response option and enter the enrollment numbers in the appropriate field.
Only select one of the enrollment ranges if the site actually provides capacity
information for year 2.

21.

Overall Capacity
• Description: Some sites may only report overall program capacity (i.e., the total
number of offenders the site plans on accommodating over the course of the project
through SVORI funds). For these sites/programs, please use this field to enter
enrollment information. As stated for items 17-20, if a site only reports overall
program capacity, do not divide by 3 to calculate year 1 and year 2 estimates (we will
only use year-specific estimates if they are provided by the site); for items 17-20,
please check the option for “overall capacity/annual provided”.
• Potential Questions/Probes: “What is the total number or offenders your site plans on
accommodating over the course of your SVORI funding period?”
• Database field format/special instructions: check-box (less than 50, 51-100, 101-150,
151-200, more than 200, unknown); select the appropriate category

22.

Annual Capacity
• Description: Some sites may only report annual program capacity (i.e., the number of
offenders the site plans on accommodating each year over the course of the project
through SVORI funds). For these sites/programs, please use this field to enter
enrollment information. Do not calculate annual capacity based on other enrollment
information, and do not use annual capacity to create overall program capacity or year

•
•

1 or year 2 program capacity (for items 17-20, please check the option for “overall
capacity/annual provided”).
Potential Questions/Probes: “How many offenders does your site plans on
accommodating each year over the course of your SVORI funding?”
Database field format/special instructions: check-box (less than 50, 51-100, 101-150,
151-200, more than 200, unknown); select the appropriate category

23.

Current program enrollment – pre-release
• Description: We would like to know the number of offenders currently enrolled in
the pre-release phase of the site’s SVORI program (or whatever the most recent
enrollment numbers are that the site maintains). This number should only include
those actually placed into the pre-release programming component (exclude those
assessed and/or accepted but waiting placement). If a site only provides total current
program enrollment (i.e., not separated by pre- or post-release stages), please enter
the information in the pre-release field and do not repeat it in the post-release field.
• Potential questions/probes: “What is the current number of offenders enrolled in the
pre-release phase of your SVORI program?” and “What date do these numbers
reflect?”
• Database field format/special instructions: check-box (not applicable, 0-25, 26-50,
51-75, 76-100, 101-125, 126-150, more than 150, unknown) and “forced date format”
(mm/dd/yy); using the check-box, select the category that applies (for sites that have
not begun enrolling clients yet, select “not applicable”); using the “forced date” field,
enter the date that these numbers reflect (mm/dd/yy)

24.

Current program enrollment – post-release
• Description: We would like to know the number of offenders currently enrolled in
the post-release phase of the site’s SVORI program (or whatever the most recent
enrollment numbers are that the site maintains). This number should only include
those actually placed into the post-release programming component (exclude those
assessed and/or accepted but waiting placement). If a site only provides total current
program enrollment (i.e., not separated by pre- or post-release stages), please enter
the information in the pre-release field and do not repeat it in the post-release field
• Potential questions/probes: “What is the current number of offenders enrolled in the
post-release phase of your SVORI program?” and “What date do these numbers
reflect?”
• Database field format/special instructions: check-box (not applicable, 0-25, 26-50,
51-75, 76-100, 101-125, 126-150, more than 150, unknown) and “forced date format”
(mm/dd/yy); using the check-box, select the category that applies (for sites that have
not begun enrolling clients yet, select “not applicable”); using the “forced date” field,
enter the date that these numbers reflect (mm/dd/yy)

Program Information
This section covers the actual content of the site’s SVORI program, including services available
and other program components. Several of the items distinguish between what is planned for the
“pre-release” phase and the “post-release” phase of the program.

25.

Current stage of implementation
• Description: We expect substantial variability in how far along the sites are in
implementing their SVORI programs (especially depending on whether they are using
SVORI funds to expand an existing program or to implement a brand new program).
We have identified some key “progress points” (roughly based on the steps outlined
in the site’s solicitation) and would like to know which ones the sites have completed:
selected SVORI population (this does not mean that the site’s actual SVORI
participants must be identified but rather that the site has identified the population to
be served by the program), identified key partners/agencies, identified program
components and services, have the service delivery systems in place (i.e., worked out
MOUs and other background arrangements), and have already enrolled participants.
• Potential questions/probes: “We’d like to know more about your site’s current stage
of implementation. Can you tell me if each of the following has been completed:
target population selected, key partners/agencies identified, program
components/services identified, service delivery systems in place, participants already
enrolled? For each of the activities I just mentioned, can you tell me the date that the
activity was complete or, if it hasn’t yet happened, the anticipated date of
completion?”
• Database field format/special instructions: check-box (funding received, SVORI
population selected, key partners/agencies identified, program components/services
identified, service delivery systems in place, participants already enrolled) and
“forced date format”; use the check-boxes to select all activities that have been
completed (do not check activities that have not been completed); use the date fields
to enter the date that they were completed and, for activities that have not yet been
completed, the anticipated date that they will take place; for activities that were
already in place before the SVORI initiative and for which the site does not know the
date of completion, try to get an approximation

26.

Post-Release reentry authority
• Description: While offenders are incarcerated, they are clearly under the authority of
the Department of Corrections. However once they are released, the “authority” may
be transferred to another agency, such as the judiciary (i.e., for sites that are operating
reentry courts and have worked out arrangements for reentry authority to be
maintained by the courts), a community board, or an independent parole board. We
would like to know which agency has the authority over the offender once s/he is
released to the community as part of their SVORI program. In cases in which a
smaller division/unit within a larger organization is the reentry authority (i.e.,
Division of Juvenile Services under the umbrella of DOC), please use the “parent”
organization (which will be the legally binding one).
• Potential questions/probes: “In your site, who has post-release reentry authority?
The post-release reentry authority is the agency that has the ability to impose
conditions of a reentry plan and graduated sanctions and/or revocation of release if an
offender fails to comply with those conditions.” (Probe: Is your post-release reentry
authority held by the judicial branch, Department of Corrections, an independent
parole board, or a community board?)

17
•

Database field format/special instructions: check-box (Department of Corrections,
judicial branch, independent parole board, community, unknown); select all that
apply; if necessary, use the “other” open-ended text field to enter a succinct
description of another type of reentry authority

27.

Relationship to existing reentry efforts
• Description: The sites may be using SVORI funds to implement a brand new
program (where none existed before) or to expand existing programs. We would like
to know the extent to which their reentry initiative is related to existing
efforts/programs. That is, were their SVORI grant funds used to develop and
implement a new program or expand an existing program?
• Potential questions/probes: “Is your site using SVORI funds to implement a new
program, or to expand an existing program (i.e., such as filling service gaps)
Database field format/special instructions: check-box (new program, expansion of
existing program, unknown); select each category that applies

28.

Pre-release programming - duration
• Description: We would like to know the average length of time the site’s pre-release
programming component is expected to take to complete. It is understood that prerelease programming may vary tremendously (especially if a needs assessment is
conducted and programming is tailored to individual needs) but we are looking for
the average amount of time needed to complete the pre-release programming
component of their SVORI program.
• Potential questions/probes: “What is the average duration of the pre-release
programming component of your reentry initiative?” (probe: “What is the typical
length of time it is expected to take the average offender in your reentry initiative to
complete the institutional phase of the program?”)
• Database field format/special instructions: check-box (unknown, less than 1 month,
1-3 months, 4-6 months, 7-9 months, 10-12 months, more than 12 months); select the
category that applies; if necessary, use the “other” open-ended text field to describe
situations in which the programming duration varies depending on the sentence
length
o If the duration crosses categories (i.e., 6-12 months), use the “other” field and
enter the appropriate range (“6-12 months”)

29.

Pre-release programming - timeframe
• Description: We also would like to know the timing at which the pre-release
programming takes place. For example, even though the average duration of prerelease programming may be 6 months, this programming could take place 6 months
prior to release or 1 year prior to release. We are looking for the average number of
months prior to release at which pre-release programming generally begins.
• Potential questions/probes: “When does pre-release programming generally begin?”
(probe: What is the typical number of months prior to release at which pre-release
programming generally begins?”
• Database field format/special instructions: check-box (unknown, less than 1 month,
1-3 months, 4-6 months, 7-9 months, 10-12 months, more than 12 months); select the

18
category that applies; if necessary, use the “other” open-ended text field to describe
situations in which the programming timeframe varies depending on the sentence
length
o If the duration crosses categories (i.e., 6-12 months), use the “other” field and
enter the appropriate range (“6-12 months”)
30.

Pre-release components and services
• Description: We would like to know the different components and services that make
up the site’s pre-release program. Specifically, we are interested in knowing whether
the pre-release phase of their program includes a risk assessment (formal assessment
of an offender’s risk of recidivism or revocation), needs assessment (formal
assessment of an offender’s treatment needs), case management, treatment plan ,
release/reentry plan, substance abuse treatment, mental health counseling, medical
services, dental services, employment skills/vocational training, education, housing
assistance, parenting skills, domestic violence counseling, life skills training, anger
management, or any other types of services. Note: the treatment plan and release
plan may be quite similar, so go with what the site provides (if they say one or the
other then code only what they say). In theory, a reentry plan is more inclusive of a
treatment plan (including post-release treatment needs as well) but do not add this
unless the site says it.
• Potential questions/probes: “Please tell me the which of the following are included
for offenders in the pre-release phase: risk assessment, needs assessment, case
management, treatment plan, release plan, substance abuse treatment, mental health
counseling, medical services, dental services, employment skills/vocational training,
education, housing assistance, parenting skills, domestic violence counseling, life
skills training, anger management. Does your SVORI program include any other
program components or services that I didn’t mention?”
• Database field format/special instructions: check-box (risk assessment, needs
assessment, case management, treatment plan, release plan, substance abuse
treatment, mental health counseling, medical services, dental services, employment
skills/vocational training, education, housing assistance, parenting skills, domestic
violence counseling, life skills training, anger management, unknown); select all that
apply; if necessary, use the “other” field to enter a succinct description of other
services available

31.

Transfer to pre-release facility
• Description: Some sites may transfer all (or the majority of) SVORI participants to a
special pre-release facility for special services (or other purposes) prior to release,
while offenders are still serving out their sentence. For example, offenders may be
transferred to a jail, correctional work release facility, residential treatment facility, or
other pre-release facility. We are interested in knowing whether the site uses such a
facility for most or all of its SVORI participants.
• Potential questions/probes: “Does your site transfer most or all of the SVORI
participants to a pre-release facility for special programming or services prior to
release?”

19
•

Database field format/special instructions: Check-box (yes, no, unknown) and openended text field; use check-box to select appropriate category; use open-ended text
field (“specify”) to provide a succinct description of how the program uses transfers
to pre-release facilities

32.

Post-release programming - duration
• Description: We would like to know the average length of time their post-release
programming component is expected to take to complete. Once again, there may be
variation in program duration among offenders (and some offenders may technically
be required to be in the program for the entire length of their parole) but we are
looking for the average amount of time needed to complete the post-release
component of their SVORI program. Also, some programs may divide their postrelease phase into sub-phases (in which they have an intensive initial phase followed
by a less intensive phase) and we are interested in knowing whether they have
separate sub-phases among the post-release phase.
• Potential questions/probes: “What is the average duration of the post-release
component of your reentry initiative?” (probe: “What is the typical length of time it
is expected to take the average offender in your reentry initiative to complete the
post-release phase of the program?”)
• Database field format/special instructions: Check-box (less than 3 months, 4-6
months, 7-9 months, 10-12 months, 13-24 months, more than 24 months); select the
category that applies; if necessary, use the “other” text box to enter additional
information about sub-phases used by the site in the post-release phase

33.

Post-release components and services
• Description: We would like to know the different components and services that make
up the site’s post-release phase. Specifically, we are interested in knowing whether
the post-release phase of their program includes a risk assessment (formal assessment
of an offender’s risk of recidivism or revocation), needs assessment (formal
assessment of an offender’s treatment needs), case management, treatment plan,
release/reentry plan, substance abuse treatment, mental health counseling, medical
services, dental services, employment skills/vocational training, education, housing
assistance, parenting skills, domestic violence counseling, life skills training, anger
management, or any other types of services. Note: the treatment plan and release
plan may be quite similar, so go with what the site provides (if they say one or the
other then code only what they say). In theory, a reentry plan is more inclusive of a
treatment plan (including post-release treatment needs as well) but do not add this
unless the site says it.
• Potential questions/probes: “Please tell me the which of the following are included
for offenders in the pre-release phase: risk assessment, needs assessment, case
management, treatment plan, release plan, substance abuse treatment, mental health
counseling, medical services, dental services, employment skills/vocational training,
education, housing assistance, parenting skills, domestic violence counseling, life
skills training, anger management. Does your SVORI program include any other
program components or services that I didn’t mention?”

20
•

Database field format/special instructions: check-box (risk assessment, needs
assessment, case management, treatment plan, release/reentry plan, substance abuse
treatment, mental health counseling, medical services, dental services, employment
skills/vocational training, education, housing assistance, parenting skills, domestic
violence counseling, life skills training, anger management, unknown); select all that
apply; if necessary, use the “other” field to enter a succinct description of other
services available

34.

Key partnering agencies (note: this information will be entered in the “organization table”
in the database, rather than as a separate field with the other programmatic information)
• Description: We would like to know the names of all of the key partnering agencies
involved in the site’s SVORI initiative. These are agencies involved in the planning,
implementation (i.e., service delivery), and/or management of the site’s program.
These agencies do not need to be funded directly from SVORI (they could be part of
the SVORI program but funded from separate sources. Even if the workplan includes
a long list of partner agencies, if these are agencies that will be involved in delivering
services to offenders, then please enter them in the “organization” table. If the
workplan includes a “laundry list” and the agencies do not appear to be involved in
service delivery, then do not enter them. If the workplan doesn’t specify the level of
involvement, wait until the telephone interview before entering them. In cases in
which one or more smaller division/unit within a larger organization are key
partnering agencies and are involved in providing separate services, please enter each
division/unit in the “organization” table.
• Potential questions/probes: “Can you tell me the key partnering agencies involved in
the planning, implementation, and management of your SVORI program?” (note: you
might want to ask if the site contact has a directory or some type of document that
he/she could fax to you because you will want to get addresses, telephone numbers,
and fax numbers). “For each agency you named, can I please have the address and
telephone number?” “For each agency you named, can you tell me 1) whether you
have an MOU established with that agency (and if no, whether an MOU is in
progress) and 2) whether there has been previous collaboration between the lead
SVORI agency and the agency prior to receiving your 2003 SVORI grant award (and
if yes, whether this collaboration could be described as significant or minimal)?”
• Database field format/special instructions: list the key agencies as organizations in
the “organization table” of the database (enter addresses and telephone numbers for
each agency) and be sure to flag the “organization type” as “partner agency”; for each
partner agency, indicate the status of their MOU in the “MOU” field (“yes”, “no but
MOU in progress”, or “no”); for each partner agency, indicate whether previous
collaboration existed between the agency and the lead SVORI agency in the
“previous collaboration” field (“yes – significant collaboration”, “yes – minimal
collaboration”, and “no)

35.

Key stakeholders (note: this information will be entered in the “client table” in the
database, rather than as a separate field with the other programmatic information)
• Description: We would like to know the names of the individuals you consider to be
“key stakeholders” in your SVORI program. In many cases, these individuals will be

•

the main representatives from the agencies described for item 31. Even if the
workplan includes a long list of stakeholders, if the stakeholders are with agencies
that will be involved in delivering services to offenders, then please enter them in the
“client” table. If the workplan includes a “laundry list” and the stakeholders do not
appear to be involved in service delivery, then do not enter them. If the workplan
doesn’t specify the level of involvement, wait until the telephone interview before
entering them.
Potential questions/probes: “Whom do you consider to be the key stakeholders
involved in your SVORI program. We are interested in key representatives from the
agencies you just named as well as stakeholders who may be with other agencies?”
“For each individual, can I please have the address and telephone number?” (note:
you might want to ask if the site contact has a directory or some type of document
that he/she could fax to you because you will want to get addresses, telephone
numbers, fax numbers, and e-mail address).
Database field format/special instructions: list the key stakeholders as clients in the
“client table” of the database (enter contact information for each client and be sure to
link the client to the appropriate organization); identify the individual as a “key
partner”

36.

Formal Steering Committee/Management Organization
• Description: We would like to know if the site uses a formal steering committee or
some other “high level” management organization. This group would typically
include senior-level agency representatives (likely the key stakeholders identified in
the previous item) who would meet to provide guidance on planning and
implementation to the SVORI program.
• Potential questions/probes: “Does your site have a formal steering committee or
other type of management organization for your SVORI program?” “(if yes) How
often does this group meet and what is its primary function?”
• Database field format/special instructions: drop-down box (yes, no, unknown); use
the drop-down box to select whether the site has a formal steering committee or other
high level management organization

37.

SVORI-induced change in business
• Description: We would like the site contact’s opinion of how SVORI funding has
influenced the way of “doing business” in the site (i.e., the types of services that are
now available, the collaboration among agencies, the number of offenders who
receive services, etc.). This piece of information will only come from the telephone
interviews.
• Potential questions/probes: “Can you tell me how SVORI funding has changed the
way you do business in your site?”
• Database field format/special instructions: open-ended text field; enter succinct
description

38.

Programmatic Changes
• Description: This is not a stand-alone piece of information that we want you to
specifically look for or ask about. This database field will be used to document major

•

changes in the site’s program over time (including target population, enrollment
numbers, services/components, etc.). Do not worry about any major changes that
occurred between the original application and the workplan (only focus on major
changes occurring subsequently to the workplan). Because the database will only
reflect the most “current” information, this field will be used to document information
that has changed. For example, if in the telephone interview you learn that something
that was described in the workplan is now outdated, you would briefly describe the
change. This field should only be used to note substantial changes in the site’s
SVORI population, enrollment numbers, or services provided.
Database field format/special instructions: open-ended; enter a succinct description

Other Evaluability Issues
This section covers other issues that will help inform the decision regarding whether the site
would be an appropriate candidate for our subset of impact evaluation sites.
39.

Local Evaluation Efforts
• Description: While the SVORI does not require a local evaluation, some sites are
planning on using some of their funds for a local evaluation of their program. We
would like to know whether they have any plans for a local evaluation and if, so, who
is leading the local evaluation effort.
• Potential questions/probes: “Is your site planning on conducting a local evaluation
for your SVORI program?” (probe if yes: “Who do you expect to lead this local
evaluation effort?”)
• Database field format/special instructions: drop-down box (yes, no, unknown); select
whether a local evaluation is planned or not; also, if you receive the name of a local
evaluator, be sure to enter this name and contact information (try to obtain) in the
“client table” and flag this individual as the “local evaluator”

40.

MIS flagging SVORI participants
• Description: For the subset of sites in which we conduct our longitudinal offender
study, we will need a way to identify SVORI participants (and appropriate
comparison subjects) prior to their being release from prison. We are interested in
knowing whether the site contact knows of any management information system
(MIS) that contains offender-level data and somehow allows for the identification of
SVORI participants prior to their release from prison. In some cases, the MIS might
contain all offenders incarcerated in state prisons or on parole (SVORI and nonSVORI offenders). For example, the State Department of Corrections (DOC) may
have a comprehensive database of all offenders under state supervision (i.e., those on
parole, probation, and incarcerated in state prisons), with a “flag” for SVORI
participants. In other cases, the site may have set up a database that only includes
information SVORI participants.
• Potential questions/probes: “Do you know of any centralized databases or
management information systems that identify offenders as SVORI participants prior

•

to their release from prison?” (if yes) “Can you identify a contact person who knows
about this MIS?” (get contact information)
Database field format/special instructions: check-box (no MIS that identifies SVORI
participants, SVORI-specific MIS, other MIS that identifies SVORI participants,
unknown); select each category that applies; enter a “client” record for the MIS
contact person and select “MIS contact” as the person’s “role”

41.

Potential comparison subjects – within the site
• Description: For our longitudinal offender study, in addition to identifying the
SVORI participants, we are going to need to identify a comparison group of offenders
within each site (i.e., inmates released from prison who are not part of the SVORI
program). We are interested in any relevant information about where to recruit a
comparable group of offenders within the site (i.e, within the same geographical
parameters as their SVORI program) that is not receiving SVORI services. Even
though the SVORI site contact may not know much about it, it would be very helpful
to get the site contact’s opinion about what an appropriate comparison group for the
SVORI participants in their site would be. But it is fine if the project director cannot
provide any information about this. Also, you may be able to use the workplan
descriptions about local evaluation efforts (involving a comparison group) to find
appropriate information for this field. Be sure to use your own knowledge about the
site (rather than relying strictly on the PD’s opinion).
• Potential questions/probes: “Do you have any ideas about what a comparable group
of offenders within your site but not receiving SVORI services would be?”
• Database field format/special instructions: open-ended text field; enter succinct
information

42.

Potential comparison subjects – outside the site
• Description: We may need to go outside the geographical parameters of the SVORI
site to obtain a large enough comparison group (especially for sites that state that they
will be able to serve all eligible offenders). Once again, even though the SVORI site
contact may not know much about it, it would be very helpful to get the site contact’s
ideas about an appropriate comparison group for their SVORI participants but that are
located outside of the site. For example, they may know of a geographic region with
comparable demographic characteristics, crime rates, etc. as their site or similar
correctional institutions (if they are identifying SVORI participants based on
correctional institutions). Once again, it is fine if the project director cannot provide
any information about this, and be sure to use the workplan descriptions about local
evaluation efforts (involving a comparison group) to find appropriate information for
this field.
• Potential questions/probes: “Do you have any ideas about what a comparable group
of offenders outside your site would be?”
• Database field format/special instructions: open-ended text field; enter succinct
information

24
After you have covered all of these items in the telephone interviews, be sure to thank the site
contact for his/her help and tell them that we will be in touch with them regarding future SVORI
national evaluation activities.

SVORI Evaluation “Short List” Site Visit Protocol
Purpose and Overview of Site Visits
o Purpose is to determine which “short list” programs appear most suitable for
inclusion in the impact evaluation
o Note: the site visits are primarily “feasibility” visits – for sites that are selected
for the impact evaluation, we will regularly gather additional information as
needed
o Be sure to stress to the PDs that the purpose is not to evaluate or judge their
programs, we are simply trying to select the programs that meet our criteria
for inclusion in the cross-site evaluation
o Site visits will be approximately 1.5-2 days, depending on scheduling issues and the
geographical dispersion of the programs within the site
o The site visits will consist of semi-structured interviews and (if possible) observations
of programmatic activities (such as steering committee meetings)
o Note: We definitely do not want to meet with/interview any offenders, due to
human subjects protection concerns, so if the PD suggests this, please
indicate that among the sites ultimately selected for the impact evaluation we
will likely hold SVORI participant focus groups, but that we do not intend to
meet with them during the current site visits.
o The interviews will be conducted in “groups” (approximately 1 hour in duration), with
the following stakeholders:
o Project director (and whatever project staff they would like to include)
o Local evaluation (if the site is doing a local evaluation)
o DOC institutional program staff
o DOC research/MIS staff
o Community supervision staff (parole/probation)
o Community service providers
o An ideal ordering of the meetings would consist of:
o Afternoon of Day 1
Project director (and whatever project staff they would like to include)
Local evaluator
o Morning of Day 2
DOC institutional program staff
DOC research/MIS folks
o Afternoon of Day 2
Community supervision staff
Community Service providers
Final site visit “wrap up” meeting with Project Director

Site Visit Prep Work
The steps below should be completed for each program within a site. Short list
programs and staff assignments are provided in a spreadsheet (posted on the database
under “Site Liaison Materials”).
1. For all programs within a site, the staff member who conducted the original PD
interview will send an e-mail to the PD(s) indicating that we have developed a short list
of programs that we would like to visit in person and that their program(s) is one of
them. The staff member should convey that the primary goal of the site visit is to gather
additional information that would enable us to determine whether the program would be
appropriate for inclusion in our intensive evaluation and indicate that while on site, we
would like to speak with the PD (and key project staff), local evaluator, key DOC
institutional programming staff, key DOC research and MIS staff, key community
supervision staff, and key service agency representatives.
• for sites in which the staff member will be going on the site visit him/her-self, s/he
should indicate in the e-mail that s/he will be following up shortly with a phone call to
talk about dates for the site visit
• for sites in which the staff member will not be going on the site visit him/her-self, the
staff member should indicate that another RTI or UI staff member will be contacting
them to set up the site visit (after the staff member has e-mailed the PD, he/she
should let the "backup" site visit staff member know that the PD has been sent the email)
2. Once the PD(s) have received this notification, the "backup" site visit staff member
should contact the PD by telephone and:
• Reiterate the information conveyed in the e-mail
• Obtain dates in October and November that would NOT work for the site (note:
10/13 is Columbus day and some gov't offices will be closed), as well as dates that
would be particularly good (i.e., special SVORI events such as steering committee
meetings, orientations, etc. - in geographically dispersed sites staff should
particularly ask about occasions in which key people will be in the same place at the
same time). The PD may need to consult with the individuals with whom you will be
meeting while on site, but we don't want this step to drag on very long.
• note: the site visits will take approximately a day and a half to 2 days
• if the lead and/or backup already have a date in mind, they should convey this to
the PD and find out whether this is acceptable
• determine which cities in the state need to be visited (in order to keep the site visit
manageable, we will likely just visit the city where the central office is located and
perhaps 1 satellite program)
• this will be based on the number of sites and programs within the state (the PD
for each program on our list will need to be contacted)
• if a "low priority" program is not located in close proximity to the "high priority"
program, we do not necessarily need to visit that program in person (we could try
to get whatever information is available at the state office)
• if a program serves multiple counties, it will most likely not be necessary to visit
each county

3
•

•

this step could get complicated, but the point is to try to get a feel for where we
need to be while on site (keeping in mind that we would like to do as much as
possible within the city where the central office is located, only traveling to a
satellite program as necessary)
determine the "head" agency individual at the state who should be notified of the site
visit (i.e., DOC commissioner) and get contact information for this person
• notify the PD that we will be sending a letter to this individual notifying them of
our upcoming site visit (a copy of this letter can be shared with the PD)

3. Once the backup staff member has gotten the dates from the PD(s) and determined
the cities that need to be visited, the staff member should convey this to the lead staff
member and work with the lead staff member (and PD[s]) to finalize the site visit date.
• if scheduling conflicts prevent the backup staff member from being able to attend the
site visit, the staff member should contact Chris L to arrange for alternate coverage.
4. Once the site visit date has been finalized, the backup staff member should
download the head grantee agency letter from the website and mail this to the head
individual identified in #2 above. Note that several fields in this letter will need to be
filled in by the backup (please enter your name and contact information in the field for
“site liaison”).
5. Once the head grantee agency letter has been mailed, the backup staff member
should work with the PD to schedule the meetings with the individuals listed in #1
(group meetings are preferable).
• our preference would be for the PD to arrange these interviews but since we don't
want to burden the PD's, the backup should volunteer to contact the individuals to
set up the meetings
• the PD's may want something in writing to give the interviewees ahead of time; staff
members can cut and paste relevant sections from the head grantee agency letter
for this purpose
• note: we should allow at least 1 hour per "group" meeting (and longer for the PD
meeting)
6. The backup and lead staff members should coordinate their travel plans and make
airline, rental car, and hotel reservations (be sure to get gov’t rates). The backup staff
member should acquire maps, directions, contact information (etc.) from the PD(s) and
prepare an itinerary of the individual meetings for the site visit (note: in many cases, the
PD will probably prepare an itinerary for you). The backup should provide the PD with
RTI/UI staff member’s contact information while on site (i.e., cell phone number, hotel
name and number), and, if the PD was not involved in scheduling the individual
meetings, a copy of your itinerary. Other “heads up” materials to be sent to the PD
ahead of time include the 2-page overview of the cross-site evaluation (posted on the
website) and a list of the site visit questions (you will need to cut and paste these from
the protocol but be sure not to include the “front” and “back” material or the “site visit
wrap-up” material).

4
7. Prior to the site visit, both the backup and lead staff member should familiarize
themselves with all site materials (i.e., workplans, database information, 1-page
narratives) and the site visit protocol. The backup should download relevant documents
from state websites, such as lists of correctional institutions (and any available
information about enrollment, transfers, etc. at the institutions), state MIS information,
information about parole in the state, etc. Other relevant documents to bring include
business cards, a copy of the SVORI award notice (if necessary in order to qualify for
government rates), copies of the cross-site evaluation overview (to distribute to the PD),
and a re-formatted list of the site visit questions below (formatted conducive to notetaking) with any existing information derived from the telephone interviews already
inserted.

Site Visit Activities
While on the site visit, the lead staff member will lead the meetings and the backup will
take notes and help as needed. The lead and backup should keep a list of outstanding
issues/discrepancies to discuss with the PD at the final debriefing.
Below are the topics and questions that should be covered in each stakeholder meeting:
Project Director/staff
1. Program phases and components
o How many phases does your project have?
o What services and components are available to SVORI participants in each
phase (include case management, assessment, release planning,
supervision, etc.)?
o In the pre-release phase, do you know the extent of variation among
participation institutions regarding what services they deliver through SVORI
(note: this question pertains to institutional-level variation rather than
individual-level variation)?
o Are SVORI participants transferred to a special facility for pre-release
programming (i.e., while they are still serving their sentence)?
o Does the post-release phase involve group/institutional housing for
participants?
o When does each phase begin (for a given participant)?
o How long do offenders spend in each phase?
o Have all phases been developed (i.e., finalized decisions about the content of
each phase, service contracts in place, etc.)?
o Have participants been enrolled in each phase yet?
o Can participants be enrolled in phases independently of one another (e.g.,
can someone enter the community phase of the program without having
completed the institutional phase)?
2. Intake case processing/pipeline
o When are offenders screened as ‘paper eligible’? (by “paper eligible”, we
mean an initial screening conducted in a standardized way)
o Who screens potential participants?
o What are the eligibility criteria (if not clear from existing information about the
site)?
o What pre-release institutions do SVORI participants come from?
o What geographic areas can SVORI participants return to (if not clear from
existing information about the site)?

6
o What is the recruitment and admission process?
o Is entry voluntary or mandatory? (if it differs among participants, in what
situations is it voluntary/mandatory?)
o If voluntary, what incentives are used to encourage participation?
o What proportion of eligible offenders are ultimately enrolled in the program?
o (If demand exceeds supply) How do you determine which offenders to admit
into the program?
o (if demand exceeds supply) Would your program consider random
assignment?
o What type of screening information do you track (i.e., offender-level
information on screening scores, etc.)?
o How soon prior to release are participants officially identified as SVORI
participants?
o How soon prior to release do participants actually begin receiving SVORI
services?
3. Enrollment
o How many participants does your program expect to serve over the course of
your project (once a number is provided, determine what time period the PD
is talking about)?
o How many participants does your program expect to serve annually (once a
number is provided, determine what time period the PD means [i.e., 10/039/04, 10/04-9/05, etc.])?
o How many participants are currently enrolled in the program (total and by
phase, if possible)?
o How many total participants does your program expect to serve during
the time period from 5/04 through 10/05 (the “enrollment” period for our
offender study)?
o Do you expect your enrollment to be relatively evenly distributed by month (or
is the program planning on enrollment en masse)?
4. Tailoring and Coordination of Services
o How are offenders matched to available services (e.g., are some mandatory
and others available based on institutional availability or client self-selection)?
o What is your program’s approach to coordinating services for individual
participants (i.e., individual case management approach, team approach,
etc.)?
o What is your program’s approach to coordinating services among agencies
(i.e., aggregate-level coordination)?

7
o Who is responsible for coordinating services for SVORI participants in the
pre-release phase?
o Who is responsible for coordinating the transition from the pre-release phase
to the post-release phase (i.e., community “reach in”)?
o Who is responsible for coordinating services in the post-release phase?
5. Comparison group construction
o In the pre-release phase, what are the differences in the services (and service
coordination) received by SVORI participants and non-SVORI offenders?
o In the community phase, what are the differences in the services (and service
coordination) received by SVORI participants and non-SVORI offenders?
o Do you have any thoughts about how what an appropriate comparison group
to the SVORI participants would be (both within the geographical boundaries
of the site and outside of the geographical boundaries)?
6. Key agencies and stakeholders
o What agencies or organizations do you consider to be “key partnering
agencies”?
o What agencies/organizations do you have MOU’s with?
o What individuals do you consider to be key partners for the SVORI project?
7. Barriers and Solutions
o What has been necessary to get SVORI “done?”
o What have been the difficulties and barriers?
o Have there been any major changes in the scope of your program?
o Why did you decide to serve this specific population (i.e., what factors
influenced the decision to focus on this target population?)
o Would you categorize your SVORI program as “UP and running”?
o Do you forsee any major changes to your program in the future?
Local Evaluator
1.

Local Evaluation
o What do you have planned for your local evaluation?
o What comparison group are you using for your local evaluation?

2. Management information system
o Do you know what type of databases are available for offenders under state
supervision?
o Do you know what data elements these databases contain (try to obtain print
out)?

8
o Do you know what offenders are included in this database (i..e., only
incarcerated offenders, those on probation, etc.)?
o Do you know whether there is a SVORI-specific MIS?
o (if not) Is there another MIS that identifies SVORI participants?
o Do you know how often is the database updated?
3.

Institutional Data
o Are you aware of any state-level policies regarding conducting research in
state prisons?

Comparison Group Construction
o Do you have any thoughts about how what an appropriate comparison group
to the SVORI participants would be (both within the geographical boundaries
of the site and outside of the geographical boundaries)?

5. National Evaluation
o If the program were selected for the cross-site evaluation, would you have
any concerns about the program’s participation in both the cross-site and
local evaluations?

DOC institutional program staff (note: while we intend to initially obtain this
information from central DOC institutional program staff, variability among
individual institutions in the processes below may necessitate that we obtain the
information from individual facilities, if feasible)
1. Case Flow Information (only cover if staff are involved with SVORI
assessment/screening/intake)
o When are offenders screened as ‘paper eligible?’ (by “paper eligible”, we
mean an initial screening conducted in a standardized way)
o Who screens potential participants?
o What are the eligibility criteria (if not clear from existing information about the
site)?
o What pre-release institutions do SVORI participants come from?
o What geographic areas can SVORI participants return to (if not clear from
existing information about the site)?
o What is the recruitment and admission process?
o Is entry voluntary or mandatory? (if it differs among participants, in what
situations is it voluntary/mandatory?)
o If voluntary, what incentives are used to encourage participation?
o What proportion of eligible offenders are ultimately enrolled in the program?

9
o (If demand exceeds supply) How do you determine which offenders to admit
into the program?
o (if demand exceeds supply) Would your program consider random
assignment?
o What type of screening information do you track (i.e., offender-level
information on screening scores, etc.)?
o

How soon prior to release are participants officially identified as SVORI
participants?

o How soon prior to release do participants actually begin receiving SVORI
services?
2. Tailoring and Coordination of Services
o What services and components are available to SVORI participants during
the pre-release phase?
o In the pre-release phase, do you know the extent of variation among
participation institutions regarding what services they deliver through SVORI
(note: this question pertains to institutional-level variation rather than
individual-level variation)?
o How does do the pre-release services for SVORI participants differ from what
is available to non-SVORI offenders?
o How are offenders matched to available service (e.g., are some mandatory
and others available based on institutional availability or client self-selection)?
o What is your program’s approach to coordinating services for individual
participants (i.e., individual case management approach, team approach,
etc.)?
o What is your program’s approach to coordinating services among agencies
(i.e., aggregate-level coordination)?
o Who is responsible for coordinating services for SVORI participants in the
pre-release phase?
o Who is responsible for coordinating the transition from the pre-release phase
to the post-release phase (i.e., community “reach in”)?

DOC research/MIS staff
1. Management information system
o What type of databases are available for offenders under state supervision?
o What data elements do these databases contain (try to obtain print out)?

10
o What offenders are included in this database (i..e., only incarcerated
offenders, those on probation, etc.)?
o Is there a SVORI-specific MIS?
o (if not) Is there another MIS that identifies SVORI participants?
o How often is the database updated?
2. Institutional Data
o Do you have any institution-level enrollment data for the pre-release facilities
included in the SVORI program (try to get total enrollment, as well as any
breakdowns by age, race, gender, etc.)?
o Do you have any institution-level data on transfer rates?
o Are you aware of any state-level policies regarding conducting research in
state prisons?
3. Comparison Group Construction
o Do you have any thoughts about how what an appropriate comparison group
to the SVORI participants would be (both within the geographical boundaries
of the site and outside of the geographical boundaries)?

Community service providers
1. Staff involvement
o Can you describe how your agency’s staff are involved in SVORI?
2. Service provision
o What services are available to SVORI participants through your agency?
o How do these services differ for non-SVORI offenders?
o How are offenders matched to services?
o Who brokers service availability and referral (i.e., coordinates services)?
o Who monitors service use?
3. Interagency linkages and cooperation
o Can you describe how your agency communicates w/ others regarding
SVORI participants’ status?
o How would you describe the degree of information sharing?
o How would you describe the level of collaboration among key partnering
agencies in the SVORI project?
5. MIS information
o Does your agency utilize an automated system which captures information
on the services received by SVORI participants?
o What data elements are contained in that system?

11
o How often is the data updated?
o Does this system contain individual-level or aggregate data?
o Would it be possible to share this information with external researchers (if
the site were to be selected for the impact evaluation)?

Community supervision agents
1. Staff involvement
o Can you describe how your agency’s staff are involved in SVORI?
2. Service provision
o What services are available to SVORI participants through your agency?
o How do these services differ for non-SVORI offenders?
o How are offenders matched to services?
o Who brokers service availability and referral (i.e., coordinates services)?
o Who monitors service use?
3. Interagency linkages and cooperation
o Can you describe how your agency communicates w/ others regarding
SVORI participants’ status?
o How would you describe the degree of information sharing?
o How would you describe the level of collaboration among key partnering
agencies in the SVORI project?
4. Community Supervision Conditions
o Are SVORI participants on some type of formal supervision after release?
o If so, what type of supervision?
o What is the average length of time for post-release supervision?
o How (if at all) are the expectations of SVORI integrated into post-release
conditions?
o How does supervision differ for SVORI participants and regular parolees?
o Are there specific rewards or sanctions used for SVORI?
o (if yes) What are these?
5. MIS information
o Does your agency utilize an automated system which captures information
on the services received by SVORI participants?
o What data elements are contained in that system?
o How often is the data updated?
o Does this system contain individual-level or aggregate data?

12
o Would it be possible to share this information with external researchers (if
the site were to be selected for the impact evaluation)?

Site Visit “Wrap-Up” with PD
o Use this time to ask for clarification on any unresolved questions or major
discrepancies you noticed. Also, ask the PD the following questions:
o What strengths do you think your program could bring to the cross-site
evaluation?
o What concerns do you have about participation in the cross-site evaluation?
o Regarding our next steps in site selection, let the PD know that after we have
completed site visits this fall to each of the 28 grantees on our “short list”, we will, in
consultation with NIJ, select about half of these sites for the impact evaluation. In
making these decisions, we will examine the short list sites with an eye for selecting
a variety of programs across different regions of the country that represent distinct
approaches to reentry, represent a variety of population types, and that have
enrollment sufficient to support our survey field operations. We hope to make our
final decisions regarding site selection in January, and they will be notified as soon
as any decisions are made. We may also need to contact them in the meantime, in
case we need any additional information to make our decision.
o Convey to the PD that you enjoyed learning about their program and were glad to
have the opportunity to visit.

Post-Site Visit Activities
After the site visit, the backup staff member should follow-up with the PD to make sure
that any outstanding materials have been received (i.e., program materials) and send a
thank-you letter to the PD.
Any hard copy materials obtained during the site visit (i.e., program brochures,
documents, etc.) should be photocopied (2 copies) so that both RTI and UI can keep a
copy in our respective central filing cabinets. RTI’s SVORI files are in Chris’s office and
UI’s are in Meghan’s office.
As soon after (or even during, if possible) the site visit, the backup staff member will
enter the information gathered during the site visit in the “site visit information” section of
the database. After the lead staff member has reviewed and approved/edited the
information, the backup should enter a completion date for “site visit complete”.
Then, the backup should generate an auto-report containing the site visit information
entered in the database and edit the report to generate a 5-page summary for each
program. The summary should contain the following sections:
• Current program status
• Overview of the program (intake, phases, key components by phase, service
coordination, supervision)
• enrollment/case flow information
• potential comparison group members
• MIS and service data availability
After the lead has reviewed and approved/edited the report, please e-mail the report to
Laura and Chris for posting on the website.

SVORI MIS –Tables and Data Elements for SVORI Program
Information
Program Table (i.e., programmatic information for each program)
Target Population (form heading)
• Population type
o Drop down box: adults, juveniles
• Age range
o 2 drop down boxes – one for “lower age range” and one for “upper age
range” (for each box, the values should be “no limit” and the numbers
12-99)
• Gender
o Drop down box: males, females
• Offense Type(s)
o Drop-down box: sex offenses, other violent offenses, drug offenses, other
• Criminal history (i.e., number of previous offenses or incarcerations)
o Drop-down box: first time offenders, repeat offenders, other
• Special populations
o Drop down box: substance abusers, mentally ill, co-occurring (dual
diagnosis), developmentally disabled, physically disabled, other
• Geographical parameters (e.g., released into a specific county or area/quadrant)
o Open-ended text box
• Other inclusion criteria
o Open-ended text box
• Exclusion criteria
o Drop-down box: sex offenders, severely mentally ill, other
Enrollment/Program Entry Procedures (form heading)
• Screening procedures – when
o Open-ended text box
• Screening procedures – who
o Open-ended text box
• Screening procedures – what
o Drop-down box: substance abuse problems, mental illness, risk of
recidivism, educational needs, vocational/employment needs, housing
needs, other issues
• Screening procedures – formal risk assessments
o Drop-down box: Level of Service Inventory-Revised (LSI-R), Correctional
Offender Management Profiling for Alternative Sanctions (COMPAS),
Salient Factor Score , CMC (aka Wisconsin System), Risk of Reconviction
(ROC) and Criminogenic Needs Inventory (CNI), Community Risk/Needs
Management Scale (CRNMS), Case Needs Identification and Analysis
(CNIA), Rapid Risk Assessment for Sexual Offense Recidivism (RRASOR),

•
•
•
•

Static 99, Minnesota Sex Offender Screening Tool, Sexual Violence Risk –
20 (SVR-20), Hare Psychopathy Checklist –Revised (PCL-R), Violence
Risk Assessment Guide (VRAG), Youth Level of Service/Case Management
Inventory, Other (specify)
Stage of enrollment
o Open-ended text box
Procedures for program entry
o Open-ended text box
Degree of coercion for participation
o Drop-down box: voluntary, mandatory, other
Incentives for participation
o Open-ended text box

Program Information (form heading)
• Current stage of implementation
o Drop-down box: target population selected, key partners/agencies
identified, program components/services identified, service delivery
systems in place, participants already enrolled
• Post-release reentry authority
o Drop down box: traditional judge (judicial branch), administrative law
judge, parole board (executive branch), community board, other
• Date of enrollment of first participant
o Open-ended text box (Mark – can we also have a drop down box
associated with this text box with the following categories: anticipated,
actual)
• Monthly caseflow
o Open-ended text box (Mark – can we also have a drop down box
associated with this text box with the following categories: anticipated,
actual)
• Total program capacity
o Open-ended text box
• Current program enrollment
o Open-ended text box
• Goals/objectives of program
o Open-ended text box
• Compliance with SVORI 3 phase model
o Open-ended text box
• Relationship to existing reentry efforts
o Drop-down box: new program, expansion of existing program, filling
service gaps
• Pre-release programming - duration
o Open-ended text box
• Pre-release programming - timeframe
o Open-ended text box
• Pre-release program components

•

•
•

•

•
•
•
•

•

•
•
•

o Drop-down box: risk assessment, needs assessment, case management,
treatment plan development, other)
Pre-release program services
o Drop-down box: substance abuse treatment, mental health counseling,
medical services, dental services, employment skills/ vocational training,,
education, housing assistance, parenting skils, domestic violence
counseling, life skills training, anger management, other
Post-release programming - duration
o Open-ended text box
Post-release program components
o Drop-down box: risk assessment, needs assessment, case management,
treatment plan development, other (Mark – for the “other” category, can
we have an associated open-ended text box?)
Post-release program services
o Drop-down box: substance abuse treatment, mental health counseling,
medical services, dental services, employment skills/ vocational training,,
education, housing assistance, parenting skils, domestic violence
counseling, life skills training, anger management, other (Mark – for the
“other” category, can we have an associated open-ended text box?)
Post-release - supervision type
o Drop-down boxes: parole, probation, other (Mark – for the “other”
category, can we have an associated open-ended text box?)
Post-release supervision contacts
o Open-ended text box
Post-release supervision conditions
o Open-ended text box
Post-release sanctions available
o Drop-down box: jail time, community service, more intensive supervision
type, more supervision contacts, curfew/travel/other restrictions, more
frequent drug testing, increased treatment intensity, writing assignment,
jury box, fines, other (Mark – for the “other” category, can we have an
associated open-ended text box?)
Post-release rewards available
o Drop-down box: decreased community service requirements, less intensive
supervision type, fewer supervision contacts, fewer curfew/travel/other
restrictions, less frequent drug testing, decreased treatment intensity,
forgoing fines, less time in court, decreased length of supervision,
vouchers, trinkets, certificates, graduation ceremony, praise, other
Restitution requirement
o Open-ended text box
Participant payment for program/treatment
o Open-ended text box
Formal steering committee/management organization
o Open-ended text box

Other Evaluability Issues (form heading)
• Local evaluation efforts
o Open-ended text box
• MIS flagging SVORI participants
o
•

•
•

•
•
•
•
•

Open-ended text box

MIS – content
o Drop-down box: offender name, offender address, pre-release facility,
offender demographic information, offense history, supervision type,
supervision contacts, UA results, court contacts, re-arrests
MIS – accuracy and speed
o Open-ended text box
MIS – maintaining agency
o Open-ended text box
Other follow-up study participant identification info
o Open-ended text box
Possibility of random assignment – sample size issues
o Open-ended text box
Possibility of random assignment – feasibility
o Open-ended text box
Potential comparison subjects – within site
o Open-ended text box
Potential comparison subjects – outside site
o Open-ended text box

SVORI PD Interview 3/05

SVORI Project Director Interview—2005

[Site Name]
[Program Name]
Completed by: ____________________

Screening and Enrollment
The first questions have to do with how SVORI participants are currently identified and enrolled.
1. According to the information that was provided or confirmed
for the National Portrait, your program eligibility criteria are:
[import from National Portrait].
Y

a. [If yes] Do you use the MIS to generate a list of eligible
SVORI participants?

b. [If yes] Does the MIS contain a "flag" for SVORI
participants or otherwise identify offenders who are
participating in SVORI?

(1) [If no] Does your program maintain a complete
electronic list of all individuals who are enrolled in
SVORI?

3. Do you receive referrals for potential SVORI participants?

Is this correct?
a. [If no] What are the eligibility criteria you are currently
using to determine an offender’s eligibility for your
SVORI program?

2. Does your agency (DOC or DJJ) maintain an electronic
management information system (MIS) or other type of
database containing information on offenders under the
jurisdiction of the agency?

a. [If yes] Who makes these referrals? Please check all
that apply.

Facility staff
Community corrections staff
Offenders (self-referral)
Other (specify at left)

SVORI PD Interview 3/05

4. Are all offenders who meet the eligibility criteria accepted
into the SVORI program (or, if your program is voluntary,
invited to participate in the program)?

a. [If no] What are some reasons for rejecting an offender
who meets all of the eligibility criteria? Please check all
that apply.

Insufficient capacity
Offender has highly specialized needs
Offender is too much of a risk (likely to
fail)
Offender’s crime is too notorious
Offender will likely not be released by
parole board
Other (specify at left)

b. [If no] Approximately what proportion of eligible
offenders are NOT accepted into the program (or, if
your program is voluntary, invited to participate)?

N
F
U
O
M
A

5. Is program participation voluntary?

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)
Y

a. [If yes] Approximately what proportion of eligible
offenders decide NOT to participate?

b. [If yes] Has this changed during the course of the
program, and, if so, how?

The percentage has not changed
The percentage has decreased
The percentage has increased

SVORI requires too much time or effort
SVORI interferes with their ability to
participate in other programs (e.g.,
work release)
SVORI involves too much oversight
post-release
They don’t think they need the services
Other (specify at left)

[If yes] What do you think is the main reason that
offenders decline to participate? Please check only
one.

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

6. What are the consequences of dropping out during the prerelease phase? Please check all that apply.

None
Institutional infraction lodged
Lose privileges
Not be permitted in other programs
Lengthen time until release date
Other (specify at left)

7. Approximately what proportion of enrolled participants end
up dropping out prior to release?

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

SVORI PD Interview 3/05

8. What are the consequences of dropping out during the
post-release phase? Please check all that apply.

None
Returned to prison
Technical violation filed but held in
abeyance
Graduated sanction imposed
Additional conditions imposed
Other (specify at left)

9. Approximately what proportion of enrolled participants end
up dropping out post-release?

N
F
U
O
M
A

10. Once they are enrolled, can offenders be terminated from
the program (i.e., by program staff) during the pre-release
phase?

(None
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(76-99%, most)
(All)
Y

a. [If yes] To date, approximately what proportion of
enrolled participants have been terminated from the
program prior to release?

b. [If yes] Of those terminated prior to release, what was
the main reason for termination? Please check only
one.

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

Transferred to another facility
Drug use
Behavioral infractions
Failure to participate/noncompliance
with program requirements
Poor attitude
Other (specify at left)

11. Once they are enrolled, can offenders be terminated from
the program during the post-release phase?

a. [If yes] To date, approximately what proportion of
enrolled participants have been terminated after
release?

b. [If yes] Of those terminated after release, what was the
main reason for termination? Please check only one.

Transferred outside the post-release
geographical area of the program
Drug use
Committed technical violation
Committed new crime
Reincarcerated
Failure to comply with program
requirements
Poor attitude
Other (specify at left)

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

SVORI PD Interview 3/05

Assessment Tools
Now we’d like to know about the current assessment practices in your state.

Pre-Release Assessment Tools
First we’d like to know about any assessments that are currently administered prior to release.
Throughout this survey, when we refer to “comparable non-SVORI” offenders, we mean individuals
comparable to SVORI participants in terms of age, needs, and risk criteria but who are not actually enrolled
in the program.
12. Please indicate which of the following assessments are used with offenders while they are incarcerated
prior to release. For each type of assessment, please indicate whether the assessment is used with
SVORI offenders only (S), comparable non-SVORI offenders only (C), both SVORI and comparable
non-SVORI offenders (B), or none (N).
a. Risk assessment

b. Needs assessment

d. Substance abuse assessment

e. Medical/dental screening

g. IQ test

h. Literacy/educational assessment

Employment/vocational assessment

Sex offender assessment

Other (specify:
________________________________________)

Classification assessment (supervision level)

Psychology/mental health assessment

13. Does your state use the Level of Service Inventory (LSI) or a
variation on it (LSI-R, Y-LSI, YLS/CMI, YO-LSI) as part of the
pre-release assessment process (during incarceration)?

Post-Release Assessment Tools
14. Please indicate which of the following assessments are used with offenders after release. For each
type of assessment, please indicate whether the assessment is used with SVORI offenders only (S),
comparable non-SVORI offenders only (C), both SVORI and comparable non-SVORI offenders (B), or
none (N).
a. Risk assessment

b. Needs assessment

d. Substance abuse assessment

e. Medical/dental screening

Classification assessment (supervision level)

Psychology/mental health assessment

g. IQ test

SVORI PD Interview 3/05

h. Literacy/educational assessment

Employment/vocational assessment

Sex offender assessment

Other (specify:
________________________________________)

15. Does your state use the Level of Service Inventory (LSI) or a
variation on it (LSI-R, Y-LSI, YLS/CMI, YO-LSI) as part of the
post-release assessment process (following incarceration)?

Program Focus
The next questions ask about your program’s focus, in terms of target population and programming
priorities.
16. Would you say that your program primarily focuses its resources
and efforts on working with the offender prior to release (Pre),
after release (Post), or emphasizes pre- and post-release equally
(Both)? Please check only one.

Pre
Post
Both

17. For your pre-release programming, is your SVORI
program serving all facilities in the state or targeting select
facilities only? Please check only one.

All facilities
Select facilities only

18. For your post-release programming, is your SVORI program
primarily serving individuals who are returning to all communities
within the state or targeting select communities within the state?
Please check only one.

All communities
Select communities

19. Is your program primarily serving the general "serious and
violent" offender population or targeting a subset of offenders
with specific service needs? Please check only one.

20. Would you classify your program’s service provision as general,
in that you attempt to provide all needed services for participants,
or targeted, in that you focus on a specific service or small set of
specific services? Please check only one.

General “serious and violent”
offender population
Subset of offenders with specific
service needs
Other (specify at left)
General service provision
Targeted service provision
(specify at left)

21. Is the post-release phase of your program run primarily by a
government agency or a private agency? Please check only one.

Government agency
Private agency

22. Would you say your program is using SVORI funds primarily to
fill service gaps, expand existing services, or start a new
program? Please check only one.

Fill service gaps
Expand existing services
Start a new program

SVORI PD Interview 3/05

23. When thinking about providing programming or services to
offenders, what are the top three areas on which your program
focuses its resources and efforts? Please rank the three areas
by putting a “1” next to most important area, “2” next to second
most, and “3” next to the third most. (Rank only three.)

Rank
____ Employment and
vocational training
____ Physical health
____ Mental health
____ Substance abuse
____ Family support/unification
____ Community integration
____ Education and skills
building
____ Other (specify at left)

24. Besides recidivism, what outcomes does your program hope to
affect? Please list your program’s top three outcomes.

1.
2.
3.

25. If you were to be given more federal funding for reentry
programming, would you use the funds primarily to fill service
gaps, expand existing services, start a new program, or serve a
population not eligible for SVORI under the current funding
guidelines? Please check only one.

26. If you were to be given more federal funding for reentry
programming, which three programming areas would you
consider the three most important? Please rank the three areas
by putting a “1” next to most important area, “2” next to second
most, and “3” next to the third most. (Rank only three.)

Fill service gaps
Expand existing services
Start a new program
Serve a population not eligible
for SVORI under the current
funding guidelines

SVORI PD Interview 3/05

Services
Next we’d like to know about services that offenders in your state are currently receiving during incarceration
and after release. For both SVORI enrollees and comparable non-SVORI offenders (individuals comparable
to SVORI enrollees in terms of age, needs, and risk criteria but who are not actually in the program), please
circle the letter corresponding to (1) the proportion who receive or are referred to the service while they are still
incarcerated (pre-release), (2) whether the pre-release service is provided by faith-based organizations (yes
or no), (3) whether the pre-release service is provided by other community-based organizations (yes or no),
(4) the proportion who receive or are referred to the service after they are have been released (post-release),
(5) whether the post-release service is provided by faith-based organizations (yes or no), and (6) whether the
post-release service is provided by other community-based organizations (yes or no).
Pre-Release

Provided by other
community-based
organization?

N (None)
F (A few, 1–25%)
U (Just under half,
26–50%)
O (Just over half,
51–75%)
M (Most, 76–99%)
A (All)

Provided by faith-based
organization?

Provided by other
community-based
organization?

Service Type
27. Risk assessment
a. SVORI
b. Non-SVORI
28. Needs assessment
a. SVORI
b. Non-SVORI
29. Treatment/release
plan
a. SVORI
b. Non-SVORI
30. AA/NA
a. SVORI
b. Non-SVORI
31. Counseling sessions
(e.g., individual or
group; please do not
include drug education
classes)
a. SVORI
b. Non-SVORI
32. Comprehensive drug
treatment programs
(e.g., residential,
therapeutic
communities, etc.)
a. SVORI
b. Non-SVORI

Proportion
Receiving?

Provided by faith-based
organization?

Proportion
Receiving?

Post-Release

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

N (None)
F (A few, 1–25%)
U (Just under half,
26–50%)
O (Just over half,
51–75%)
M (Most, 76–99%)
A (All)

SVORI PD Interview 3/05

Pre-Release

Provided by other
community-based
organization?

N (None)
F (A few, 1–25%)
U (Just under half,
26–50%)
O (Just over half,
51–75%)
M (Most, 76–99%)
A (All)

Provided by faith-based
organization?

Provided by other
community-based
organization?

Service Type
33. Mental health services
a. SVORI
b. Non-SVORI
34. Medical services
a. SVORI
b. Non-SVORI
35. Dental services
a. SVORI
b. Non-SVORI
36. Education/GED/
tutoring/literacy
a. SVORI
b. Non-SVORI
37. Vocational training
a. SVORI
b. Non-SVORI
38. Employment referrals/
job placement
a. SVORI
b. Non-SVORI
39. Resume and
interviewing skills
development
a. SVORI
b. Non-SVORI
40. Work release program
a. SVORI
b. Non-SVORI
41. Cognitive skills
development/behavior
al programming
a. SVORI
b. Non-SVORI
42. Life skills training
a. SVORI
b. Non-SVORI
43. Legal assistance
a. SVORI
b. Non-SVORI

Proportion
Receiving?

Provided by faith-based
organization?

Proportion
Receiving?

Post-Release

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

N (None)
F (A few, 1–25%)
U (Just under half,
26–50%)
O (Just over half,
51–75%)
M (Most, 76–99%)
A (All)

SVORI PD Interview 3/05

Pre-Release

Provided by other
community-based
organization?

N (None)
F (A few, 1–25%)
U (Just under half,
26–50%)
O (Just over half,
51–75%)
M (Most, 76–99%)
A (All)

Provided by faith-based
organization?

Provided by other
community-based
organization?

Service Type
44. Assistance obtaining
identification (e.g.,
driver’s license, social
security card)
a. SVORI
b. Non-SVORI
45. Assistance obtaining
benefits and
completing
applications (e.g.,
Medicaid, disability
benefits)
a. SVORI
b. Non-SVORI
46. Financial support/
emergency assistance
(e.g., housing, clothing)
a. SVORI
b. Non-SVORI
47. Domestic violence
services (e.g., victim
and/or perpetrator)
a. SVORI
b. Non-SVORI
48. Parenting skills
development
a. SVORI
b. Non-SVORI
49. Family reunification
a. SVORI
b. Non-SVORI
50. Family counseling
a. SVORI
b. Non-SVORI
51. Anger management/
violence counseling
a. SVORI
b. Non-SVORI
52. Peer support groups
a. SVORI
b. Non-SVORI

Proportion
Receiving?

Provided by faith-based
organization?

Proportion
Receiving?

Post-Release

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

N (None)
F (A few, 1–25%)
U (Just under half,
26–50%)
O (Just over half,
51–75%)
M (Most, 76–99%)
A (All)

SVORI PD Interview 3/05

Pre-Release

Provided by other
community-based
organization?

N (None)
F (A few, 1–25%)
U (Just under half,
26–50%)
O (Just over half,
51–75%)
M (Most, 76–99%)
A (All)

Provided by faith-based
organization?

Provided by other
community-based
organization?

Service Type
53. 1-on-1 mentoring
a. SVORI
b. Non-SVORI
54. Housing placements
or referrals
a. SVORI
b. Non-SVORI
55. Transportation
a. SVORI
b. Non-SVORI
56. Other service
(specify):

Proportion
Receiving?

Provided by faith-based
organization?

Proportion
Receiving?

Post-Release

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

N (None)
F (A few, 1–25%)
U (Just under half,
26–50%)
O (Just over half,
51–75%)
M (Most, 76–99%)
A (All)

n/a
n/a

a. SVORI
b. Non-SVORI
57. Other service
(specify):

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

a. SVORI
b. Non-SVORI
58. Other service
(specify):

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

a. SVORI
b. Non-SVORI

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

F
F

U
U

O
O

M A
M A

Y
Y

N
N

Y
Y

N
N

59. Of all of the services you indicated (in questions 27–58) are
offered in your state, which three have been enhanced the most
as a result of SVORI funding?

1.
2.
3.

SVORI PD Interview 3/05

Program Components
The next set of questions pertains to other components of your program. For each component, we’d like to
know how it currently applies to both SVORI participants and comparable non-SVORI offenders. Once
again, when we refer to “comparable non-SVORI” offenders, we mean individuals comparable to SVORI
participants in terms of age, needs, and risk criteria but who are not actually enrolled in the program.
60. For any offenders in your state, does a representative from the
post-release supervision agency begin working with them while
they are still incarcerated?

a. [If yes] Does this happen for none, a few (1-25%), just under
half (26-50%), just over half (51-75%), most (76-99%), or all
of the SVORI enrollees?

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

b. [If yes] Does this happen for none, a few (1-25%), just under
half (26-50%), just over half (51-75%), most (76-99%), or all
of comparable non-SVORI offenders?

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

61. Are any offenders in your state placed on post-release
supervision?

a. [If yes] How many SVORI participants are on some type of
post-release supervision: none, a few (1-25%), just under half
(26-50%), just over half (51-75%), most (76-99%), or all?

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

b. [If yes] How many of the comparable non-SVORI offenders
are on some type of post-release supervision: none, a few (125%), just under half (26-50%), just over half (51-75%), most
(76-99%), or all?

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

[If yes] For the SVORI participants, is the pre-release
supervision agent the same person who supervises them
post-release?

62. Does your state use any reentry courts to manage returning
prisoners?

a. [If yes] Are reentry courts used for SVORI offenders (S),
comparable non-SVORI offenders (C), or both (B)?

b. [If yes] Is the reentry plan imposed by the court as a condition
of the offender’s release?

63. Has your SVORI program created a set of graduated sanctions
specifically for SVORI?

64. Has your SVORI program created a set of rewards specifically for
SVORI?

SVORI PD Interview 3/05

65. Which of the following activities are available to SVORI
participants in your state? Please check all that apply.

66. Do any offenders in your state participate in “restorative justice”
activities?

Animal training/care
Habitat for Humanity
Community beautification/
landscaping
Community service
Weed & Seed
Restitution
Victim mediation
Victim awareness/education
Y

a. [If yes] Prior to release, are these activities used for SVORI
offenders (S), comparable non-SVORI offenders (C), both
(B), or none (N)?

b. [If yes] After release, are these activities used for SVORI
offenders (S), comparable non-SVORI offenders (C), both
(B), or none (N)?

Items 67 and 68 ask about Community Accountability Panels and Offender-Specific Reentry Teams,
respectively. Community Accountability Panels are a group of agency and/or community members who
meet regularly to review the status of returning offenders. The offender appears before this board to have
his or her case reviewed, and the panel makes recommendations. The members of this panel are the same
(for the most part) for all offenders who appear before it. Offender-Specific Reentry Teams are groups
consisting of agency representatives (i.e., supervision, service providers) and/or community members. The
team composition is unique to each individual offender. The team meets to review the offender’s progress
and make recommendations.
67. For any offenders in your state, are Community Accountability
Panels or Boards utilized in the reentry process?

a. [If yes] Prior to release, are Community Accountability
Panels used for SVORI enrollees (S), comparable nonSVORI offenders (C), both (B), or none (N)?

b. [If yes] After release, are Community Accountability Panels
used for SVORI enrollees (S), comparable non-SVORI
offenders (C), both (B), or none (N)?

[If yes] Which of the following are represented on/members of
the Community Accountability Panel? Please check all that
apply.

d. [If yes] Is the composition of the Community Accountability
Panel different during the pre- and post-release phases?
(Please select “n/a” if a Community Accountability is not used
both prior to and after release.)

Faith-based organization
Other community service
providers
Law enforcement
Community Corrections/
Supervision
Corrections agency
Former prisoner representative
Victim
Family members or other
community members
Other (specify at left)

n/a

SVORI PD Interview 3/05

68. For any offenders in your state, are offender-specific reentry
teams used? (See definition on previous page.)

a. [If yes] Prior to release, are Offender-Specific Reentry
Teams used for SVORI enrollees (S), comparable nonSVORI offenders (C), both (B), or none (N)?

b. [If yes] After release, are Offender-Specific Reentry Teams
used for SVORI enrollees (S), comparable non-SVORI
offenders (C), both (B), or none (N)?

[If yes] What agencies or organizations have representatives
on the Team? Please check all that apply.

d. [If yes] Is the composition of the Offender-Specific Reentry
Team different during the pre- and post-release phases?
(Please select “n/a” if offender-specific reentry teams are not
used both prior to and after release.)
69. Does your state use video-conferencing technology to facilitate
the involvement of individuals and organizations in the reentry
process?

Faith-based organization
Other community service
providers
Law enforcement
Community
Corrections/Supervision
Corrections agency
Former prisoner representative
Victim
Family members or other
community members
Other (specify at left)

n/a

a. [If yes] Prior to release, is video-conferencing used to
facilitate communication across SVORI partnering agencies,
with individual offenders, or for some other reason? Please
check all that apply. (If video-conferencing is not used prerelease, please check “n/a.”)

Across SVORI partnering
agencies
With individual offenders
Other (specify at left)
n/a

b. [If yes] After release, is video-conferencing used to facilitate
communication across SVORI partnering agencies, with
individual offenders, or for some other reason? Please check
all that apply. (If video-conferencing is not used post-release,
please check “n/a.”)

Across SVORI partnering
agencies
With individual offenders
Other (specify at left)
n/a

[If yes] Is video-conferencing used for SVORI enrollees (S),
comparable non-SVORI offenders (C), or both (B)?

70. For prisoners in your state, do any individuals in pre-release
facilities attend curriculum-based classroom programs prior to
release?
a. [If yes] Is this curriculum completed by SVORI offenders (S),
comparable non-SVORI offenders (C), or both (B)?

SVORI PD Interview 3/05

b. [If yes] What topics are addressed in the program(s)? Please
check all that apply.

Basic education/GED/college
courses
Cognitive skills
Computer skills
Basic vocational training
Employment issues
Money management
Family issues
Time management
Substance abuse issues
Health/nutrition
Mental health
Finding a place to live
Where to go for legal assistance
Other (specify at left)

Faith-based organizations only
Other community-based
organizations
Both faith- and other communitybased organizations
Neither type of organization

[If yes] Do the programs involve staff from faith-based
organizations, other community-based organizations, both
faith-based and other community-based organizations, or
neither type of organization?

The next questions are about individuals and organizations that may be involved in the reentry process in
your correctional system in a routine or systematic way.
71. For any offenders in your state, are family members routinely
involved in the reentry process?

a. [If yes] Prior to release, are family members routinely
involved for SVORI enrollees (S), comparable non-SVORI
offenders (C), both (B), or none (N)?

b. [If yes] After release, are family members routinely involved
for SVORI enrollees (S), comparable non-SVORI offenders
(C), both (B), or none (N)?

72. For any offenders in your state, is a victim routinely involved in
the reentry process?

a. [If yes] Prior to release, are victims routinely involved for
SVORI enrollees (S), comparable non-SVORI offenders (C),
both (B), or none (N)?

b. [If yes] After release, are victims routinely involved for
SVORI enrollees (S), comparable non-SVORI offenders (C),
both (B), or none (N)?

73. For any offenders in your state, is law enforcement routinely
involved in the reentry process?

a. [If yes] Prior to release, is law enforcement routinely
involved for SVORI enrollees (S), comparable non-SVORI
offenders (C), both (B), or none (N)?

b. [If yes] After release, is law enforcement routinely involved
for SVORI enrollees (S), comparable non-SVORI offenders
(C), both (B), or none (N)?

SVORI PD Interview 3/05

74. For any offenders in your state, are former prisoners routinely
involved in the reentry process?

a. [If yes] Prior to release, are former prisoners routinely
involved for SVORI enrollees (S), comparable non-SVORI
offenders (C), both (B), or none (N)?

b. [If yes] After release, are former prisoners routinely involved
for SVORI enrollees (S), comparable non-SVORI offenders
(C), both (B), or none (N)?

75. Are any offenders in your state offered the option of having a
mentor during the reentry process?

a. [If yes] Prior to release, are mentors offered to SVORI
enrollees (S), comparable non-SVORI offenders (C), both (B),
or none (N)?

b. [If yes] After release, are mentors offered to SVORI
enrollees (S), comparable non-SVORI offenders (C), both (B),
or none (N)?

76. Of all the program components covered in this section (questions
60–74), which three have been enhanced the most as a result of
SVORI funding?

1.
2.
3.

Coordination
Service Coordination
The next set of questions pertains to different methods of service coordination. For each type of service
coordination strategy, we’d like to know whether you offer it and the extent to which the strategy has been
affected by SVORI.
77. Does your program provide case management to offenders prior
to release?

a. [If yes] Please indicate the proportion of SVORI offenders
who receive case management during the pre-release
period.

b. [If yes] Who provides the pre-release case management for
SVORI participants? Please check all that apply.

Facility staff
Grantee agency staff (other than
facility staff)
Faith-based organization
Other community organization or
service provider
Other (specify at left)

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

SVORI PD Interview 3/05

[If yes] Please indicate the proportion of comparable, nonSVORI offenders who receive case management during the
pre-release period.

N
F
U
O
M
A

78. Does your program provide case management to offenders after
release?
a. [If yes] Please indicate the proportion of SVORI offenders
who receive case management during the post-release
period.

N
F
U
O
M
A

b. [If yes] For SVORI participants, is the pre-release case
manager the same person who will work with them postrelease?
c.

[If yes] Who provides the post-release case management for
SVORI participants? Please check all that apply.

d. [If yes] Please indicate the proportion of comparable, nonSVORI offenders who receive case management during the
post-release period.

b. [If yes] How has the use of this practice changed as a result
of SVORI funding? Is there no change (NC) as a result of
SVORI, is it a new practice (N), or has the use of the practice
been expanded or enhanced (E)?
80. Does your program have an individual or set of individuals who
work to develop or build service provider networks (sometimes
termed a boundary-spanner)?

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)
Y

Supervision agency
Grantee agency staff (other than
supervision agent)
Other community organization or
service provider
Faith-based organization
Other (specify at left)

79. Does your program use a “continuity of care” model in which a
case manager, supervision officer, or service provider is involved
with an individual from the pre-release facility to the community?
a. [If yes] Who provides the continuity of care? Please check all
that apply.

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)

N
F
U
O
M
A

(None)
(A few, 1-25% )
(Just under half, 26-50% )
(Just over half, 51-75%)
(Most, 76-99%)
(All)
Y

Supervision officer
Case manager
Service provider
Other (specify at left)

SVORI PD Interview 3/05

a. [If yes] What are some responsibilities of this (these)
individual(s)? Please check all that apply.

Building relationships with
community agencies
Educating community service
providers about the unique
needs of former prisoners
Encouraging providers to
prioritize or begin serving
returning prisoners
Ensuring the availability of
service providers able and willing
to accept referrals
Other (specify at left)

81. Does your program use a one-stop shop within which a variety of
treatment providers are available to provide referrals or services
to offenders in a single location?

a. [If yes] Are representatives from the post-release supervision
agency (e.g., parole officers) located in the one-stop shop?

b. [If yes] Are faith-based organizations among the providers
available in the one-stop shop?

[If yes] How has the use of this practice changed as a result
of SVORI funding? Is there no change (NC) as a result of
SVORI, is it a new practice (N), or has the use of the practice
been expanded or enhanced (E)?

82. Does your program use a “wrap-around” approach where a broad
set of interested agencies are involved in developing and
delivering a comprehensive, individualized treatment plan that
takes into account the offender’s entire social network?
a. [If yes] What types of agencies are involved in this process?
Please check all that apply.

b. [If yes] How has the use of the wrap-around approach
changed as a result of SVORI funding? Is there no change
(NC) as a result of SVORI, is it a new practice (N), or has the
use of the practice been expanded or enhanced (E)?

Law enforcement
Facility staff
Post-release supervision
Employment
Health
Mental health
Substance abuse
Education
Faith-based
Other (specify at left)

SVORI PD Interview 3/05

Program Coordination
Think of the primary agencies you work with to serve SVORI offenders
83. How often does phone or e-mail contact occur between SVORI
program staff and the primary agencies?

84. Since SVORI funding began, has the frequency of phone or
e-mail contact among the agencies increased, decreased, or
stayed the same?

Increased
Decreased
Stayed the same

85. How often are meetings held between SVORI program staff and
the primary agencies to discuss the quality and content of the
overall services provided?

86. Since SVORI funding began, has the frequency of meetings
among the agencies discussing the quality and content of the
overall services increased, decreased, or stayed the same?

Increased
Decreased
Stayed the same

87. How often are meetings held between SVORI program staff and
the primary agencies to discuss services to individual SVORI
offenders?

88. Since SVORI funding began, has the frequency of meetings
among the agencies to discuss services to individual offenders
increased, decreased, or stayed the same?

Increased
Decreased
Stayed the same

89. How often are meetings held between SVORI program staff and
the primary agencies to strategize about the implementation of
approaches to serve SVORI offenders? (For example, shared
decision-making about offender accountability and how the
system will address it.)

90. Since SVORI funding began, has the frequency of meetings to
strategize about the implementation of approaches to serve
offenders increased, decreased, or stayed the same?

Increased
Decreased
Stayed the same

91. How often do SVORI program staff and the primary agencies
contact one another to facilitate referrals for SVORI participants?

92. Since SVORI funding began, has the frequency of agency
contact with one another to facilitate referrals for offenders
increased, decreased, or stayed the same?

Increased
Decreased
Stayed the same

Weekly
Monthly
Quarterly
Annually
Not at all

93. Please indicate whether you strongly agree (SA), agree (A), neither agree nor disagree (N), disagree
(D), or strongly disagree (SD) with each of the following statements about your SVORI program:
a. A core group of SVORI staff is responsible for handling the
day-to-day implementation of program (grant) activities.

SVORI PD Interview 3/05

b. Information sharing about specific offenders across
partnering agencies has improved as a result of SVORI.

d. Partnering agencies have developed a common vision of
reentry as a result of SVORI.

e. Partnering agencies have created common goals related to
reentry as a result of SVORI.

Communication across partnering agencies has improved as
a result of SVORI.

SVORI is a collaborative effort among different agencies.

Current Program Status
94. Would you say your SVORI program is fully operational? By “fully
operational” we mean that the program is up and running and,
although the program may evolve, all of the program components
are currently being implemented.

N
[If no, skip to 94d]

a. [If yes] When would you say your program became fully
operational? (month/year)

_____/_____

b. [If yes] When did you enroll your first participant?
(month/year)

_____/_____

[If yes] How long did it take to get your program up and
running once all of the federal funds were released?

< 3 months
3–5 months
6–8 months
9–11 months
12+ months

d. [If no] Please describe what part(s) of your program still
need(s) to be implemented and explain the reasons for the
delay.

e. [If no] Provide an estimate of the earliest date by which your
program will be fully operational.

Estimate: _____/_____

95. How many total SVORI participants had you enrolled by
12/31/04?

Number:

96. How does this number compare with your original projections?

Fewer than originally projected
About the same as originally
projected
More than originally projected

97. How many SVORI participants are currently enrolled in the prerelease phase of your program?

Number:

a. As of what date?
98. How many SVORI participants are currently enrolled in the postrelease phase of your program?
a. As of what date?

(month/year): _____/_____
Number:
(month/year): _____/_____

SVORI PD Interview 3/05

The next set of questions pertains to issues that you may have encountered regarding recruiting or enrolling
SVORI participants.
99. Please indicate the extent to which you agree or disagree that the following issues have limited the
number of offenders you were able to enroll. Please indicate whether you strongly agree (SA), agree
(A), neither agree nor disagree (N), disagree (D) , or strongly disagree (SD) with each of the following
statements:
a. Not enough offenders are being screened for potential
eligibility. (Select “n/a” if your program does not have a
screening process.)

n/a

b. The agency’s management information system (MIS) or
electronic database does not include the data we need to
determine if someone is eligible. (Select “n/a” if your agency
does not have an MIS.)

n/a

The agency’s MIS is difficult to use or is hard to access.
(Select “n/a” if your agency does not have an MIS.)

n/a

d. We have had difficulty obtaining information on eligible
offenders from the facilities. (Select “n/a” if facilities are not
involved with the identification of eligible participants.)

n/a

e. Accurate current information about release dates for potential
participants has not routinely been available.
f.

Accurate current information about post-release plans (e.g.,
post-release area of residence) has not routinely been
available.

g. Our program’s eligibility criteria have been too stringent.
h. Inadequate referrals have been made by staff at the facilities.
(Select “n/a” if facility staff are not responsible for making
referrals in your program.)

n/a

Facility or agency policies have made it difficult to transfer
eligible offenders to other facilities for SVORI programming or
to prevent the transfer of SVORI participants to facilities that
do not offer SVORI programming. (Select “n/a” if participants
are not transferred for programming or if SVORI is offered at
all facilities.)

n/a

Offenders have been identified but decline to participate.
(Select “n/a” if your program is not voluntary.)

n/a

Offenders have been identified too late to complete prerelease programming (i.e., too close to release date). (Select
“n/a” if your program does not provide pre-release
programming.)

n/a

We have not had the resources to serve the number of
offenders that are identified.

SVORI PD Interview 3/05

m. Please describe any other obstacles to recruitment or enrollment that you have encountered in your
program.

Issues Surrounding SVORI Implementation
100. Please indicate the extent to which you agree or disagree with the following statements about issues
that might have arisen regarding SVORI program implementation. Please indicate whether you
strongly agree (SA), agree (A), neither agree nor disagree (N), disagree (D) , or strongly disagree
(SD) with each of the following statements:
a. We have encountered resistance from…
(1) top administrators at the facilities.

(2) supervisors at the facilities.

(3) line staff at the facilities.

(4) top administrators at the post-release supervision
agency.

n/a

(5) supervisors at the post-release supervision agency.

n/a

(6) line officers of the post-release supervision agency.

n/a

(7) some of the SVORI partner agencies in the community.

(8) members of the community to which SVORI offenders
return (the ‘not in my backyard’ syndrome).

b. Existing agency regulations or policies have made it difficult
to implement SVORI.

There has been poor communication within agencies.

d. There has been poor communication between agencies.

e. We have experienced turf battles.

g. The available funding has been poorly allocated.

h. We have had insufficient staff available.

Staff training has been inadequate.

Staff turnover has been high.

There has been inadequate availability of services for
referrals we have made.

Funding for reentry is inadequate.

SVORI PD Interview 3/05

Sustainability
101. Do you consider the current political climate in your community
to be favorable to your reentry programming?

102. Are there other reentry initiatives under way in your state?

103. What are your plans for your reentry program once SVORI
funds are no longer available? Please check all that apply.

Discontinue the program
Continue the program
Expand the program
Replace the program

a. [If you plan to continue or expand the program] Do you think
that you have sufficient resources to continue the initiative at
the current level?

b. [If you plan to continue or expand the program] Will your
initiative continue beyond the grant period without additional
funds from the federal government?

Have used/
currently using

Planning to use/
continue using

[If you plan to continue or expand the program] Are you
currently working on ways to sustain the initiative beyond the
grant period?

d. [If you plan to continue or expand the program] For each of
the following strategies, please indicate whether (1) you have
used or are currently using the strategy (Y/N), and (2)
whether you are planning to use (or continue to use) the
strategy in the future (Y/N).
(1) Pursue additional federal funding

(2) Pursue additional state funding

(3) Pursue additional funding from local sources

(4) Pursue additional funding from other sources
(Specify:
__________________________________________)

(5) Reallocate resources within the current agency

(6) Reallocate resources across the partnering agencies

(7) Communicate with policy makers about the program

(8) Conduct a local evaluation

(9) Develop a Web site to convey information about the
program

(continued)
(10)Develop printed materials to convey information about
the program

Have used/
currently using
Y

Planning to use/
continue using
Y

SVORI PD Interview 3/05

(11)Work with the media (e.g., press releases, conferences,
interviews, newspaper articles)

(12)What other strategies are you using or planning to use to
sustain or expand your program? (Specify:
_____________
_____________________________________________
_____________________________________________
__________)

104. Please indicate whether you strongly agree (SA), agree (A), neither agree nor disagree (N), disagree
(D), or strongly disagree (SD) with the following statements:
a. The SVORI initiative is not worth continuing.

b. Reentry programming is no better now than it was before
SVORI.

SVORI is helpful to the current target population.

d. SVORI would be helpful to all returning offenders.

105. Finally, from your perspective, what characteristics of your SVORI program make it particularly unique
or innovative?

Thank you so much for taking the time to complete this survey.
If we need to follow up on any of the responses, whom is the most appropriate person for us to contact?
Name: ________________________
Phone No.: ____________________
Please make a photocopy of this survey and mail the original to RTI using the Federal Express mailing label.
If you have misplaced the label, please contact Mark Pope at (919) 485-5701.

SVORI Program Director Interview—2006

SVORI PD Interview 3/06

«SiteName»
«TargetName»
Completed by: ____________________
Date completed: ____________________
Program Status
The first set of questions pertains to the status of your SVORI program.
1. When would you say all of the planned elements of your SVORI
program became fully operational (month/year)?

2. Does your program still have a SVORI program director?
3. Over the course of your grant, how many individuals have held the
SVORI program director position?
4. Have you applied for a no-cost extension for your original SVORI
grant?

Program became fully
operational on ___/____
Program has not become fully
operational
Y

Number:

5. What is the current end date of your SVORI grant (including any
no-cost extensions you have received or will receive on your
SVORI grant)? Please do not include extensions as a result of any
supplementary funds you may have received from other sources.

(month/year): _____/_____

6. What was the original end date of your SVORI grant?

(month/year): _____/_____

Enrollment
The next questions pertain to your program’s enrollment.
7. When did you enroll your first participant (month/year)?

8. As of 3/1/2006, what was the total cumulative enrollment in your
SVORI program (i.e., how many individuals did you enroll in your
program from its inception to 3/1/06?)

9. How does this number compare with your original projections?

We enrolled our first
participant on ___/____
We have not enrolled any
participants
Number:

Fewer than originally
projected
About the same as originally
projected
More than originally projected

10. How many SVORI participants are currently enrolled in the prerelease phase of your program?

Number:

11. How many SVORI participants are currently enrolled in the postrelease phase of your program?

Number:

SVORI PD Interview 3/06

12. Are you still enrolling new participants into your program?
a.

[If yes] How long do you expect to continue enrolling new
participants into your program?

b. [If no] When did you stop enrolling new participants into your
program?
13. Did your SVORI grantee agency (e.g. your Department of
Corrections or Juvenile Justice agency) set an enrollment target for
your program?
a.

[If yes] Did your SVORI grantee agency monitor progress
toward this target?

14. Did the top administrators at your SVORI grantee agency (e.g.,
DOC or DJJ) set implementation goals for your program?
a.

[If yes] Did your SVORI grantee agency monitor progress
toward these goals?

15. Which phase of your program was more difficult to implement?
(Please check only one.)
16. What were the three
most significant factors
that limited the number
of participants you were
able to enroll in your
program? Please rank
these three factors by
putting a “1” next to the
most significant factor
“2” next to second most
significant, and “3” next
to the third most
significant. (Please rank
only three.)

Plan to continue enrolling
until approximately
(month/year) ___/____
Plan to continue enrolling
indefinitely
(month/year): _____/_____

Pre-release
Post-release

Rank
____ Not screening enough offenders for potential eligibility
____ Your program’s eligibility criteria being too restrictive (i.e., not enough
eligible offenders available)
____ The federal funding agency’s eligibility criteria being too restrictive
____ Your pre-release agency’s management information system (MIS) or
electronic database not including the data needed to determine if
someone is eligible
____ Your pre-release agency’s MIS being difficult to use or hard to access
____ Accurate current information about release dates for potential
participants not routinely being available
____ Accurate current information about post-release plans not routinely
being available
____ Inadequate referrals by facility staff
____ Facility or agency policies making it difficult to deliver SVORI
programming
____ Offenders declining to participate
____ Offenders being identified too late to complete post-release
programming (i.e., too close to release date)
____ Inadequate resources to serve the number of offenders identified by
facility staff

SVORI PD Interview 3/06

Services
The next questions ask about your program’s programming priorities and desired outcomes.
17. When thinking about providing
programming or services to offenders,
how has your program focused its
resources and efforts overall throughout
the course of your program? Please rank
the areas by putting a “1” next to your top
focus, “2” next to the second focus, “3”
next to the third focus, “4” next to the
fourth focus, and “5” next to the fifth
focus. (Please rank all areas.)

Rank
____ Assessment, Coordination, and Supervision Services
(e.g., risk/needs assessments, treatment/release plan
development, post-release supervision)
____ Transition Services (e.g., housing placements/referrals,
assistance obtaining identification and benefits, legal
assistance, financial support/emergency assistance,
peer support, mentoring)
____ Health Services (e.g., substance abuse treatment,
counseling, mental health services, anger
management/violence counseling, medical services,
dental services)
____ Employment, Education, and Skills Development
Services (e.g., education/GED/tutoring/literacy
services, vocational training, employment referrals/job
placement, resume/ interviewing skills, work release,
cognitive skills development/behavioral programming,
life skills)
____ Family services (e.g., family reunification, family
counseling, parenting skills, domestic violence
services)

18. If you were to be given more federal
funding for reentry programming, how
would you focus your resources? Please
rank the areas by putting a “1” next to
your top focus, “2” next to the second
focus, “3” next to the third focus, “4” next
to the fourth focus, and “5” next to the
fifth focus. (Please rank all areas.)

SVORI PD Interview 3/06

19. Besides recidivism, what outcomes does
your program hope to affect for individual
participants? Please rank the three most
important outcomes by putting a “1” next
to most important outcome, “2” next to
second most, and “3” next to the third
most. (Please rank only three.).

Rank
____
____
____
____
____
____
____
____
____

20. If someone were evaluating the
effectiveness of your SVORI program,
what measurable outcomes do you think
it would be fair to use to determine
program effectiveness? (Please check all
that apply.)

Reduced substance use
Improved physical and/or mental health
Employment
Educational attainment
Housing
Family reunification/functioning
Community integration/connectedness
Improved decision-making or self-sufficiency
Other (please specify in the box at the left)

Reduced recidivism
Reduced substance use
Improved physical and/or mental health
Employment
Educational attainment
Housing
Family reunification//functioning
Community integration/connectedness
Improved decision-making or self-sufficiency
Other (please specify in the box at the left)

SVORI PD Interview 3/06
Next we’d like to know about services that offenders in your state are currently receiving during incarceration and after release. For each service type
in the table below, please indicate the following by circling the appropriate letter:
Post-release services
– Whether post-release services of this type have changed (N, S, NC,
NA) as a result of SVORI [if the service is not available to any
offenders, circle NA and skip the following two steps related to
proportions served]
– The proportion (N, F, U, O, M, A) of SVORI participants who receive
or are referred to the service post-release [circle the letter on the “a”
line]. If you are not currently serving any SVORI participants postrelease, please leave the “a” line blank.
– The proportion (N, F, U, O, M, A) of the general serious and violent
offender (General SVO) inmate population who receive or are referred
to the service post-release [circle the letter on the “b” line”]

Pre-release services
– Whether pre-release services of this type have changed (N, S, NC,
NA) as a result of SVORI [if the service is not available to any
offenders, circle NA and skip the following two steps related to
proportions served]
– The proportion (N, F, U, O, M, A) of SVORI participants who receive
or are referred to the service pre-release [circle the letter on the “a”
line]. If you are not currently serving any SVORI participants prerelease, please leave the “a” line blank.
– The proportion (N, F, U, O, M, A) of the general serious and violent
offender (General SVO) inmate population who receive or are
referred to the service pre-release [circle the letter on the “b” line”]
Pre-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

21. Case management
a. SVORI
b. General SVO population
22. Risk assessment
a. SVORI
b. General SVO population
23. Needs assessment
a. SVORI
b. General SVO population
24. Treatment/release plan
a. SVORI
b. General SVO population
25. Formal post-release supervision
a. SVORI
b. General SVO population

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

Proportion receiving the
service:
N
F
U
O
M
A

(None, but service available)
(A few, 1–25%)
(Just under half, 26–50%)
(Just over half, 51–75%)
(Most, 76–99%)
(All)

S NC NA

Post-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A

N F U O M A
N F U O M A
N

not applicable

N (None, but service
available)
F (A few, 1–25%)
U (Just under half, 26–50%)
O (Just over half, 51–75%)
M (Most, 76–99%)
A (All)

S NC NA

N F U O M A
N F U O M A
N

Proportion receiving the
service:

S NC NA
N F U O M A
N F U O M A

SVORI PD Interview 3/06
Pre-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

26. In-person contact from the postrelease case manager or
supervision officer while the
offender is still incarcerated
a. SVORI
b. General SVO population
27. Reentry courts
a. SVORI
b. General SVO population
28. Video-conferencing
a. SVORI
b. General SVO population
29. Offender-specific reentry teams
(groups consisting of agency
representatives and/or community
members that review and develop
a plan for the offender)
a. SVORI
b. General SVO population
30. AA/NA
a. SVORI
b. General SVO population
31. Counseling sessions for drug or
alcohol use (e.g., individual or
group; please do not include drug
education classes)
a. SVORI
b. General SVO population

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

Proportion receiving the
service:
N
F
U
O
M
A

(None, but service available)
(A few, 1–25%)
(Just under half, 26–50%)
(Just over half, 51–75%)
(Most, 76–99%)
(All)

Post-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

Proportion receiving the
service:
N (None, but service
available)
F (A few, 1–25%)
U (Just under half, 26–50%)
O (Just over half, 51–75%)
M (Most, 76–99%)
A (All)

S NC NA
not applicable
N F U O M A
N F U O M A

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A

S NC NA

N F U O M A
N F U O M A

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

N F U O M A
N F U O M A

S NC NA
N F U O M A
N F U O M A

SVORI PD Interview 3/06
Pre-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

32. Comprehensive drug treatment
programs (e.g., residential,
therapeutic communities, etc.)
a. SVORI
b. General SVO population
33. Mental health services
a. SVORI
b. General SVO population
34. Anger management/violence
counseling
a. SVORI
b. General SVO population
35. Education/GED/tutoring/literacy
a. SVORI
b. General SVO population
36. Employment referrals/job
placement
a. SVORI
b. General SVO population
37. Resume and interviewing skills
development
a. SVORI
b. General SVO population
38. Cognitive skills development/
behavioral programming
a. SVORI
b. General SVO population
39. Life skills training
a. SVORI
b. General SVO population

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

Proportion receiving the
service:
N
F
U
O
M
A

(None, but service available)
(A few, 1–25%)
(Just under half, 26–50%)
(Just over half, 51–75%)
(Most, 76–99%)
(All)

S NC NA

Post-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

N F U O M A
N F U O M A

N (None, but service
available)
F (A few, 1–25%)
U (Just under half, 26–50%)
O (Just over half, 51–75%)
M (Most, 76–99%)
A (All)

S NC NA

N F U O M A
N F U O M A
N

Proportion receiving the
service:

S NC NA
N F U O M A
N F U O M A

SVORI PD Interview 3/06
Pre-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

40. Pre-release curriculum
a. SVORI
b. General SVO population
41. Assistance obtaining
identification (e.g., driver’s
license, social security card)
a. SVORI
b. General SVO population
42. Assistance obtaining benefits and
completing applications (e.g.,
Medicaid, disability benefits)
a. SVORI
b. General SVO population
43. Financial support/emergency
assistance (e.g., housing, clothing)
a. SVORI
b. General SVO population
44. Parenting skills development
a. SVORI
b. General SVO population
45. Family reunification
a. SVORI
b. General SVO population
46. Peer support groups
a. SVORI
b. General SVO population
47. One-on-one mentoring
a. SVORI
b. General SVO population

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

Proportion receiving the
service:
N
F
U
O
M
A

(None, but service available)
(A few, 1–25%)
(Just under half, 26–50%)
(Just over half, 51–75%)
(Most, 76–99%)
(All)

Post-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

not applicable

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

S NC NA

N F U O M A
N F U O M A
N

N (None, but service
available)
F (A few, 1–25%)
U (Just under half, 26–50%)
O (Just over half, 51–75%)
M (Most, 76–99%)
A (All)

S NC NA
N F U O M A
N F U O M A

Proportion receiving the
service:

S NC NA

N F U O M A
N F U O M A
N

N F U O M A
N F U O M A

S NC NA
N F U O M A
N F U O M A

SVORI PD Interview 3/06
Pre-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

48. Housing placements or referrals
a. SVORI
b. General SVO population
49. Transportation
a. SVORI
b. General SVO population

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

Proportion receiving the
service:
N
F
U
O
M
A

(None, but service available)
(A few, 1–25%)
(Just under half, 26–50%)
(Just over half, 51–75%)
(Most, 76–99%)
(All)

S NC NA

Post-Release
How has the service
changed as a result of
SVORI?
N
S
NC
NA

(Newly implemented)
(Substantially enhanced)
(No substantial change)
(Service not available)

N (None, but service
available)
F (A few, 1–25%)
U (Just under half, 26–50%)
O (Just over half, 51–75%)
M (Most, 76–99%)
A (All)

S NC NA

N F U O M A
N F U O M A

N F U O M A
N F U O M A
N

not applicable

Proportion receiving the
service:

S NC NA
N F U O M A
N F U O M A

50.

Please describe your program’s approach to service coordination.
We may post your response on your program’s profile on the SVORI Multi-Site Evaluation website. Please check here if you do not want your
response posted:

51.

Please describe any programming delivered to SVORI participants once the formal post-release supervision phase is complete (i.e., the “Sustain
and Support” phase described in the original SVORI solicitation).
We may post your response on your program’s profile on the SVORI Multi-Site Evaluation website. Please check here if you do not want your
response posted:

SVORI PD Interview 3/06

Organizational Context
The next set of questions asks about organizational context, including interagency communication and
collaboration.
52. What were the most significant barriers to implementation
that your program encountered? Please rank the top three
barriers by putting a “1” next to biggest barrier “2” next to
second biggest, and “3” next to the third biggest. (Please
rank only three.)

Rank
____
____
____
____
____
____
____
____
____
____
____

Existing agency regulations or policies
Turf battles
Inadequate funding
Poor allocation of available funding
Insufficient staff
Inadequate staff training
Staff turnover
Inadequate availability of services
Poor intra-agency communication
Poor inter-agency communication
Other (please specify in the box at left)

53. Please complete the table below, indicating whether each of the following agencies or community-based
organizations (CBO) has been involved in your SVORI programming and the extent to which you agree or
disagree with the statements about the agency’s involvement. (Please complete the entire row for each
agency/CBO, even if you answer “no” in the first column.)
Has this
agency/
CBO been
involved
in your
SVORI
program?
a.

Pre-release supervision
agency (e.g., DOC/DJJ)
b. Post-release supervision
agency
c. Faith-based organizations
d. Substance abuse agencies
or CBO’s
e. Mental health agencies or
CBO’s
f. Family/social services
agencies or CBO’s
g. Law enforcement agency
h. Housing agencies or
CBO’s
i. Employment agencies or
CBO’s
j. Vocational training
agencies or CBO’s
k. Technical institutions,
community colleges, and
universities
l. [Juvenile programs only]
Local school systems

Do you strongly agree (SA), agree (A), neither agree nor disagree
(N), disagree (D), or strongly disagree (SD) with the following?
We have
encountered
resistance from this
agency/CBO as we
implemented
SVORI.

Support for SVORI
from this agency/
CBO has been
strong.

This agency/CBO
made major
contributions
toward SVORI
programming.

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SA A N

D SD

SA A

D SD

SA A

D SD

SVORI PD Interview 3/06

54. Please complete the table below, indicating whether you strongly agree (SA), agree (A), neither agree nor
disagree (N), disagree (D), or strongly disagree (SD) with the following statements about the groups below.
We have encountered
resistance from this group as
we implemented SVORI.
a.

Support for SVORI from this
group has been strong.

Top administrators at the prerelease facilities

b. Supervisors at the pre-release
facilities

Line staff at the pre-release facilities

d. Top administrators at the postrelease supervision agency
e.

Supervisors at the post-release
supervision agency

Line staff at the post-release
supervision agency

Members of the community to
which SVORI participants are
returning

55. Please indicate whether you strongly agree (SA), agree (A), neither agree nor disagree (N), disagree (D), or
strongly disagree (SD) with each of the following statements about your SVORI program:
a.

Information sharing about specific offenders across partnering
agencies has improved as a result of SVORI.

b. Communication across partnering agencies has improved as a
result of SVORI.

d. Partnering agencies have created common goals related to reentry
as a result of SVORI.

SVORI is a collaborative effort among different agencies.

The original SVORI partnering agencies are still very involved in
SVORI.

The culture within your SVORI grantee agency (e.g., DOC or DJJ) is
supportive of reentry programs in general.

h. The culture within your SVORI grantee agency is supportive of
SVORI.

The current political climate in your community is favorable to
reentry programming in general.

Support for SVORI from the state legislature has been strong.

Partnering agencies have developed a common vision of reentry as
a result of SVORI.

u. Support for SVORI from the executive branch of the state
government has been strong.

SVORI PD Interview 3/06

Sustainability and Lessons Learned
The final set of questions addresses program sustainability, local evaluation efforts, technical assistance, and
lessons learned.
56. Since you received your original SVORI grant, has your
SVORI program received funding from any of the following
sources? (Please check all that apply.)

57. Are there other reentry initiatives (besides SVORI) under
way in your state?
58. Are you planning to continue any elements of your SVORI
program once SVORI funds are no longer available?
a.

[If yes to 58] Which elements are you planning to
retain?

b. [If yes to 58] Are you planning to expand your
program?
b1. [If yes to 58b] In which of the following ways are
you planning to expand your program? Please
check all that apply.

Supplemental SVORI funds from the
Federal government
Funds other than SVORI funds from
the Federal government
Funds from state agencies other than
your SVORI grantee agency (e.g.,
DOC or DJJ)
Funds (additional or reallocated) from
your SVORI grantee agency
Funds from local government(s)
Funds from non-profit, not-for-profit,
or other private organizations
Other (please specify in the box at left)
Y

N
[skip to 58c]

Steering committee
Other partnerships formed through
SVORI
Staff hired through SVORI
Curriculum developed through SVORI
Service coordination approach
Approach for screening offenders for
eligibility
Specific pre-release services enhanced
through SVORI
Specific post-release services
enhanced through SVORI
Other (please specify in the box at left)
Y

Expand pre-release programming to
additional facilities
Expand post-release programming to
additional communities
Expand offender eligibility criteria
Offer more pre-release services
Offer more post-release services
Lengthen the duration of the prerelease phase
Lengthen the duration of the postrelease phase
Hire more staff
Other (please specify in the box at left)

SVORI PD Interview 3/06

[If no to 58] What are the main reasons that you are
not planning to continue your SVORI program?
(Please check all that apply.)

59. In order to take reentry programming (not just SVORI
programming) “to scale” in your state (i.e., provide
comprehensive reentry services to all returning offenders in
the state), which factors are necessary in addition to state
or local funding for reentry programming? Please rank the
top three areas by putting a “1” next to what you consider
to be the most important factor, “2” next to the second
most important, and “3” next to the third most important.
(Please rank only three.)

Insufficient funding
Lack of support from your SVORI
grantee agency (e.g., DOC or DJJ)
Lack of support from other partnering
agencies
Too many barriers to program
implementation/operation
Insufficient numbers of eligible
participants
Program model was not viewed as
successful
Other (please specify in the box at left)
Rank
____ Support from elected state officials
____ Support from top administration at
DOC/DJJ
____ Support from other partnering
agencies
____ Support from the community
____ An effective model for service
coordination
____ An accessible, easy-to-use
management information system
(MIS) containing detailed information
on offenders
____ Policies that make reentry
programming part of the agency’s
standard operating procedure
____ Other (please specify in the box at
left)

60. Please indicate whether your SVORI partnership has engaged in the following sustainability strategies.
a.

Held sustainability planning meetings

b. Assessed progress achieved compared with original
goals

d. Developed a sustainability plan

Extended MOAs with partnering agencies

Sought out other partnering agencies

Pursued additional federal funding

h. Pursued additional state funding

Pursued additional funding from local sources

Pursued additional funding from private funding
sources

Reallocated resources within your SVORI grantee
agency (e.g., DOC or DJJ) in order to continue SVORI

Assessed resource needs

SVORI PD Interview 3/06

Reallocated resources across the partnering agencies
in order to continue SVORI

m. Cross-training of staff

n. Other (please specify: ________________________
__________________________________________)

[If yes] Have any reports been produced from your
local evaluation?

b. [If yes] Have you communicated the results of your
local evaluation/analyses to local, state, or federal
policy makers?

61. Has your program conducted a local evaluation or made
an attempt to document the success of the program in
affecting offender outcomes such as recidivism?
a.

[If yes] For each outcome below, please indicate whether the outcome was evaluated in your local
evaluation, and, for each outcome that was evaluated, whether the analyses demonstrated a positive
program effect.
[If yes] Did the analyses
demonstrate a positive program
Was the Outcome
Outcome
effect?
Evaluated?
c1. Service utilization
Y
N
Y
N
c2. Recidivism
Y
N
Y
N
c3. Substance use
Y
N
Y
N
c4. Physical or mental health
Y
N
Y
N
c5. Educational attainment
Y
N
Y
N
c6. Employment
Y
N
Y
N
c7. Housing
Y
N
Y
N
c8. Family unification/support
Y
N
Y
N
c9. Community integration
Y
N
Y
N
c10. Other (please specify: _________)
Y
N
Y
N

62. Has your program engaged in communication/public
relations designed to convey information about the
program to the public?

63. For each of the following types of technical assistance (from the SVORI technical assistance provider),
please indicate whether you needed it, whether you received it, and if you received it, how helpful it was
(very helpful, somewhat helpful, not at all helpful).
Did you need the
assistance?

Did you receive
the assistance?

Y (Yes)
N (No)

Type of Assistance

Y (Yes)
N (No)

[If yes] How helpful
was the assistance?
V (Very helpful)
S (Somewhat helpful)
N (Not at all helpful)

a. Assistance with federal fiscal reporting

b. Assistance with performance
measurement (GPRA) reporting

c. Assistance forming a steering committee

d. Assistance with staff training

SVORI PD Interview 3/06

Did you need the
assistance?

Did you receive
the assistance?

Y (Yes)
N (No)

Type of Assistance

Y (Yes)
N (No)

[If yes] How helpful
was the assistance?
V (Very helpful)
S (Somewhat helpful)
N (Not at all helpful)

e. Assistance with evidence-based program
selection

f. Assistance with risk/needs assessments

g. Assistance with substance use
programming

h. Assistance with mental health
programming

i. Assistance with employment
programming

Assistance with housing programming

Assistance with family/community
integration programming

Local evaluation assistance

m. Other assistance (please specify:
__________________________________
___________________________)

n. Other assistance (please specify:
__________________________________
___________________________)

o. Other assistance (please specify:
__________________________________
___________________________)

64. What is the key component of your SVORI program that you think has made the biggest difference for
program participants?
We may post your response on your program’s profile on the SVORI Multi-Site Evaluation website. Please
check here if you do not want your response posted:

SVORI PD Interview 3/06

65. What components of your SVORI program did not appear to work?
We may post your response on your program’s profile on the SVORI Multi-Site Evaluation website. Please
check here if you do not want your response posted:

66. What have been the most significant organizational or systems-level changes as a result of SVORI?
We may post your response on your program’s profile on the SVORI Multi-Site Evaluation website. Please
check here if you do not want your response posted:

Thank you very much for taking the time to complete this survey.
If we need to follow up on any of the responses, who should we contact?
Name: _________________________
Phone No.: _____________________
Email address: __________________
In order to update our records, please provide the contact information for the individual responsible for your
program’s local evaluation (if applicable).
Name: ________________________
Phone No.: ____________________
Email address: _________________

Please make a photocopy of this survey and mail the original to RTI by March 31, 2006, using the Federal
Express mailing label. If you have misplaced the label, please contact Mark Pope at (919) 485-5701.

2007 Sustainability E-mail Survey for all Project Directors
1. Is your agency (DOC or DJJ) continuing any activities that were initially created
with SVORI funding (e.g. additional staff, expanded programming)?
a. Please indicate whether any aspects of the following core elements
initiated or enhanced through SVORI are being continued [a) screening
and assessment processes; b) case management; c) coordination of
services between pre-release and post-release; d) specific programs or
services]
b. Are you continuing all of the activities that were initiated or enhanced
through SVORI?
2. Is your agency currently implementing other reentry components or initiatives
that were not created with SVORI funding?
3. Funding for reentry activities originally created with SVORI funding (SVORIspecific activities)…
a. Has your agency reallocated funds internally to continue all or some
SVORI-specific activities?
b. Has your state legislature provided a direct allocation (SVORI-specific
line item) to your agency to continue all or some SVORI-specific
activities?
4. Funding for other reentry activities or initiatives…
a. Has your agency reallocated funds internally to support other (nonSVORI) reentry activities or initiatives?
b. Has your state legislature provided a direct allocation (reentry-specific line
item) to implement other (non-SVORI) reentry activities or initiatives?

6/07/05

SVORI Implementation Assessment Site Visit Protocol for Impact
Evaluation Sites (Spring/Summer 2005)

Overview and Purpose
The site team will conduct an implementation assessment site visit to each impact
evaluation site in the spring and summer of 2005 (June-August). For multi-program
sites, the site visit will focus on only the programs involved in the impact evaluation.
The primary purpose of the site visits is to get the perspective of a variety of key
stakeholders on several important implementation assessment topics. The secondary
purpose is to identify and resolve any problems related to impact data collection.
The site visits will last approximately 2-3 days and will consist of semi-structured
interviews with key stakeholders. The site team will use the information gathered during
the site visit to produce a site visit report.
This protocol describes pre-site visit activities, outlines the key stakeholders to be
interviewed and the topics to be covered in each interview, lists the questions to be
asked, and provides guidance on the site visit report generated from the site visit.

Pre-Site Visit Activities
The site team should contact the Project Director to convey the purpose of the site visits
and to identify potential dates for the site visit. The site team should send the PD a
description of the key stakeholder roles (using their prior knowledge of the site to
customize the list) and a copy of the “site visit topics by stakeholder” table (both of
which are provided in the following section), so that the PD can help identify the
individuals with whom you need to meet and customize the topics to be covered during
each interview. When discussing potential interviewees, please be sure to discuss with
the PD which types of “line staff” can most appropriately answer the questions on
treatment/supervision of non-SVORI (comparison) offenders. If the staff that provide
treatment/supervision to SVORI offenders do not have an understanding of business as
usual, work with the PD to identify similar line staff who work with non-SVORI offenders.
Ideally, the PD will schedule the actual interviews for the site team. The PD should
allow 2 hours for your meeting with him/her, 30 minutes for the local evaluator and key
1

6/07/05
research staff meetings, and approximately 1 hour for the other interviews. The
interview schedule should also factor in any necessary driving time for the site team,
particularly if remote facilities are to be visited (in situations in which Lucinda has
indicated a need for troubleshooting interviewing issues in a certain facility). The
interviews can either be individual or in small groups of related stakeholders. The PD
does not need to attend the meetings (and in fact our preference would be for the PD
not to attend the meetings in order to ensure more candid responses). The site team
should plan to have the first meeting of the site visit be with the PD, and, if possible,
should schedule a debriefing meeting with the PD after the other interviews have taken
place.
Although it can be advantageous to schedule the site visit to coincide with a prescheduled meeting of relevant stakeholders (such as a steering committee meeting),
the disadvantage of this approach is that it becomes difficult for the site team to have
detailed discussions with each stakeholder group (particularly if other business needs to
be conducted or if the stakeholders are expected to wait around while the site team
interviews other stakeholders). If the PD suggests this approach, be sure to talk
through these details ahead of time, in order to ensure that you are able to have oneon-one (or small group) meetings with the stakeholders whom you need to interview.
Once the site visit and the individual meetings have been scheduled, the site team
should cut and paste from the list of questions in the “Interview Guide” section of this
protocol to create a customized interview guide for each stakeholder or group of
stakeholders (be sure to use “paste special” and select “unformatted text” in order to
avoid the item numbers from being automatically renumbered). The determination of
the specific questions to include in each customized interview guide will be based on
the “site visit topics by stakeholder” table and the site teams’ understanding of the
stakeholders’ ability to cover the specified topics.1 This step may take some detailed
discussions with the PD regarding the individuals with whom he/she has scheduled
meetings (you will want to maximize your limited time on site, requiring a good
understanding of the individuals with whom you will be meeting). The site team should
bring the customized interview guides on site and use the guides to facilitate the
interview and to take notes on the stakeholders’ responses. The site team will also
1

As described in a later section, when you meet with service providers (line staff and agency heads), for the
questions on pre- and post-release services available, you only need to discuss the service type that they are involved
in delivering.

6/07/05
need to bring several copies of the informed consent forms developed for the site visits
to pass out to the interviewees prior to the interview.
Also prior to the site visit, the site team should review the site’s PD survey, the previous
site visit information entered in the database, and any other relevant information about
the site. In addition, the site team should talk to Lucinda and the field supervisor
assigned to the site to get a general understanding of how data collection is going at the
site and (if necessary) to ask any questions you may have about the “site problems”
spreadsheet circulated by Lucinda on 5/27/05.

Key Stakeholders to be Interviewed
While on site, the site team will conduct either individual or small group interviews with
the following key stakeholders:
o SVORI Project Director
o Person in charge of programs and/or services at the grantee agency
o This would be the highest level person (e.g., the director of programs and
services) at the grantee agency (e.g., DOC, DJJ) who knows, in broad terms
what is happening in the pre-release component of SVORI, as well as prerelease services for comparable, non-SVORI offenders2
o Pre-release case managers (line staff)
o We are interested in talking with a subset of the actual staff members who
provide pre-release case management to SVORI participants. If numerous
facilities are involved, try to identify staff from the facilities that serve the largest
number of SVORI participants.
o If the SVORI case managers only provide case management to SVORI
participants and do not know what comparable, non-SVORI offenders receive,
you will need to try to identify the people that provide pre-release case
management to comparison offenders3
2

Ideally, this will be the specific comparison group that we are using for the site in the impact evaluation.

Pre-release line staff who only work with comparison offenders will only be asked questions 11, 18, 20-21, and 36.

6/07/05

o Pre-release service providers (line staff) (if different from pre-release case
managers)
o We are interested in talking with a subset of the actual staff members who
provide pre-release services to SVORI participants. If numerous facilities are
involved, try to identify staff from the facilities that serve the largest number of
SVORI participants.
o If the SVORI service providers only provide services to SVORI participants and
do not know what comparable, non-SVORI offenders receive, you will need to
try to identify the people that provide pre-release services to comparison
offenders4
o Head of the agency that provides post-release supervision
o We want to talk to the highest level person at the agency (this may be the
agency head or it may be another high-level staff member) who knows, in
broad terms, what is going on in terms of the post-release supervision provided
to SVORI and comparison offenders
o Post-release supervision officers (line staff)
o We are interested in talking with a subset of the actual staff members (such as
parole officers) who provide post-release supervision to SVORI offenders. If
numerous counties are involved, with specific supervision officers for each
county, try to identify staff from the counties that serve the largest number of
SVORI participants.
o If the SVORI supervision officers only supervise SVORI participants and do not
know what comparison offenders receive, you will need to try to identify some
supervision officers that supervise comparison offenders5
o Head of the agency that provides post-release case management (if different
from supervision agency or if there is no supervision agency)
o In some sites, this will be the same agency that provides post-release
supervision; once again, we want to talk to the highest level person at the

Pre-release line staff who only work with comparison offenders will only be asked questions 11, 18, 20-21, and 36.
Post-release line staff who only work with comparison offenders will only be asked questions 12, 18, 20-21, and
36.
5

6/07/05
agency (this may be the agency head or it may be another high-level staff
member) who knows, in broad terms, what is going on in terms of the postrelease case management provided to SVORI and comparison offenders.
o Post-release case managers (line staff) (if different from supervision agency or if
there is no supervision agency)
o We are interested in talking with a subset of the actual staff members who
provide post-release case management to SVORI offenders. If numerous
counties are involved, with separate case managers in each county, try to
identify staff from the counties that serve the largest number of SVORI
participants.
o If the SVORI case managers only provide case management to SVORI
participants and do not know what comparison offenders receive, you will need
to try to identify the people that provide post-release case management to
comparison offenders6 if case management is not provided by the parole
officers
o Heads of the key agencies that provide post-release services
o We want to talk to the highest level person at each agency (this may be the
agency head or it may be another high-level staff member at the agency) who
knows, in broad terms, what is going on in terms of the post-release services
received by SVORI and comparison offenders. If numerous agencies are
involved, try to identify the ones that are most involved in serving SVORI
participants (i.e., the services utilized by most SVORI participants or agencies
that exclusively or primarily serve SVORI participants).
o If the “SVORI” agencies do not provide services to comparison offenders, try to
identify similar agencies that provide services to comparable, non-SVORI
offenders
o Post-release service providers (line staff)
o We are interested in talking with a subset of the actual service providers who
work with SVORI participants in the post-release phase. If service providers
from numerous agencies are involved, try to identify the agencies that are most

Post-release line staff who only work with comparison offenders will only be asked questions 12, 18, 20-21, and
36.

6/07/05
involved in serving SVORI participants (i.e., the services utilized by most
SVORI participants or agencies that exclusively or primarily serve SVORI
participants).
o If particular agencies do not provide services to comparison offenders, try to
identify service providers from similar agencies who work with comparable,
non-SVORI offenders.7
o Local evaluator
o If the site is conducting a local evaluation, we would like to talk with the lead
local evaluator
o MIS/research staff
o If you work with any MIS or research staff at the site who can speak to issues of
caseflow, we would like to talk with these individuals.
o Other integral staff
o If there are any other staff that are integral to the initiative and who can speak
to the topics to be covered in the site visit protocol (i.e., individuals who do not
fit into one of the categories above but who are integral to the project, such as
boundary spanners), we would like to talk with these individuals.
o Facility contacts
o In addition to the interviews conducted for the implementation assessment,
while on site the site team should try, if feasible, to visit the pre-release facilities
identified by Lucinda as having issues that need to be resolved.

The topics to be covered during the interviews are summarized in the “topics by
stakeholder” table below. Once again, the actual interview topics to be covered with
each type of stakeholder will be customized for each site, depending on the individual’s
knowledge of the topic to be covered. The topics to be covered for “other integral staff”
are not identified in the table because they will be site specific.

Post-release line staff who only work with comparison offenders will only be asked questions 12, 18, 20-21, and
36.

6/07/05

Facility Contacts

Key Research Staff

Local Evaluator

Post-release Service
Providers

Post-release Service
Agency Heads

Post-release Case
Managers

Post-release Case
Management Agency
Head

Post-release Supervision
Officers

Post-release Supervision
Agency Head

Pre-release Service
Providers

Manager

Pre-release Case

Programs Division Head

Grantee Agency

Approach for site visit

Implementation Assessment Topics for Site Visits

Case Flow
Perceptions regarding why
case flow expectations have/
have not been met
Strategies for overcoming
enrollment barriers

3
3

3
3
3

Program Focus
Degree of “targeting”
Rationale for program focus
Goals of program

3
3
3

Services and Components
Pre-release services/
components
Post-release services/
components
Differences in service provision
between SVORI and nonSVORI offenders
Impact of SVORI on service
provision

Service Coordination
Strategies for facilitycommunity linkages
Impact of SVORI on service
coordination
Barriers/solutions related to

Facility Contacts

Key Research Staff

Local Evaluator

Post-release Service
Providers

Post-release Service
Agency Heads

Post-release Case
Managers

Post-release Case
Management Agency
Head

Post-release Supervision
Officers

Post-release Supervision
Agency Head

Pre-release Service
Providers

Manager

Pre-release Case

Programs Division Head

Grantee Agency

Approach for site visit

6/07/05

service coordination

Program Coordination
Organizational structure of
program
Partnering agencies
Approach for inter-/intraagency coordination
Organizational context

3
3
3

3
3

3
3
3

Sustainability
Perceptions of support for
SVORI
Sustainability strategies
Other reentry initiatives
Perceived success of program
Local evaluation results
Other Topics
Data collection issues

3
3

*Service providers only need to be asked about the specific services they deliver.

6/07/05

Interview Guide
This section contains the questions designed to elicit information regarding each topic of
interest. As stated above, you will not be asking all questions of all stakeholders (you
will use the “topics by stakeholder” table to create a customized interview guide for each
stakeholder). As referenced in the key stakeholder descriptions above, line staff who
only work with non-SVORI offenders will only be asked questions 11 (pre-release line
staff), 12 (post-release line staff), 18, 21-22, and 36. In addition, these questions will
need to be slightly reworded, as shown in the footnotes for the specific items affected.
For all respondents who are affiliated with the SVORI program, when asking the
questions please use the actual name of the SVORI program at the site, rather than
saying “the program” or “SVORI”.
As you will notice, some of the questions were already asked in the PD survey and are
therefore not earmarked to be covered with the PD during the site visit. We do,
however, want to get other key stakeholders’ opinions on the topic. For questions that
were covered in the PD survey, the PD’s responses will be imported into the relevant
items below (Mark will generate an Interview Guide for each program and post this on
the project website). These responses will assist the site team in understanding what is
happening at the site. At the conclusion of the site visit, the site team might want to
discuss any inconsistencies for these items with the PD to try to reconcile major
differences.
Finally, please note that one of the goals of the site visit is to get at the consistency of
stakeholder opinions, so please try to ask the earmarked questions of all relevant
stakeholders (even if you feel that you have learned the “right” answer from a previous
interviewee).
Informed Consent
Be sure to hand the interviewee the informed consent form prior to beginning the
interview. You do not need to read the form (and it is not designed for the respondent
to sign it), but you should say “This form provides some information about the study and
the interview. Please note that this interview is voluntary and that you can choose to
not answer any question. We will use the information you provide to generate a site
visit report, but we will not attribute responses to individual respondents.”

6/07/05
Case Flow
Perceptions regarding why case flow expectations have/have not been met
1. Do you feel that the program has met its case flow expectations? [import PD Survey
item 96]
o (if yes) How do you think the program has achieved its case flow expectations?
(probes: In many programs, the actual case flow has been much lower than
expected, so we are really interested in learning how some programs have been
able to achieve estimated case flow. In your opinion, what strategies or
situations have made this possible?)
o (if no) Why do you think the program has not achieved its case flow
expectations? [import PD survey item 99] (probes: In many programs, the actual
case flow has been much lower than anticipated, so we are really interested in
learning more about this issue.
o What do you see as the major issue(s)?
o How have the program’s eligibility criteria affected the number of people
that are enrolled?
o How has the quality or setup of the agency’s management information
system (MIS) or offender tracking system affected enrollment?
o How have your state’s incarceration patterns or sentencing policies
affected enrollment?
o How has the amount of time it took the program to get up and running
affected enrollment?
o How has staff recruitment or retention affected enrollment?
o How has agency and facility staff “buy-in” affected enrollment?
o [if applicable] How has the referral process affected enrollment?
o [if applicable] How has the voluntary nature of the program affected
enrollment)?
Strategies for Overcoming Enrollment Barriers
2. What strategies has the program used to overcome enrollment barriers?
o (if any) Which strategies have been the most effective?
10

6/07/05
Program Focus
Degree of “Targeting”
3. Would you say that the program primarily focuses its resources and efforts on
working with the offender prior to release, after release, or emphasizes pre- and
post-release equally? [import PD survey item 16]
4. Is the program primarily serving the general “serious and violent” offender
population or targeting a subset of offenders with specific service needs? [import PD
survey item 19]
o (if subset) What type of population would you say the program is primarily
targeting?
5. Would you classify the program’s service provision as general, in that it attempts to
provide all needed services for participants, or targeted, in that it focuses on a
specific service or small set of specific services? [import PD survey item 20]
o (if targeted) What type of service would you say the program primarily focuses
on?
6. Is the post-release phase of the program run primarily by a government agency or a
private agency? [import PD survey item 21]
o What is the government or private agency that is primarily responsible for the
post-release component?
7. Would you say the program is using SVORI funds primarily to fill service gaps,
expand existing services, or start a new program? [import PD survey item 22]
8. What would you say are the top 3 programmatic areas on which the program
focuses its resources and efforts: [import PD survey item 23]
o employment/vocational training
o physical health
o mental health
o substance abuse
o family support/unification
o community integration
11

6/07/05
o education and skills building
o or something else?
Rationale for Program Focus
9. What was the rationale behind the current focus of the program, in terms of the
population that is served and the main components of the program? (probe: What
factors led the lead agency/steering committee to focus on this specific population
and to offer the current components of the program?)
Goals of Program
10. What would you say are the top three goals of the program?
Services and Components
Pre-Release Services and Components
•

For the site teams’ reference, general information about pre- and post-release
services and components is covered in the PD interview in items 27-58 and 60-75.
PD survey responses to these items will be provided in a supplementary worksheet
for you. It might be helpful for you to review the answers to this survey ahead of
time to have a better understanding of the types of services that are received by any
SVORI offenders. However, please note that the purpose of questions 11 and 12
below is to obtain more detail regarding what specifically is being delivered through
that service/component type.

11. I am going to list several categories of program services and components and I
would like to know whether the service or component is available at all for SVORI or
non-SVORI offenders8 prior to release. Then, for the services and components
that are available, I would like to know what specifically is being offered prior to
release. For example, if you indicate that substance abuse treatment is available, I’d
like to know what types of treatment components or programs are offered.
o As you read each service/component type in the table below, check the ones

For line staff who only work with non-SVORI offenders, please word the question as: “…I would like to know
whether the service or component is available at all for the offenders with whom you work prior to release.”

6/07/05
that the respondent indicates are available at all for SVORI or non-SVORI
offenders prior to release. Then, go back to the services/components that are
checked and ask about what, specifically, is offered for SVORI participants and
for non-SVORI participants (and record this information in the “specifics”
columns).
o For line staff service providers, as you read the service/component types, find
out which ones they feel comfortable providing the specifics for and only cover
these topics in detail.
Pre-Release
Services/Components

Offered?

Specifics for SVORI

Specifics for non-SVORI

Services

Risk assessment

Needs assessment

Treatment plan/release
plan

Substance abuse
treatment

Mental health services

6/07/05
Pre-Release
Services/Components

Offered?

Medical services

Dental services

Educational/GED/tutoring/
literacy

Vocational training

Employment referrals/job
placement

Resume and interviewing
skills development

Work-release

Specifics for SVORI

Specifics for non-SVORI

6/07/05
Pre-Release
Services/Components

Offered?

Cognitive skills
development/behavioral
programming

Life skills training

Legal assistance

Assistance obtaining
identification (e.g., driver’s
license, social security
card)

Assistance obtaining
benefits and completing
applications (e.g.,
Medicaid, disability
benefits)

Financial
support/emergency
assistance (e.g., housing,
clothing)

Domestic violence
services (e.g., victim
and/or perpetrator)

Specifics for SVORI

Specifics for non-SVORI

6/07/05
Pre-Release
Services/Components

Offered?

Parenting skills
development

Family reunification

Anger
management/violence
counseling

Peer support groups

1-on-1 mentoring

Housing placements or
referrals

Specifics for SVORI

Specifics for non-SVORI

Components

Reentry courts

6/07/05
Pre-Release
Services/Components

Offered?

Sanctions

Rewards

Programs such as animal
training, Habitat for
Humanity, community
beautification, community
service, Weed & Seed,
restitution, victim
mediation, victim
awareness

Restorative justice
activities

Community Accountability
Panels9

Offender-specific reentry
teams10

Specifics for SVORI

Specifics for non-SVORI

Community Accountability Panels are a group of agency and/or community members who meet regularly to
review the status of returning offenders; the offender appears before this board to have his or her case reviewed, and
the panel makes recommendations
10
Offender-specific reentry teams are groups consisting of agency representatives and/or community members; the
team composition is unique to each individual offender. The team meets to review the offender’s progress and make
recommendations.

6/07/05
Pre-Release
Services/Components

Offered?

Video-conferencing
technology

Pre-release curriculum

Routine involvement of
family members

Routine involvement of
victims

Routine involvement of
law enforcement

Routine involvement of
former prisoners

Mentoring

Any other services or
components?

Specifics for SVORI

Specifics for non-SVORI

6/07/05
Post-Release Services and Components
12. I am going to list several categories of program services and components and I
would like to know whether the service or component is available at all for SVORI or
non-SVORI offenders after release.11 Then, for the services and components that
are available, I would like to know what specifically is being offered after release.
For example, if you indicate that substance abuse treatment is available, I’d like to
know what types of treatment components or programs are offered.
o Please note that the services and components covered in question 12 are
identical to those covered in question 11 except the list does not include work
release or pre-release curriculum and includes transportation and post-release
supervision
o As you read each service/component type, check the ones that the respondent
indicates are available at all for SVORI or non-SVORI offenders after release.
Then, go back to the services/components that are checked and ask about
what, specifically, is offered for SVORI participants and for non-SVORI
participants (and record this information in the “specifics” columns).
o For line staff service providers, as you read the service/component types, find
out which ones they feel comfortable providing the specifics for and only cover
these topics in detail.

Post-Release
Services/Components

Offered?

Specifics for SVORI

Specifics for non-SVORI

Services

Risk assessment

Needs assessment

6/07/05
Post-Release
Services/Components

Offered?

Treatment plan/release
plan

Substance abuse
treatment

Mental health services

Medical services

Dental services

Educational/GED/tutoring/
literacy

Vocational training

Specifics for SVORI

Specifics for non-SVORI

6/07/05
Post-Release
Services/Components

Offered?

Employment referrals/job
placement

Resume and interviewing
skills development

Cognitive skills
development/behavioral
programming

Life skills training

Legal assistance

Assistance obtaining
identification (e.g., driver’s
license, social security
card)

Assistance obtaining
benefits and completing
applications (e.g.,
Medicaid, disability
benefits)

Specifics for SVORI

Specifics for non-SVORI

6/07/05
Post-Release
Services/Components

Offered?

Financial
support/emergency
assistance (e.g., housing,
clothing)

Transportation

Domestic violence
services (e.g., victim
and/or perpetrator)

Parenting skills
development

Family reunification

Anger
management/violence
counseling

Peer support groups

Specifics for SVORI

Specifics for non-SVORI

6/07/05
Post-Release
Services/Components

Offered?

1-on-1 mentoring

Housing placements or
referrals

Specifics for SVORI

Specifics for non-SVORI

Components

Post-release supervision

Reentry courts

Sanctions

Rewards

Programs such as animal
training, Habitat for
Humanity, community
beautification, community
service, Weed & Seed,
restitution, victim
mediation, victim
awareness

6/07/05
Post-Release
Services/Components

Offered?

Restorative justice
activities

Community Accountability
Panels12

Offender-specific reentry
teams13

Video-conferencing
technology

Routine involvement of
family members

Routine involvement of
victims

Specifics for SVORI

Specifics for non-SVORI

Community Accountability Panels are a group of agency and/or community members who meet regularly to
review the status of returning offenders; the offender appears before this board to have his or her case reviewed, and
the panel makes recommendations
13
Offender-specific reentry teams are groups consisting of agency representatives and/or community members; the
team composition is unique to each individual offender. The team meets to review the offender’s progress and make
recommendations.

6/07/05
Post-Release
Services/Components

Offered?

Routine involvement of
law enforcement

Routine involvement of
former prisoners

Mentoring

Any other services or
components?

Specifics for SVORI

Specifics for non-SVORI

Differences in Service Provision Between SVORI and Non-SVORI
13. Would you say that SVORI participants are prioritized over non-SVORI offenders for
any services or components?
o (if yes) How are SVORI participants prioritized?
14. Would you say that SVORI participants receive a higher quality of service compared
to non-SVORI offenders for any services? (probes: Are there differences in terms of
the follow-through of referrals between SVORI and non-SVORI offenders? Are there
differences in the intensity or length of the services delivered to SVORI and nonSVORI offenders? Are there differences in the degree of aftercare received by
SVORI and non-SVORI offenders? Are there differences in the qualifications of the
staff who deliver services to SVORI and non-SVORI participants?)
o (if yes) In what ways would you say that SVORI participants receive higher

6/07/05
quality services?
15. Are there differences between SVORI and non-SVORI offenders in terms of who is
responsible for delivering the services? (probes: For example, does involvement of
faith-based or other community-based service organizations differ for SVORI and
non-SVORI offenders?)
Impact of SVORI on Service Provision
16. Are there any other ways in which service provision has been influenced by SVORI?
(Note that there is a separate set of questions on service coordination, so
coordination issues do not need to be covered here)
17. Are there any other differences in service provision between SVORI and non-SVORI
offenders?

Service Coordination
Strategies for Facility-Community Linkages
18. What is the program’s strategy for facilitating the “inside-outside” linkage14 (probes:
What are you doing to facilitate an individual transitioning from incarceration to
home? In what ways are community and post-release supervision agencies
“reaching in” to the institutions to work with offenders prior to release? How are
facility staff working with and communicating with community and post-release
supervision agencies prior to release?)
Impact of SVORI on Service Coordination
19. How has SVORI funding influenced the way that services are coordinated for
individual offenders? (probe: Has SVORI funding changed the way organizations
work (or work together) to ensure offenders get services they need? Has case

For line staff who only work with non-SVORI offenders, please replace “the program’s strategy” with “the
agency’s strategy”.

6/07/05
management been influenced by SVORI funding? Have approaches such as
continuity of care, development of service provider networks, one-stop shops, or
“wrap-around” provision been implemented or enhanced as a result of SVORI
funding?)
Barriers/Solutions Related to Service Coordination
20. What barriers to effective service coordination has the program15 experienced?
21. What strategies have been implemented to address these barriers?
Program Coordination
Organizational Structure of Program
22. How would you describe the organizational structure of the program, in terms of
program leadership, advisory groups, and committees?
Partnering Agencies
23. What are the specific agencies involved in the program?
Approach for Inter-/Intra-agency Coordination
24. What is the approach for coordination among the agencies involved in SVORI?
25. How has SVORI changed the way agencies coordinate with one another?
26. What is the approach for coordination within your agency, as it relates to SVORI?
27. How has SVORI changed the way staff within your agency coordinate with one
another?
Organizational Context
15

For line staff who only work with non-SVORI offenders, please replace “the program” to “your agency”.

6/07/05

28. How have the goals and expectations of the agencies involved in SVORI influenced
the development of the program?
29. How have the various partners invested in SVORI?
Sustainability
Perception of Support for SVORI
30. How would you describe the level of support for the SVORI program from:
o The grantee agency?
o The post-release supervision agency (if offenders are under parole
supervision)?
o The post-release case management agency (if different from post-release
supervision agency)?
o The post-release service provider agency(ies)?
o The state legislature?
o The community(ies) to which offenders are returning?
Sustainability Strategies
31. What are your plans for SVORI once grant funding ends? [import PD survey item
103]
32. What strategies have been undertaken to sustain the program once federal funding
ends? [import PD survey item 103d]
33. What additional strategies are you considering implementing to sustain the program
once federal funding ends? [import PD survey item 103d]
Other Reentry Initiatives
34. What other reentry initiatives are underway in your state?
o (if any) How do these initiatives relate to SVORI? (prompt: Did they precede
SVORI or come after? How well-integrated are these efforts?)
28

6/07/05

Perceived Success of Program
35. In your opinion, to what extent have the goals and objectives established by the
program been met?
36. In your opinion, how successful has the program16 been at:
o Improving family and community reintegration among participants?
o Reducing substance use among participants?
o Preventing recidivism among participants?
37. What would you say have been the biggest improvements at the system-level as a
result of SVORI funding?
Local Evaluation Results
38. What has your local evaluation shown in terms of the effectiveness of the program?
39. Are the results from your evaluation being used to make any programmatic
changes?
Other Topics
Data Collection Issues
40. [Add questions or discussion starters here for any topics related to data collection
issues the field interviewers are experiencing in specific facilities.]

For line staff who only work with non-SVORI offenders, please replace “the program” with “your agency” and
replace “participants” with “offenders.”

6/07/05

Post Site-Visit Activities
After the site visit, the site team should send an e-mail or letter to the stakeholders with
whom they met, thanking them for their time and providing your contact information in
case they have any questions about the evaluation. The site team may also need to
engage in telephone follow-up for individuals with whom they were not able to meet
while on site, or obtain outstanding information.
The site team should also update the information on “Clients” and “Organizations” in the
database to reflect the most recent information for the site.
Finally, the site team should prepare a site visit report summarizing the information
learned during the site visit. The report should include a list of the individuals (and
roles) interviewed while on site and should be organized using the major headings in
this protocol. The report should be as concise as possible (10-15 pages) and should
not attribute responses to individual stakeholders. A template for the report is included
on the following page. In addition, the site visit reports from the pilot site visits will be
available as a resource.

6/07/05

SVORI – [Site name] Site Visit (dates of visit)

People interviewed -(name-affiliation/title)

Case Flow – (Perceptions regarding why case flow expectations have/ have
not been met, Strategies for overcoming enrollment barriers)

Program Focus – (Degree of “targeting”, Rationale for program focus,
Goals of program)

Services/Components – (Specifics for services and components received
by SVORI and non-SVORI offenders prior to release and after release,
Differences in service provision between SVORI and non-SVORI offenders,
Impact of SVORI on service provision)

Service Coordination – (Strategies for facility-community linkages, Impact
of SVORI on service coordination, Barriers/solutions related to service
coordination)

Program Coordination – (Organizational structure of program, Partnering
agencies, Approach for inter-/intra- agency coordination, Organizational
context)

Sustainability – (Perceptions of support for SVORI, Sustainability
strategies, Other reentry initiatives, Perceived success of program, Local
evaluation results)

Appendix B. Impact Site
Program Descriptions

B-1

Florida DJJ – Study Design

The FL DJJ reentry initiative, Going Home, targets Dade County youth who are committed for a
violent felony or burglary to any of the following facilities: Bay Point Schools, Inc. (has 3
campuses), Everglades Youth Development Center, Florida City Youth Center, Southern Glades
Youth Camp, and Wings for Life South Florida (girls). The youth must be at least 16 years of
age by the time they leave the residential facility and not have a High School diploma or have
completed a GED by the time they leave the residential facility. YLS/CMI scores are also used
to determine program eligibility.
For the comparison group will consist of youth who meet the eligibility criteria1 except for
county of admission / post-release geographical parameters. Comparison youth will release to
either Broward or Palm Beach County.

YLS/CMI scores are not available on juveniles who are not participating in the Going Home
initiative. MAYSI scores are available on all juveniles in the state and the MAYSI provides a
similar measure of functioning to that which is provided by the YLS/CMI. However,
comparison subjects are not being matched to treatment subjects based on their MAYSI scores at
this time.

THE URBAN INSTITUTE 2100 M STREET, N.W. / WASHINGTON D.C. 20037
Meghan Salas
Research Associate II

phone: 202-261-5272
fax: 202-296-2252
e-mail: MSalas@ui.urban.org

DATE: January 12, 2005
SUBJ: Site Design Memo - Indiana
Facilities that are Included:
We are drawing our sample from the following facilities:
1) Chain O’Lakes
2)

Indiana Women’s Prison

Miami

Plainfield

Putnamville

Rockville

Westville

To date, SVORI participants in Indiana are receiving no pre-release programming that is
specific to SVORI participants. Allen County Community Corrections (the grantee
agency) is piloting some pre-release programming in one Indiana facility, but at this point
these services are not offered to all SVORI participants and shouldn’t affect many, if any,
of our sample. Therefore, we do not consider someone officially enrolled in the SVORI
program until they have been released to Allen County Community Corrections.

The Treatment Group:
The treatment group is made up of those persons (both males and females) who are
incarcerated in one of the seven facilities listed above who are returning to Allen County and agree
to participate in the Community Transition Program (CTP) at Allen County Community
Corrections.
Ted Gallmeier from Allen County Community Corrections sends weekly files to Meghan
Salas listing those who are eligible for the reentry program in Allen County and those who have
denied the program. She keeps track of this information and sends those eligible who have not (to
our knowledge) denied the program on to Lucinda for fielding. There is a bit of lag time between
when someone is determined eligible and when ACCC finds out whether or not they have denied
the program. The system is imperfect, but we are doing our best to catch as many of the eligible
participants. If Meghan Salas finds out later that one of the people who she indicated as a SVORI
program participant ends up denying the program, she notifies Lucinda that this person should not
be contacted for a follow-up interview.

Eligibility and the Decision to Participate in the Reentry Program at ACCC
Almost all persons returning from DOC to Allen County are eligible for programming at
ACCC unless they have been terminated from reentry twice, they have a pending suit
against ACCC (4-5 right now), they have severe medical requirements (if they have a
really serious physical or mental handicap), or if they are serving less than a two-year
sentence.
Additionally, Indiana recently signed new legislation in terms of eligibility for CTP.
Now prisoners do not have the choice to select early release. Instead, they are given 10
days after notification (60 days pre-release) to either accept the program or to petition to
the courts to allow them to opt-out with no penalty. If they decline to participate without
getting proper approval from the court, they will lose their earned-time. As a result, the
program is no longer purely voluntary, it is now “voluntarily coerced” according to
ACCC staff. [** This is the information that Meghan Salas receives via email on a
weekly basis from Ted Galmeier]
Stan Pflueger of ACCC mentioned that the new legislation might exacerbate the already
existing disincentive to participate. People who accept the program are, in a sense,
volunteering to do three times at much time on electronic monitoring to get out 60 days
early. Therefore, those with a short release are more likely to decline. The early release
varies by crime. The parole board in almost every case has waived the balance of the
parole time if someone goes through the reentry program.

The Comparison Group:
The comparison group is made up of those persons (both males and females) who are
incarcerated in one of the seven facilities listed above who are returning to Marion County and
match the selection criteria used for the Reentry program at Allen County Community Corrections
(with the exception of post-release county of residence). Also, participants will be excluded if
they are enrolled in the Community Transition Program or the Community Chaplain Program in
Marion County.
Whether or not one of the comparison subjects enrolls in the Community Transition
Program in Marion County or the Community Chaplain Program is not known by UI or the DOC
until this person has been released and we receive the Actual Release Confirmation File from the
DOC. Since this only effects a very small percentage of the total eligible comparison subjects, we
do not anticipate that it will be a big issue. Currently, the procedure is as follows: If Meghan Salas
finds out that an eligible comparison subject from a previous month, has entered the CTP program
in Marion County, she sends Lucinda an update so that this person is not contacted for a follow-up
interview.

THE URBAN INSTITUTE 2100 M STREET, N.W. / WASHINGTON D.C. 20037
Janine M. Zweig, Ph.D.
Senior Research Associate

phone: 202-261-5338
fax: 202-659-8985
e-mail: JZweig@ui.urban.org

TO:
Christy Visher, Pam Lattimore
DATE: January 13, 2005
SUBJ: Iowa design--Revised
The Iowa SVORI project supports a same-same design and serves prisoners who are returning to
Polk County (Des Moines) and who meet the following criteria: (1) being ages 18 to 40 for
males, and 18-48 for females (an increase from original design), and (2) having one year or more
to serve on their sentence at the time of their release. The SVORI (or Keys) programming is a
12-week, in prison program for offenders to participate in as close to their time of release as
possible. It is conducted in the state’s three medium-security facilities: Fort Dodge Correctional
Facility, Fort Dodge (Webster county); Iowa Correctional Institution for Women, Mitchellville,
(Polk county); and Newton Correctional Facility, Newton, (Jasper county). By design, all Keyseligible prisoners are supposed to be transferred from higher-level security facilities to one of
these three facilities (in the case of the male facilities) in order to receive programming before
they are released. Although by design SVORI participants are supposed to be directly released
to community supervision with a dedicated parole officer and intensive case management
services provided by the Des Moines Area Community College, some SVORI participants (about
20 percent) end up being sent by the parole board to work release facilities for three to six
months before release to the community. There are two work-release facilities in Polk County –
one for men and one for women – run by the Fifth Judicial District (the community supervision
agency with the dedicated SVORI parole officer).
Participants who are eligible for KEYS are randomly assigned in those facilities to the program
or comparison group. At the end of the first three cycles of enrollment, the Iowa numbers were
lower than were originally projected and as a result the control group was smaller than we had
hoped. The numbers for the Iowa site were n=171, of which 118 are treatment. This includes
those participants who may have refused participation or those participants that we have missed.
In an effort to increase the numbers of the treatment and control groups the Iowa team
conducted a conference call with site representatives at which time we proposed that the current
period of data collection be extended for another round of KEYS programming (for a total of
five rounds) as opposed to the four rounds we planned for. With the implementation of our
proposal, Round 5 of KEYS programming will be May 30, 2005-August 29, 2005. There will
likely be some lag for people in the treatment and control group for these rounds to be released
and that those paroled to work release rather than directly to the community will experience a
greater lag. The site is working hard to get participants “program ready” meaning the offender
has completed all other treatment provided by the facility that has been identified as a need by
their facility counselor or the parole board before being eligible for KEYS. The age criteria have
been increased to 40 for males and 48 for females.

Our site visit took place from March 30th through April 2nd and was very productive. The
process for offender flow into the study was designed and the interviewing logistics were
articulated. Specific decisions were made in collaboration with the Iowa Department of
Corrections (DOC), the Criminal and Juvenile Justice Planning (CJJP) agency, and the Fifth
Judicial District . We have also met with the parole board, DOC officials, the local
evaluators/MIS staff (these are government agency people, not independent evaluators), the Fifth
Judicial District directors (the supervision agency), and have visited the three KEYS facilities
and the two work-release facilities.

Kansas Adults
The Shawnee County Reentry Program (SCRP) is a voluntary program serving adults
who are incarcerated in any correctional facility and who are returning to Shawnee
County (Topeka). All males who are incarcerated in facilities other than the Lansing
Correctional Facility (LCF) are transfered to LCF before beginning the program. There
is only one female facility in Kansas, which is located in Topeka (TCF). Additional
eligibility criteria for SCRP include the following: 18-34 years old; convicted of a
serious, violent, or sex offense; and identified at least 12 months prior to earliest possible
release date (to allow for completion of pre-release programming). Some SVORI
participants will have post-release supervision, while others will be max-outs. We are
interviewing all SVORI participants who are released during the data collection period.
Our original estimate of the number of SVORI participants who would be released
between 9/04 and 11/05 was 88-100 (6-7 per month). However, actual numbers have
been much lower: a total of 11 SVORI participants have been released during the first
five months of data collection.
Our comparison group includes offenders from LCF and TCF who are 18-34 years old,
have been convicted of a serious, violent, or sex offense, and who are returning to
Sedgwick County (Wichita). The DOC does not have a field in their database that
indicates where offenders will reside after release from prison, so we are using pre-prison
county of residence to identify a sample of potential comparison subjects.

Kansas Juveniles
All juveniles returning to one of three judicial districts in south central Kansas who have
a conditional release period of at least six months can participate in the Going Home
Initiative. The three judicial districts to which participants return comprise five counties:
Sedgwick (Wichita), Butler, Elk, Greenwood, and Cowley.* The GHI participants can
come from any of the four juvenile correctional facilities (JCF’s) in Kansas. The four
JCF's serve different populations: Beloit JCF houses female offenders; Larned JCF
houses offenders with serious mental health and/or substance abuse issues; Atchison JCF
houses the youngest offenders; and Topeka JCF houses the most serious offenders (it is
also the largest JCF).
For the purposes of the national evaluation, only GHI participants from two of the four
juvenile correctional facilities, Beloit (females) and Topeka (males), are being
interviewed. The major factors that were taken into account when deciding to include
only two of the four facilities were: [1] there likely would not be enough participants at
the three "specialty" facilities to conduct an adequate study of their populations, and
pooling them involves too much risk of type II error (e.g., there are not enough of any of
them to make an adequate test of moderation of any intervention effect); [2] if we take
the notion of "severe and violent offenders" at face value, only the kids at the fourth
facility fit the definition.
The comparison group includes offenders at Topeka JCF and Beloit JCF who have a
conditional release period of at least six months and are returning to Shawnee (Topeka),
Johnson, or Wyandotte County (Kansas City area).**

*The GHI program used to serve offenders returning to one of five judicial districts in the northeastern and
south central regions of the state. The post-release geographic areas included the Kansas City area in the
northeast and Wichita and surrounding rural areas in the south central region of the state. Around the time
we started data collection, the two GHI staff in the Kansas City area resigned, and the program is no longer
operating in that part of the state. GHI participants returning to the two judicial districts in the northeast
were supposed to account for half of our total sample.
**If and when new GHI staff are hired to work with offenders returning to the northeast region (Johnson
and Wyandotte counties) and the program resumes operations in the Kansas City area, we will interview
offenders returning to the northeast as SVORI participants, and our comparison group will be limited to
offenders returning to the Topeka area.

THE URBAN INSTITUTE 2100 M STREET, N.W. / WASHINGTON D.C. 20037
Janine M. Zweig, Ph.D.
Senior Research Associate

phone: 202-261-5338
fax: 202-659-8985
E-mail: JZweig@ui.urban.org

TO:
Christy Visher, Pam Lattimore
DATE: March 22, 2004
SUBJ: Draft of Maine design
Maine’s SVORI program targets prisoners in all six adult facilities and two youth facilities
ranging in age from 16 to 25 and returning to four counties (Washington, Androscoggin,
Penobscot, and Knox). Initially, the facility staff members refer people to the program based
only on these two criteria, but the SVORI project director conducts a second level screening
checking a referral’s risk of recidivism (high risk) and level of need (moderate to high).
Although the second level of screening is conducted, the SVORI project takes virtually all the
referrals it receives because if a person is in prison in Maine, then they are considered high risk
(based on the very low incarceration rate for the state).
The pre-release SVORI programming (intensive transition planning with reentry specialists and
teams) starts six months before release. Right now, the program has about 60 people enrolled
(up from 23 in December and 44 in January) with 20 having been released. The project director
estimates that 177 adult SVORI participants and 96 juvenile SVORI participants will be released
from prison between June 2004 and June 2005 (information provided January 2004).
The only design option possible for Maine is a same-different design because SVORI
programming occurs in all the facilities of the state. As a result, the ME team and the project
director have identified four counties that are similar to the SVORI counties, from which we can
recruit a returning comparison sample. Specifically, the county pairings are:
1.
2.
3.
4.

Washington County (SVORI) and Piscataquis County
Androscoggin County (SVORI) and York County
Penobscot County (SVORI) and Kennebec County
Knox County (SVORI) and Lincoln County

An analysis of county characteristics shows that these pairings are quite similar in a number of
ways, including: the gender and racial/ethnic breakdown of the population; the proportions of the
population who are high school graduates and who have bachelor’s degrees or more; the median
household income; the percent of population living below poverty; and unemployment rates.
(Specific information can be provided if you would like to review it.) The project director also
believes the county pairings are similar on service availability in the community.
The project director provided us with data from CORIS (their recently initiated data system) that
includes a list of the current prison population with each prisoner’s county of origin (assumed in

most cases to be the county of return), projected release date (between June 04 through June 05),
gender, and current facility location. The information is summarized in the two attached tables.
As you can see, the numbers presented in the tables are lower than the estimated numbers
presented in January. I talked with the project director about this discrepancy and describe the
response below. (Either way, we should scrutinize carefully the enrollment numbers toward the
beginning of the study.)
For adults, the number of prisoners that SVORI will actually enroll and release, and therefore
that we will be able to enroll in the study, should be higher for three reasons: (1) good time
calculation does not show up in CORIS and some people will have good time calculated into
their release date, therefore pushing their release forward into the timeframe when we will be
enrolling sample; (2) a group of people with less than one year sentences (9 months and 1 day),
who do not currently show up in CORIS, will be released into the SVORI counties over the next
year (about 2 to 5 per month according to current SVORI referrals); and (3) the county of origin
designations may be somewhat inaccurate because a number of people have a null county field
and others end up returning to SVORI counties, although they are not from there, to receive
services.
For juveniles, the numbers in the table are totally inaccurate because the information provided
was only for youth with projected release dates in CORIS. Because the youth program is a
treatment program and participants are released based on their individual progress, projected
release dates are generally not entered into CORIS because they are ultimately inaccurate.
Addendum, (01/14/05):
As of February 2005, Cumberland County will be added as a SVORI county and the baseline
enrollment period will be extended through Sept-Oct of 2005.

THE URBAN INSTITUTE 2100 M STREET, N.W. / WASHINGTON D.C. 20037
Meghan Salas
Research Associate II

phone: 202-261-5272
fax: 202-296-2252
e-mail: MSalas@ui.urban.org

DATE: January 12, 2005
SUBJ: Site Design Memo - Maryland

Treatment Group Construction
The Enterprise Foundation is providing reentry services through the REP program. Those
enrolled in the voluntary REP program make up the treatment group for SVORI. The group is
selected from the Metropolitan Transition Center (MTC) in Baltimore. Those eligible must also
be returning to five specific Baltimore zip codes: 21213, 21215, 21216, 21217 and 21218.
Potential REP participants are made aware of the program through REP caseworkers who are
typically ex-offenders. Offenders are excluded if they committed a crime against a child. In a
formal exit interview, Enterprise ensures that caseworkers, and various representatives are
present in MTC to give the inmates information about the program and to try to develop interest
and to get them signed on for participation.
Once an inmate decides to become a part of the REP program (it is voluntary), they are enrolled
in the three pre-release programs available for REP participants as well as are given a thorough
risk and needs assessment by a REP caseworker (assessment takes over 1 hour to complete).
The three courses are Outreach (life skills training), Strive (job readiness training), and a weekly
relapse prevention meeting through Patrick Allen House.
If a REP participant moves out of their original zip code residence the REP team will still
provide services as long as participants stay within the city of Baltimore.

Comparison Group Construction
Those in the comparison group are also housed at MTC, but are returning to other Baltimore zip
codes and are not enrolled in any specific reentry programming. Andy Stritch, the Resource
Coordinator at MTC, uses the same search criteria to identify the comparison group (with the
exception of zip code).

THE URBAN INSTITUTE 2100 M STREET, N.W. / WASHINGTON D.C. 20037
Meghan Salas
Research Associate II

phone: 202-261-5272
fax: 202-296-2252
e-mail: MSalas@ui.urban.org

DATE: January 12, 2005
SUBJ: Site Design Memo - Virginia

**A decision was made in late October to stop doing offender interviews in Virginia due to
a smaller than expected sample size. Administrative data collection and analysis of the
treatment and comparison groups will still continue.

Treatment Group
The treatment group is made up of those incarcerated persons returning to Fairfax County
Virginia from all of the Virginia state prisons. Those eligible are those imprisoned for sexual or
other violent offenses. These individuals sent to the Fairfax Adult Detention Center (jail) where
they begin to receive transition services and finish out their sentence. While there, they begin to
receive services from OAR of Fairfax County.

Comparison Group
Those in the comparison group have similar characteristics to those returning to Fairfax County,
but instead are returning to the City of Alexandria and Arlington County. These individuals are
not receiving reentry programming.

Missouri DOC – Revised Study Design

01/14/05

1. Eligibility: The MO DOC reentry initiative is a mandatory program serving incarcerated
adults with a prior criminal history.
• Prior criminal history can consist of two convictions, one prior incarceration, serving
a minimum mandatory prison term (MMPT) because of a prior incarceration or
dangerous felony, or a conviction of a serious violent offense (NCIC violent and class
A or B). Selection criteria were recently extended to MMPT offenders and violent
offenders without a prior criminal history. The targeted population includes first
releases or violator returns.
• Additional selection criteria are that the adults: be between the ages of 17-34, have a
vocational skills score of 3-5 (unskilled to semi-skilled), have at least 12 months of
incarceration remaining after intake to participate in pre-release
planning/programming, and must be released returning to the community with
supervision required.
The program accepts both male and female prisoners. Males must be returning to specific
zip codes within Kansas City, which are the most populous downtown area, but a fairly
small part of the metro area:
• 64101-64109
• 64123-64133
• 64110-64113
• 64120
Females can be returning to the same Kansas City zip codes or to St. Louis city or
county.
2. Pre-release facilities: Offenders meeting criteria are identified roughly 15 months prerelease and transferred to one of five facilities for SVORI programming 12 months prior
to release. There are 3 male facilities (all located north of Kansas City) and 2 female (one
NE of Kansas City and the other NW of St. Louis):
Male:
• Crossroads Correctional Center (CRCC), a maximum security prison (C-5)
• Western Missouri Correctional Center (WMCC), a medium to maximum security
facility (C-3 & C-5)
• Western Reception Diagnostic Correctional Center (WRDCC), a minimum security
(C-2)
Female:
• Chillicothe Correctional Center (CCC), a minimum to maximum security prison (C-1
to C-5). Female SVORI participants leaving CCC will return to Kansas City for postrelease programming.
• Women’s Eastern Reception Diagnostic and Correctional Center (WERDCC), a
minimum to maximum security (C-1 to C-5). Female SVORI participants leaving
WERDCC will return to St. Louis for post-release programming.

3. SVORI Case Flow: 102 participants were enrolled in the program between April 2003 &
May-early June 2004. The Director of Research & Evaluation (our MIS contact)
indicated that enrollment could increase 25-50% due to the newly extended selection
criteria of MMPT and violent offenders without prior criminal histories [this could also
expand the comparison case flow].
They currently estimate the maximum number of SVORI participants who would be
available for enrollment in our study from 9/04 to 10/05 is 153:
Male: 96
• 12
• 45
• 39

CRCC [maximum security C-5]:
WMCC [medium to maximum security C-3 & C-5]:
WRDCC [minimum security C-2]:

Female: 57
• 12
• 45

CCC [minimum to maximum security prison C-1 to C-5]
WERDCC [minimum to maximum security C-1 to C-5]

It is possible, but unlikely to expand the treatment group to include St. Louis.
4. Comparison group: MO is among the early states participating in Transition from
Prison to Community (TPCI). This will make re-entry business-as-usual, essentially
following a model very close to their SVORI program, beginning in August 04. All
offenders in TPCI facilities will receive both pre- and post-release planning & services.
TPCI is being piloted in 5 institutions – and unfortunately, 2 are the same as the SVORIdesignated facilities: WRDCC (male) and WERCC (female). Unfortunately, WRDCC is
the facility that contributed heavily to prior estimates of comparison case flow.
Nevertheless, the decision was made by RTI/UI to proceed with including WRDCC in
the comparison sample selection, on the grounds that TPCI may not implement as it
intends.
Males:
In Fall, 2004: Applying the same eligibility criteria as for SVORI, the estimated number
of males who would be available for enrollment in our study from 9/04 to 10/05 is not
large (the exact number is still in question as the spreadsheet provided does not sum
properly within rows or columns, and does not square with summary tables provided):
•

48 [if we exclude WRDCC and limit selection to those returning to the same target
areas as for the SVORI group]:
o 7
o 4

CRCC -> Jackson County
CRCC -> St. Louis city/county

o 29
o 8
•

WMCC -> Jackson County
WMCC -> St. Louis city/county

63 [if we expand to the counties abutting the counties in which KC and St. L are
located]:
o 12
CRCC -> Jackson County & 5 surrounding counties
o 4
CRCC -> St. Louis city/county
o 39
WMCC -> Jackson County
o 8
WMCC -> St. Louis city/county

After baselining began, in order to generate a larger comparison sample, we
instructed site representatives to expand the geographic criteria to include Jackson
County.
Females:
Applying the same eligibility criteria as for SVORI, the estimated number of females
who would be available for enrollment in our study from 9/04 to 10/05 is:
•
•

15 (5 from CCC and 10 from WERDCC), if we limit selection to those returning to
the same target areas as for the SVORI group; the MIS contact estimates this could
swell to 23
69 (17 from CCC and 52 from WERDCC), if we count those returning anywhere in
the state

There is no female comparison group in MO.
5. Data: We are receiving monthly lists of both pre- and post-release individuals (so long as
they remain under supervision) from centralized MIS. Glen Meier (Research Analyst) is the
point of contact. The centralized database is the best source for actual release dates, but we
also can verify these with facility contacts. Our information can be updated at the end of each
month.
•

This information is provided in Excel spreadsheets in an automated format.

•

They provide all needed elements: name, DOC#, SSN, DOB, gender, race &
ethnicity (Hispanic/nonhispanic), current facility, date of admission, commission
offense as general NCIC classification and also MO charge code prior commitments,
custody level (which is determined largely by prison they are in), county of
admission, and expected release date

•

They also provide an institutional risk indicator: e.g., how many conduct violations
in past 6 months. They have both old and new salient factor scores used inside
prison. Also, supervising officers do risk/need scales within 30 days post-release.

•

They provide updated home addresses centrally for those in custody as home plans
are updated, and for those in the community under supervision [note that all

participants are under supervision at time of release, but some may not remain under
supervision 6 months post-release].
6. Transfer Agreement for research approval
• The transfer agreement is complete, and has been sent to affected facilities, probation
and parole, and state supervisors.
7. Letter of Support
•

We have a letter of support executed by Gary Kempker, DOC Director.

8. MOUs
•

We have executed a general MOU with the DOC Director (Gary Kempker).

•

A MOU has been executed and signed by each superintendent with customized
information specific to his/her facility.
o Crossroads Correctional Center (CRCC);
o Western Missouri Correctional Center;
o Western Reception Diagnostic Correctional Center;
o Western Reception Diagnostic Correctional Center (WRDCC)
o Chillicothe Correctional Center
o Women’s Eastern Reception Diagnostic and Correctional Center

9. Interviewer requirements (unchanged)
1.
Custody staff are mixed gender, so interviewers can be as well
2.

Cell phones are probably ok, but it would be better if they can be left locked in car

Can use laptops, and also can tape record if desired. Video recording would be a
problem for them

There are private rooms in visiting areas. The visiting areas are not open Monday &
Tuesday (and sometimes Wed) so that might be when most interviews could be
conducted

Study Design – Nevada DOC
Nevada Department of Correction’s reentry program, Going Home Prepared, is designed to
serve adult and juvenile offenders between the ages 14 and 35, both males and females.
Participants must be eligible for parole between 2003 through June 30, 2005. Offense types
targeted include category A and B felons, sex offenders whose victims are adults of the opposite
sex and offenders that have committed drug offenses combined with a history of criminal activity
are included as well. Preference is given to offenders with a history of weapons use, prison
disciplinary incidents, gang affiliation, prior convictions, habitual offending, and other
criminogenic factors. Sex offenders are excluded if their offense is against a child or senior or if
their offense includes ‘perverted things.’ Participants may be released from any of the state’s
prisons but must serve their parole sentence in Clark, Lincoln, Nye, or Esmeralda Counties
(Southern Nevada). All male participants are transferred to Southern Desert Correctional facility
for SVORI programming. Female participants receive programming at Jean Conservation Camp
and Southern Nevada Women's Correctional Facility. Participants are screened for eligibility
through MIS and project staff after which a selection committee reviews offenders’ files to make
final selections six months to a year prior to release. Participation in Going Home Prepared is
mandatory once enrolled and is made a condition of the offender’s parole. Enrollment of the
first participants began in June 2003. It is anticipated that 250 offenders will be served
throughout the duration of the grant.
The comparison group will consist of individuals that meet all eligibility criteria except for the
post-release geographic parameters1 and those screened out of SVORI for additional reasons.2
Comparison subjects may be incarcerated in any of the following facilities: High Desert State
Prison, Indian Springs Conservation Camp, Jean Conservation Camp (women’s), Southern
Desert Correctional Center, Southern Nevada Women's Correctional Center (women’s). In
addition to the above facilities, through the end of September 2004 comparison subjects could
have been incarcerated in the following Northern Nevada facilities: Lovelock Correctional
Center, Northern Nevada Correctional Center, Northern Nevada Restitution Center, Nevada
State Prison, Stewart Conservation Camp, Silver Springs Conservation Camp (women’s), Warm
Springs Correctional Center.

Interviewing was discontinued in Northern Nevada due to small numbers of eligible participants in November
2004.
2
Individuals that are determined to not be eligible due to subjective reasons (e.g., disciplinary history) are excluded
from the comparison group. Individuals that are determined to not be eligible for other reasons (e.g., enemy gang in
yard, not enough time on sentence to participate) are included as comparison subjects.

THE URBAN INSTITUTE 2100 M STREET, N.W. / WASHINGTON D.C. 20037
Rebecca Naser, Ph.D.
Research Associate

pone: 202-261-5509
fax: 202-659-8985
e-mail: RNaser@ui.urban.org

MEMORANDUM
TO: Christy Visher
DATE: January 12, 2005
SUBJ: Ohio Site Design – update1

Facilities:
Experimental and control group sample is being drawn from the following Ohio DRC facilities:
Allen Correctional Institution (Lima)
Pickaway Correctional Institution (Orient)
Richland Correctional Institution (Mansfield)
Ross Correctional Institution (Chillicothe)
Southern Ohio Correctional Facility (Lucasville)
Belmont Correctional Institute (St. Clairsville)
Mansfield Correctional Institution (Mansfield)
Grafton Correctional Institution (Grafton) – added late 2004, due to addit. funding
Chillicothe Correctional Institution (Chillicothe) – added late 2004, due to addit. funding
Madison Correctional Institution (London) – added late 2004, due to addit. funding
Franklin Pre-Release Center (Columbus)
Northeast Pre-Release Center (Cleveland)
Ohio Reformatory for Women (Marysville)
Sample members may also be drawn from a privately operated Transitional Control facility in
Cleveland:
Harbor Light (Cleveland)
Post-Release Counties:
All study participants (experimental and control) are returning to one of three Ohio Counties:
Allen County
Cuyahoga County
Franklin County

Much of the information included in this memo was drawn from a document being prepared by the Ohio Institute
of Correctional Best Practices. When it is finalized I will forward to copy to Christy and Pam.

Ohio Site Design Memo
Page 2.

Assignment to Experimental and Control Group
Ohio is using random assignment to assign prisoners who are eligible to either the experimental
or the control group. The process is as follows:
Offenders who meet the eligibility requirements for participation in the CORE program are
identified from three sources:
Reception Center – during reception, offenders who meet the criteria are identified and
the list is forwarded to the PD;
Institutions – Institutional staff review their existing population to determine if inmates
meet the criteria for the program;
Central Office – The PD generates a Future Release Report List on a monthly basis.
Names on the list are reviewed to determine eligibility.
Once a list of eligible inmates is generated, the Reentry Management Team (RMT) in
conjunction with the community partners and parole officer meets with each inmate and explains
the program. If the inmate is interested, he/she must complete the CORE Entry Questionnaire.
The completed questionnaire is forwarded to the PD for review and final determination. Once the
PD makes the final eligibility determination, the inmate’s name is placed on a list. The list is
consecutively numbered. The names (numbers) are then drawn randomly and assigned to the
experimental or control group. The RMT is notified of who will be part of the experimental
group and the team begins meeting with the inmates to develop their Reentry Accountability
Plans (RAP).
Offenders in the experimental group receive intensive case management services and treatment
programs in the domain areas that are identified as necessary, while in the institution, for twelve
months in the community while on parole or PRC, and for an additional twelve months once
released from supervision. Inmates in the control group still receive treatment services through
the normal institutional process, but will not receive the extended case management services in
the community, after release.
Prisoners are tracked for two years following release from prison even if they voluntarily
withdraw or are terminated from the CORE program.
Revisions to Eligibility
The determination of eligibility has been refined since the initial application. The initial federal
guidelines precluded sex offenses. This has since been changed and sex offenders are now
eligible for participation in the CORE program and therefore may be assigned to the
experimental and/or control group.

Ohio Site Design Memo
Page 3.

Ohio has defined the target population as:

Offenders convicted of murder, robbery, aggravated assault, burglary, theft, and arson2;
Offenders convicted of high felony level such as murder, felony 1, 2, or 3 or the offender
is considered a serious offender;
Offenders between the ages of 18-35 at the time of the interview for the program;
Static risk assessment score of 5 or higher;
More than one dynamic domain at 3 or more need level;
Offenders receiving a sentence of incarceration of twelve months or more;
Offender who will be released to Allen, Cuyahoga, or Franklin Counties;
Offenders who will be on parole or post release control for at least one year following
release from prison;
Offender is eligible to be housed at one of the facilities listed above.

In 2004 additional money became available to expand services for offenders with serious mental illness. Ohio is
using these funds to implement a pilot program called ACT which will provide intensive wrap around services for
offenders designated as seriously mentally ill. As part of the new program, starting in October 2004 sex offenders
are eligible for inclusion.

Oklahoma Adults
Offenders are eligible for PROTECT if they are returning to Oklahoma County, are
between the ages of 18 and 35, and have an LSI score of 29 or greater. Offenders from
all state prisons, private (contract) prisons, community work centers, and community
corrections centers are eligible for PROTECT. For the national evaluation, we initially
targeted only those PROTECT participants who were incarcerated in one of the state
prisons (n=18). We projected between 60 and 115 PROTECT participants would be
released from the state prisons during the baseline data collection period. However,
because case flow was lower than originally projected, the decision was made on
11/23/04 to include in the study PROTECT participants who are incarcerated in any of
the four private prisons and six community corrections centers.
Our comparison group consists of offenders who meet the PROTECT age and LSI
criteria but who are returning to TULSA County. We are using offenders’ preincarceration county of residence to identify a sample of comparison offenders likely to
return to Tulsa.

Pennsylvania
In Pennsylvania, the Erie PA Reentry Project (EPRP) serves male and female adults from
any of the secure correctional institutions (SCI’s) who are returning to Erie County.
EPRP participants are transferred to Albion SCI (males) or Cambridge Springs SCI
(females) to begin EPRP programming. Other eligibility criteria are that the offender is
between 18 and 35 years old at the time of release from the SCI and is going to be
released on state parole or state re-parole. All EPRP participants are paroled to a
community corrections facility in Erie as part of their transition back into the community.
Males are paroled to Erie Community Corrections Center (CCC), a state-run community
corrections center, and females are paroled to Gaudenzia, a contract treatment/community
corrections facility. The average length of stay for EPRP participants at the community
corrections facilities is 90 days.
For our evaluation, we will conduct baseline interviews at the community corrections
facilities, rather than at the prisons, to ensure that offenders are actually in the community
at the time of the 3-month follow-up interview. Male study participants will be identified
and interviewed at two of the community corrections facilities in Erie, PA: Erie CCC
(EPRP participants and comparisons) and Gateway Erie (comparisons only), another
treatment facility contracted by DOC. We will interview female EPRP participants at
Gaudenzia; we will not have a female comparison group in PA. We plan on interviewing
all of the EPRP participants and an equal number of male comparison subjects. The
comparison group will comprise state parolees, state re-parolees, technical parole
violators (TPV’s) with community parole center (CPC) placements ("halfway-backs"),
and pre-release cases who are between 18 and 35 years old and are returning to Erie,
Crawford, or Warren County. Several DOC and Pennsylvania Board of Probation and
Parole (PBPP) staff believe that pre-release cases are comparable to state parolees and
halfway-back cases (TPV-CPC cases) are comparable to state re-parolees. We are
including Warren and Crawford counties as post-release target areas for comparison
subjects to ensure an adequate sample size.
For study participants from PA, the "release date" will actually be the date the offender is
released FROM the community corrections facility TO the community. Baseline
interviews will be scheduled according to the offender's expected or actual DATE OF
ARRIVAL at the CCC. At two of the three facilities (Erie CCC and Gaudenzia), study
participants will be interviewed during their first week at the facility. Offenders will then
spend approximately 3 months at the facility before being released to the community. The
3-month follow-up interview will be conducted 3 months after the offender is released
from the center to the community, so there will be more time (~6 months) in between the
baseline interview and first follow-up interview than there will be in other sites. At the
third facility, Gateway Erie, some offenders will complete inpatient substance abuse
treatment before moving into a work release phase, while other offenders will go straight
into work release upon entering the facility. Because offenders at Erie CCC will not be
in inpatient treatment, the timing of baseline interviews at Gateway will be most
comparable to the timing of baseline interviews at Erie CCC if offenders at Gateway are
interviewed during the first week of work release.

SC Adults – Study Design
Study Design
The SC DC reentry initiative is a voluntary program serving adults who are incarcerated in one
of 71 medium or maximum security institutions (Kershaw, Leath [women’s], Lee, MacDougall,
Ridgeland, Tyger River [this program targets sex offenders], and Wateree River) and who are
returning to any county in the state. The only additional selection criteria are that the adults be
between the ages of 17-35 (no age restrictions for sex offenders), have post-release supervision
(TIS sentences/splits, CSP, a few who are paroled, and possibly some that are on supervised
furlough), have a prior conviction (this criteria does not apply to sex offenders), not have been
sentenced under the Youthful Offender Act, and be 12 months from the projected release date.
Our best estimate for the number of SVORI participants who would be available for enrollment
in our study from 9/04 to 8/05 is 345.
For the comparison group (n=403), we will use two subgroups: “same-same” and “differentsame”. The “same-same” group will consist of individuals from the 7 targeted facilities who
meet all eligibility criteria except post-release supervision (n=189). The “same-same” group will
also include 66 non sex offenders from Tyger River who meet all eligibility criteria (including
post-release supervision). The “different-same” group will consist of individuals who meet all
SVORI eligibility criteria (including post-release supervision) but who are incarcerated in the
following medium or maximum security comparison facilities: Camille Griffin Graham
(women’s; n=8), Allendale (n=68), and Evans (n=23). SC DC is willing to flag the comparison
group so that they are not transferred to facilities in which we are not conducting interviews.
Interviewing at Watkins Pre-Release Center began in December 2004 since many eligible study
participants are transferred there 30 days prior to release.

As of 4/01/04, participants from all 7 facilities except Kershaw and Ridgeland had been enrolled in the program.
SVORI programming at Tyger River was suspended from June – November 2004 due to staff turnover.

SC Juveniles – Study Design
Study Design
The SC DJJ reentry initiative is a non-voluntary program (it is the new “standard” way the youth
will be released) serving youth who are incarcerated and released from one of 4 maximum
security institutions in the Broad River Complex (Birchwood Institution, John G. Richards
Institution, and Willow Lane Institution [to house females]) or any of the medium security
community corrections facilities (including Camp Bennettsville I and II, Camp Sandhills, Camp
White Pines I and II, Generations Bridges Camp, and Aspen Alternative). Basically, the only
institutions that are excluded from SVORI are the 3 evaluation centers and the juvenile detention
center. The program targets youth who meet eligibility criteria, including being committed from
the following counties: Orangeburg/Calhoun, Dorchester, Spartanburg, and Florence (the later
two counties were “rolled in” with participants enrolled in October 2004).1 We estimate that
approximately 68 SVORI youth (from all 5 participating counties) will be released from October
2004 – September 2005.
For the comparison group (target n = 138), we will use primarily a “same-different” design
(youth incarcerated in the same facilities but returning to Greenville, Aiken, and Sumter
counties), but will be able to include some “same-same” youth (program-eligible youth who are
not able to enter the program because of caseload caps). The “same-same” group will likely not
be very large (and it may be impossible to even come up with an estimated number ahead of
time), but since we are going to be in the facilities anyway, it would be easy to include these
youth in our study.

Three additional counties (Kirhaw, Marion, and York) are trying to implement the same model but are not
receiving funding; these counties will not be included in either the treatment or comparison group.

Site-Specific Design for Washington “Going Home”
Overview
In Washington, SVORI offenders will come from four facilities in the state (three near
Seattle, and one in Walla Walla) and will return to Pierce and King Counties (both near
Seattle). Comparison offenders will come from six corrections facilities and eleven
work/pre release facilities in the state (near Seattle, and in or on the way to Walla Walla)
and will return to Pierce and King Counties (the same communities as the SVORI
offenders).

Background
The SVORI program in WA targets offenders returning to three counties (Pierce, King,
and Spokane), and offenders are immediately transferred to one of four facilities once
they are enrolled in the program.
Pre-release geographical considerations: Three of the four SVORI facilities are located
in or near Seattle; one SVORI facility is located in Walla Walla, which is in the southern
portion of the state but within reasonable driving distance of Seattle. Because the
majority of offenders are held in the Walla Walla facility, we cannot exclude this facility
from the study.
Post-release geographical considerations: Pierce and King Counties are located near
Seattle; Spokane County is located on the far eastern edge of the state. Based on a
snapshot of potential releases and an analysis of actual releases from last year, it
appears that there are also far more offenders returning to Pierce and King Counties
than returning to Spokane. Because Spokane is so remote, and there are relatively few
offenders returning to that community, we have made the decision to exclude offenders
returning to Spokane from the impact evaluation.
Considerations for selection of comparison offenders: There appear to be at least twice
as many offenders returning to Pierce and King Counties than the program can
accommodate. Although the DOC is not open to random assignment, their MIS person is
willing to work with us to select matched comparison offenders for each participants
enrolled in the SVORI program. The Department would agree not to later enroll any
offenders selected as our comparison subjects. These comparison offenders,
unfortunately, cannot be limited to the same four facilities as the SVORI offenders will be
in (although some comparison offenders may be held in those four facilities). Based on
an analysis of release trends and the snapshot of potential releases, we should have
enough comparison cases if we limit the data collection facilities to those in or next to
Pierce and King Counties, and the facility in Walla Walla (and one facility between

Seattle and Walla Walla). This would exclude three facilities in Jefferson, Clark, and
Franklin Counties due to geographic distance.

SVORI Multi-Site Evaluation Sample Selection Procedures:
Colorado Juvenile Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The scope of the Colorado Affirms Reentry Efforts (CARE) program evolved over the course of
the grant. For the first 1.5 years, the program’s focus was preparing youth for release through
the involvement of a parole advocate. This was accomplished through funding a new position—
parole advocates. Parole advocates worked with the regular institutional case managers to
whom youth are assigned (with the case manager converting to the youth’s parole officer after
release) prior to release and then served as a safety net after release. The parole advocates
provided more intensive, personalized support to participants. The program then shifted to
incorporate more evidence-based programs and services, and focused on delivering these
programs at one primary facility. The program delivered at this facility was built on the
“Integrated Treatment Model” and incorporates the family an integral part of the incarceration
and transition process.
Differences Between SVORI and “Business as Usual”
Although case managers work with all juveniles as part of the DYC system (and then serve as
parole officers after release), the parole advocate role was specifically funded by the CARE
grant and was not available to youth in the general DYC population. Parole advocates provided
an additional, more concentrated resource for youth during the reentry process, and youth
participating in the CARE program had access to services typically not available to other youth
due to lack of resources. Some services provided by the Parole Advocates included
participating in client manager/family meetings, participating in job readiness groups, and
working with the workforce center to secure jobs for participants. .
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
CARE initially drew male participants from the Lookout Mountain and Ridgeview facilities, then
began focusing primarily on the Everest facility. For females, the Girls’ View facility was initially
part of the CARE program; however, it was shut down and has since re-opened under a new
name (Marler Center). It had limited involvement with the CARE program. In order to be
eligible for the program, boys had to be returning to the Denver/Metro area (Denver, Adams,
Arapahoe, Douglas, and Jefferson Counties). Girls were not limited to a particular geographic
area of return in order to obtain sufficient numbers of female participants).

Other Eligibility Criteria
Youth must have scored 28+ on the Colorado Youth Level of Supervision Inventory (which has
a range of 15-40 and is administered to all youth upon entry into the DYC system) to be eligible.
CARE participants who were supervised by certain case managers were recruited for the
program.

Voluntary/Mandatory Nature of Program

According to the program director, participation in the program was voluntary.

Program Enrollment Period
The program began enrolling cases in September 2003 and stopped enrolling in April of 2006.
As of March 2006, the program had enrolled 211 individuals (cumulative enrollment) over the
course of the grant.

SVORI and Comparison Populations Selected for the Evaluation
The impact component of the multi-site evaluation of SVORI entailed the identification of a
specific group of SVORI participants and comparable, non-SVORI offenders for which
individual-level data was obtained. Table 1 (below) is intended to convey information about how
these individuals were identified. The table provides details regarding the specific SVORI and
comparison groups included in the two main data collection components of the impact
evaluation: the longitudinal offender interview data and the administrative data.1 For the
interview component, the table presents the selection criteria used to identify the appropriate
group and the specific procedures through which the sample was identified. For the
administrative data component, the table describes the specific groups for which we were able
to obtain data. For both data collection components, it should be noted that our treatment group
assignment protocol specified that individuals are considered to have enrolled in SVORI if they
were in the program for more than one day.
Finally, it is important to understand the relationship between the population identified as being
eligible for inclusion in the interview component (i.e., the individuals who were identified through
the procedures outlined in Table 1 and are included in the “master file” for the program) and the
final group actually enrolled in the baseline interview component (the “baseline sample”), as well
as subsequent interview waves.2 Table 2 provides detailed information about all cases fielded
at each interview wave, listing the number of cases assigned to each final code maintained in
the field management system developed for the interview component of the study. Data are
provided separately by group status (SVORI vs. comparison), gender, and interview wave.

Some administrative data analyses will be limited to only the sample for which interviews were
conducted (the “interview-only” sample). However other analyses will be based entirely on administrative
data and are designed to include an expanded sample of SVORI and comparison individuals.
2
Only individuals in the baseline sample were pursued for the subsequent interview waves.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Colorado Juvenile)

SVORI Population Selected for
Evaluation
Interview Component

Selection Criteria: Male
CARE participants
incarcerated in all facilities
served by the program
(Lookout Mtn, Ridgeview, and
then later Everest) who were
projected to be (and actually)
released during the baseline
enrollment period.
Sample Identification
Procedures: PD sent file for
CARE participants and
periodically provided release
date updates. When case
releases fell within our 3-month
time window, site liaison
requested the additional fields
from PD. As new cases were
enrolled in CARE, PD
forwarded information.

Administrative
Data
Component
All CARE
participants
included in the
sampling frame
for the interview
component.

Comparison Population Selected for
Evaluation
Interview Component

Selection Criteria: Male youth
within the CARE facilities who
were supervised by case
managers other than the ones
from which CARE participants
were recruited, who met all
other CARE criteria (released to
the Denver/Metro area, scored
28+ on the CLSI), and who
were projected to be (and
actually) released during the
baseline enrollment period.
Sample Identification
Procedures: PD identified
appropriate comparison
subjects based on case
manager assignments and
sent file containing information.

Administrative
Data
Component
All comparison
participants
included in the
sampling frame
for the interview
component.

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Colorado Juvenile)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Florida Juvenile Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
Going Home is designed to facilitate a smooth transition from residential placement to
community living and offer constructive alternatives for economic self-sufficiency to youth
coming out of residential programs. The program includes a pre-release curriculum, which
spans a wide array of life skills and moral reconation topics, and extensive release planning
activities. Post-release services are focused on educational and vocational support (including
job development and job coaching). Both the pre-release curriculum and the provision of postrelease support services are contracted to Transition, Inc. Youth can access the post-release
services even after the conditional release period has ended, up until the age of 21.
Differences Between SVORI and “Business as Usual”
For the most part, all youth in FL DJJ facilities have access to the same services. All youth are
given a needs assessment, which is used to develop a performance plan. However, SVORI
youth also receive a YLS/CMI assessment. This more intense level of evaluation and definition
of needs may generate increased referrals for SVORI youth. For all youth, an exit conference is
held 14 days before release, with the conditional release case manager sometimes in
attendance. All youth who are on conditional release are monitored for 6 to 9 months, during
which the treatment plan developed prior to release is reviewed and updated monthly.
“Conditional release agencies” provide case management, conduct assessments, supervise
compliance with conditional release plans, make home visits, and conduct random drug testing.
The primary difference between SVORI and non-SVORI youth is the pre-release curriculum
(which only SVORI youth receive except at the Bay Point facility, in which the curriculum is
provided to all youth) and access to enhanced post-release educational and vocational services
available through Transition, Inc. Although all youth can access these services until age 19, the
contract with Transition allows SVORI youth to access these services until the age of 21.
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
Although grant funds were divided among Miami-Dade, Hillsborough, and Duval Counties, only
Miami-Dade was selected for the impact evaluation. Youth for this site must be committed from
Dade County into one of the following facilities: Bay Point Schools, Inc., Everglades Youth
Development Center, Florida City Youth Center, Southern Glades Youth Camp, and Wings for
Life South Florida (girls).

Other Eligibility Criteria
Youth must be committed for a violent felony or burglary. In addition, youth must be at least 16
years of age by the time they leave the residential facility and not have a high school diploma or
have completed a GED by the time they leave the residential facility. YLS/CMI scores are also

used to determine program eligibility (originally, youth needed to be rated as “high” risk; later,
this was modified to “medium” or “high” risk).

Voluntary/Mandatory Nature of Program
Participation in the program is voluntary.

Program Enrollment Period
The program began enrolling participants in August of 2003. The program continued enrollment
until March of 2006. A total of 102 youth were enrolled over the course of the grant.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Florida Juvenile)

SVORI Population Selected for
Evaluation
Interview Component

Selection Criteria: All SVORI
participants at the Miami-Dade
site who were projected to be
(and actually) released during
the baseline enrollment period.
Enrollment was discontinued
for females.
Sample Identification
Procedures: Dade county sent
a master file of program
participants on an ongoing
basis. The SL identified cases
with relevant release dates.

Administrative
Data
Component
All SVORI
participants
included in
sampling frame
for interview
component

Comparison Population Selected for
Evaluation
Interview Component

Selection Criteria: Youth who
meet the same program
eligibility criteria, were
committed to the same
facilities, and were projected to
be (and actually) released
during the baseline enrollment
period, but who were committed
from Broward and West Palm
(rather than Dade) counties.
Note that YLS/CMI scores are
not available on youth who are
not participating in SVORI so
we were not able to apply this
criterion to them.

Sample Identification
Procedures: Broward and
Palm Beach County each sent
files (on an ongoing basis)
containing comparison subjects
who had the appropriate
release dates.

Administrative
Data
Component
Comparison
individuals
included in
sampling frame
for interview
component

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Florida Juvenile)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C
n/a

n/a

n/a
n/a

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Iowa Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The Iowa Going Home KEYS program included a 12-week life skills and reentry course (topics
include communication skills, conflict management, future goal setting and reentry case
planning, substance abuse treatment, computer skills and vocational training, education,
employment, mental and physical health, money management, family issues, housing, time
management, legal assistance, recreational activities) delivered in-prison (taught by staff from
Des Moines Area Community College [DMACC]) as close to participants’ time of release as
possible. During the class, post-release case managers and parole officers met with
participants and develop reentry case plans. The plans identified areas of need and began
planning services and resources to address those needs prior to release and once released.
During the course, representatives from many community-based organizations and government
agencies conducted sessions to explain how the processes work for offenders after release and
to provide contact people with whom participants can connect upon release.
Although KEYS participants were supposed to be directly released to community supervision,
approximately 20% ended up being sent by the parole board to work release facilities (there
were two in Polk County) for 3-6 months before release to the community. During the postrelease phase (which lasted for at least 6 months after release), DMACC provided intensive
case management and two dedicated parole staff members, who had offices at DMACC,
provided enhanced parole supervision (together, the DMACC case manager and dedicated PO
were the “Transition Team” and made contact while the participants were still in the institution).
Participants also participated in community Accountability Boards and were offered a weekly
peer support group to address reentry concerns and emotional issues. In addition, DMACC
offered a support group for family members of offenders.
Differences Between SVORI and “Business as Usual”
The KEYS curriculum was not available to the general prison population. The general prison
population received a prison treatment plan but had to do its own release planning (in contrast
to the release plan developed for KEYS participants and reentry services covered in the KEYS
curriculum). In addition, KEYS participants were the only ones who had DMACC case
managers and dedicated parole officers (the general population of inmates on parole are
supervised by “general” parole officers who are housed at the parole building, rather than the
DMACC campus).
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
In order to be eligible, prisoners must have been returning to Polk County (Des Moines). The
KEYS curriculum was delivered at three facilities (Fort Dodge Correctional Facility, Newton
Correctional Facility, and Iowa Correctional Institution for Women). Although the original plan
was to recruit from all facilities and have “KEYS eligible” individuals transferred to the KEYS

facilities (which are all medium security) for programming, this plan was not implemented (due
to concerns about individuals who would otherwise be released from minimum security facilities,
which are a majority of inmates, being transferred to a higher security prison only for
programmatic reasons) and KEYS participants were only identified from the three medium
security facilities listed above.

Other Eligibility Criteria
Participants must have been between the ages of 18-40 (for males) or 18-48 (for females),
although originally the program used a lower upper age limit (35 for males). Additionally,
inmates must have had one year or more to serve on their sentence at the time of their release.
Individuals with life sentences were excluded, as were people who were currently serving time
for parole violations. Originally, the program also limited participation to individuals with LSI-R
scores of 20 or above, but this criterion was relaxed over the course of recruitment and
prisoners with lower LSI-R scores were accepted. Finally, facilities only put offenders who were
considered to be “program ready” into KEYS—meaning the offender had completed all other
treatment provided by the facility that had been identified as a need by their facility counselor or
the parole board
If after going through a pre-screening process by the parole board (to verify likelihood of
release) prisoners were approved, eligible individuals were randomly assigned to receive either
KEYS programming or business as usual. Random assignment took place only during the
baseline enrollment period for the multi-site evaluation.

Voluntary/Mandatory Nature of Program
According to the program director, participation in the program was voluntary.

Program Enrollment Period
The program began enrolling participants in January of 2003. Enrollment ended in April of
2006, with a total cumulative enrollment of 490.

the procedures outlined in Table 1 and are included in the “master file” for the program) and the
final group actually enrolled in the baseline interview component (the “baseline sample”), as well
as subsequent interview waves.2 Table 2 provides detailed information about all cases fielded
at each interview wave, listing the number of cases assigned to each final code maintained in
the field management system developed for the interview component of the study. Data are
provided separately by group status (SVORI vs. comparison), gender, and interview wave.

Only individuals in the baseline sample were pursued for the subsequent interview waves.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Iowa Adult)

SVORI Population Selected for
Evaluation

Comparison Population Selected for
Evaluation

Interview Component

Administrative
Data
Component

Interview Component

Administrative
Data
Component

Selection Criteria: Individuals
randomly assigned to the
KEYS group and projected to
be (and actually) released3
during the baseline enrollment
period. Random assignment
entailed the following: once
the list of names of eligible
prisoners from each facility
were identified, the names
were sent to the parole board
for a prescreening process (to
verify likelihood of release),
cases receiving prescreen
approval were then sent to
CJJP for random assignment,
which involved an SPSS
program to randomly split the
sample of eligible and
prescreened offenders into two
groups and then identify one
group as the KEYS group and
the other as the control group
(note: the groups were not
equally distributed – KEYS
slots were filled first and then
the remaining individuals were
allocated to the control group).
An intent to treat design was
employed (drop-outs could not
be considered control group
members).

All individuals
randomly
assigned to
KEYS and on the
sampling frame
for the interview
component

Selection Criteria: Individuals
randomly assigned to the
control group and projected to
be (and actually) released
during the baseline enrollment
period. Random assignment
procedures are described in the
“SVORI population selected for
Evaluation” column.

All individuals
randomly
assigned to the
control group and
on the sampling
frame for the
interview
component.

Sample Identification
Procedures: CJJP sent data
files to site liaison.

Note that for the individuals who were released to work release facilities, “release” was defined as release from the work release facilities (not
the original institution). Therefore, although KEYS curriculum was only delivered in three prisons, SVORI interviews took place in five facilities
(the three prisons of interest and the two work-release facilities--the Fort Des Moines Community Corrections Center and the Women’s
Residential Correctional Facility--where some portion of KEYS participants were sent before being released to the community.

Table 2.

Description
Number of Cases
identified for
Evaluation
Number of Cases
Fielded
Number of Cases
Interviewed

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Iowa Adult)
Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Indiana Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The portion of the Indiana SVORI grant that went toward adult programming was small, with
funds used to support two reentry programs in Allen County: a Reentry Court and a Community
Transitions Program (CTP). Both programs are strictly post-release, with official enrollment
occurring upon release to Allen County Community Corrections (ACCC). Program participants
are those who choose to forgo the remainder of their prison sentence to be released early on
electronic monitoring.
Offenders going into the reentry court program are released from prison (to parole) more than
60 days pre-release. Those only receiving the Community Transitions Program (CTP) have an
early release date of less than 60 days and are released to probation. Many of the services
provided to these two groups at ACCC are the same, but the Reentry Court participants have
more time to access the services available (because they are under the jurisdiction of ACCC for
longer than those in the CTP program) and are responsible for appearing in court. In addition,
Reentry Court participants are followed by a reentry case manager for at least one year
whereas those on CTP only receive case management for the duration of their sentence (60
days or fewer); separate case managers are assigned to the two groups. In addition to the case
management services and intensive supervision received by both groups, numerous courses
(including behavior modification) and services (employment, education, substance abuse
treatment, etc.) are offered through ACCC. Participants are required to pay for the services
that they receive while on electronic monitoring (with DOC paying half the cost). Because the
Reentry Court program was not included in the SVORI Multi-Site Evaluation, the remainder of
this document focuses on the CTP program only.
Differences Between SVORI and “Business as Usual”
Because the program is post-release in nature, there are no differences in the pre-release
services received by SVORI and non-SVORI offenders. Out of the 92 counties in Indiana, 60
have a community corrections agency; however, few are as large and developed as the ACCC
in Fort Wayne. In general, former prisoners outside of ACCC are not supervised as closely and
do not have as many services readily available to them. In addition, the general probation
population is not typically on electronic monitoring (something that is standard for both the
Reentry Court and CTP populations, because they have chosen the early release option).
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
In order to be eligible for the program, individuals must be returning from any IDOC facility to
Allen County.

Other Eligibility Criteria

Initially, in order to be eligible prisoners had to choose to forgo the remainder of their prison
sentence to be released early to the ACCC on electronic monitoring. However, due to a new
piece of legislation, prisoners were required to either accept the program or to petition to the
courts to allow them to opt out with no penalty within 10 days of their 60 day pre-release
notification.
The following groups are excluded from participation: individuals who have been terminated
from reentry court twice, those who have a pending suit against ACCC, those who have severe
medical requirements, and those serving less than a two-year sentence.

Voluntary/Mandatory Nature of Program
Although the program director considers participation to be voluntary, the way an individual
chooses to enter the program is complicated. Originally, participants had to choose to forgo the
remainder of their prison sentence to be released early to the ACC on electronic monitoring.
However, Indiana then signed new legislation that said that prisoners do not have the choice to
select early release. Instead, they were given 10 days after their 60 day pre-release notification
to either accept the program or to petition to the courts to allow them to opt out with no penalty.
If they decline to participate without getting proper approval from the court, they will lose their
earned time.

Program Enrollment Period
The program began enrolling participants in July 2002. As of March 2006, the program had
enrolled a total of 527 individuals over the course of the grant, and planned to continue
enrollment indefinitely.

the field management system developed for the interview component of the study. Data are
provided separately by group status (SVORI vs. comparison), gender, and interview wave.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Indiana Adult)

SVORI Population Selected for
Evaluation

Comparison Population Selected for
Evaluation

Interview Component

Administrative
Data
Component

Interview Component

Administrative
Data
Component

Selection Criteria: Individuals
identified as eligible for the
CTP program (who have not
denied the program), who were
released from one of 7
designated facilities for the
evaluation (Indiana Women’s
Prison, Westville, Chain
O’Lakes, Rockville, Plainfield,
Putnamville, Miami), and who
were projected to be (and
actually) released to ACCC
during our baseline enrollment
period.

Individuals
identified as
eligible for the
CTP program
who were
released in 2004,
2005, and 2006

Selection Criteria: Individuals
incarcerated in the same 7
facilities from which CTP
participants could come but
who were returning to Marion
(rather than Allen County), who
matched the selection criteria
used for the CTP program, and
who were projected to be (and
actually) released during our
baseline enrollment period. If it
becomes known that
comparison group members
became enrolled in the CTP
program or the Community
Chaplain Program in Marion
County, the cases were
dropped from further follow-up.

From 2004-2006
release cohort
files, will subset
to individuals on
sampling frame
for interview
component (and
possibly an
alternative
expanded
sample)

Sample Identification
Procedures: A DOC staff
member sent a list of potential
CTP participants each month.
An ACCC staff member also
sent weekly files listing
individuals eligible for the CTP
program and who have not (to
their knowledge) denied the
program. As it becomes
known that someone ends up
denying the program, cases
that have already been
baselined are dropped from
further follow-up.

Sample Identification
Procedures: A DOC staff
member sends a list of potential
comparison group members
each month.

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Indiana Adult)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Kansas Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The Shawnee County Reentry Program (SCRP) attempts to provide high-risk offenders with a
smooth transition into the community. All male program participants who are incarcerated in
facilities other than the Lansing Correctional Facility (LCF) are transferred to LCF before
beginning the program. Program participants are assigned to specialized case managers, who
are highly qualified and have a smaller case load than typical parole officers. While
incarcerated, program participants meet with case managers, who conduct a risk assessment
(LSI-R) and develop a case plan. Program participants receive intensive case management
and pre-release planning. Reentry participants also receive individual housing assistance and
are prioritized for the cognitive behavior, life skills, and, for females, family transition classes
offered in the facilities. Prior to release, participants have post-release services lined up for
them by the case manager. A major program component is the Accountability Panel, which,
after release, welcomes the inmate home, hold him/her accountable for what happens in the
community, and provides options to re-incarceration when an offender commits an act that
could result in revocation. Grant funds also supported the position of boundary spanner, a staff
member that meets with agencies and improves coordination at the systems level.

Differences Between SVORI and “Business as Usual”
All inmates meet with a treatment team upon entry to a KS facility and have a “program plan”
(based on LSI-R scores) developed, including recommendations and details requirements the
inmate must complete. While inmates who are participating in SCRP are assigned to reentry
case managers with a specialized caseload, non-SCRP inmates are assigned to parole officers.
Although the parole officers do meet with inmates while they are still incarcerated, they do not
have the time or resources to spend significant time with an inmate before release. All KS DOC
inmates receive a release plan prior to release. However, SCRP participants receive more
individual attention to prepare for release, more extensive follow-through, and more advance
release preparation. Non-SCRP inmates do not meet with an Accountability Panel upon
release.
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
The program serves adults who are incarcerated in any correctional facility and who are
returning to Shawnee County. All males who are incarcerated in facilities other than the Lansing
Correctional Facility (LCF) are transferred to LCF before beginning the program. There is only
one female facility in KS, which is located in Topeka (TCF).

Other Eligibility Criteria

Additional eligibility criteria include the following: 18-34 years old (note: the upper age limit was
later dropped), a minimum of 12 months from the projected release date, and convicted of a
serious or violent offense (or a sex offense).

Voluntary/Mandatory Nature of Program
According to the program director, participation in the program is voluntary.

Program Enrollment Period
The program began enrolling participants in April of 2003. As of March 2006, the program had
enrolled 148 individuals over the course of their grant and planned to continue enrolling
indefinitely.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Kansas Adult)

SVORI Population Selected for
Evaluation
Interview Component

Selection Criteria: SVORI
participants in the Lansing and
Topeka facilities who were
projected to be (and actually)
released during the baseline
enrollment period. 8/05.
Sample Identification
Procedures: SCRP prepared
monthly files containing all
eligible cases.

Administrative
Data
Component

Comparison Population Selected for
Evaluation
Interview Component

Selection Criteria: Offenders
from the Lansing and Topeka
facilities who meet all the
SVORI eligibility criteria (18-34
years old, have been convicted
of a serious, violent, or sex
offense), projected to be (and
actually) released during the
baseline enrollment period, but
who are returning to Sedgwick
County (using pre-prison county
of residence as a proxy, since
the DOC database does not
have a field for projected post
incarceration residence).
Sample Identification
Procedures: SCRP prepared
monthly files containing all
eligible cases.

Administrative
Data
Component

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Kansas Adult)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Kansas Juvenile Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The Kansas Going Home Initiative (GHI) was designed to provide intensive support for serious
and violent juvenile offenders in the period leading up to and following release from a
correctional facility. Youth participating in the program receive intensive support and planning
services by a Community Reentry Facilitator (CRF; focused on serving the youth from program
enrollment through the period immediately following release) and a Long Term Support
Specialist (LTSS; focused on serving the youth once they are back in the community). The
work of these staff members augments the existing case management provided by facility and
community case managers, who have formal control over the cases (the CRFs and LTSSs do
not have the authority to sanction youth). The CRF’s meet with the youth monthly prior to
release and focus on release preparation. A major component of the program is family group
conferencing (involving family members, service providers, education representatives, law
enforcement officers, community case managers, and other individuals important to the youth),
with the conference taking place immediately following release. The LTSSs can maintain
contact with youth for up to a year after supervision ends.
Differences Between SVORI and “Business as Usual”
All youth in Kansas juvenile facilities receive case management by facility case managers while
they are incarcerated and community case managers during the period of conditional release
(community case managers typically do not visit the youth prior to release, but communicate
with the facility case manager). However, only youth participating in GHI receive the services of
the CRFs and LTSSs. These staff members focus on building relationships with youth and their
families, which they are able to do because they have small caseloads and because they do not
have the authority to sanction the youth (resulting in a supportive relationship). GHI participants
also benefit from having the CRFs and LTSSs participate in the “180 day staffings” on the
youth’s behalf, in which the facility treatment team revises the original service plan and plans for
release.
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
All juveniles returning to one of five judicial districts in northeastern and south central KS can
participate in the Going Home Initiative. Participants can come from any of the four juvenile
correctional facilities (JCF) in KS. The four JCF's serve different populations: Beloit JCF houses
female offenders; Larned JCF houses offenders with serious mental health and/or substance
abuse issues; Atchison JCF houses the youngest offenders; and Topeka JCF houses the most
serious offenders (it is also the largest JCF).

Other Eligibility Criteria
Youth must have a conditional release period of at least one year in order to be eligible.

Voluntary/Mandatory Nature of Program
According to the program director, participation in the program is voluntary.

Program Enrollment Period
The program began enrollment in May of 2003 and stopped enrolling participants in January of
2006. The program enrolled a total of 180 individuals over the course of the grant.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Kansas Juvenile)

SVORI Population Selected for
Evaluation
Interview Component

Selection Criteria: For the
purposes of the national
evaluation, only GHI
participants from two of the
four juvenile correctional
facilities, Beloit (females) and
Topeka (males), were
included. This decision was
made because there would not
be enough participants at the
three "specialty" facilities to
conduct an adequate study of
their populations, and pooling
them involves too much risk of
type II error. All GHI
participants from Beloit and
Topeka who were projected to
be (and actually) released
during the baseline enrollment
period were included in the
sampling frame.
Sample Identification
Procedures: SVORI PD sent
lists of all SVORI participants
at Topeka (males) and Beloit
(females) facilities to SL.

Administrative
Data
Component

Comparison Population Selected for
Evaluation
Interview Component

Selection Criteria: Because
the five judicial districts to which
GHI participants return account
for >50% of all released juvenile
offenders, we decided not to
have a comparison group of
juvenile offenders in KS.
Instead we will use data from
the Juvenile Justice Authority to
construct a historical
comparison group.
Sample Identification
Procedures: n/a (no
comparison subjects)

Administrative
Data
Component

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Kansas Juvenile)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C
n/a
n/a

n/a
n/a

n/a

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Maryland Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
For the Maryland Reentry Partnership (REP), The Enterprise Foundation is contracted to
provide case management and service coordination for male offenders. Pre-release activities
usually begin 90 days prior to release and begin with eligible offenders taking part in “exit
orientations” held for those who will be released within 90 days. During these orientations,
potential participants meet with their parole officer, community case managers (CCM),
advocate, service providers, law enforcement staff1, and Maryland Department of Public Safety
and Correctional Services (MDPSCS) staff, who inform them about the program and ask if they
would like to participate. Those who agree to participate are assigned a CCM (each REP
participant was assigned to a CCM employed at one of three community-based organizations in
the community in which the participant was returning) and an Advocate (usually an ex-offender)
who have regular contact with them post-release, link them to services, and help the participant
comply with his treatment plan. The pre-release phase of the program consists of an intense
risk and needs assessment (conducted pre-release) and two pre-release programs (life
skills/cognitive behavioral training and job readiness training). Upon release, a case plan is
created from the participants’ needs assessment. The CCM and Advocate work together to see
that the case plan is followed and collaborate with the parole officer if the participant is on
parole. Post-release activities may continue for as long as 3 years.
Differences Between SVORI and “Business as Usual”
Some pre-release activities are available to all inmates at the Metropolitan Transition Center.
All inmates at the MTC participate in an exit program called PREP, which involves community
members and agencies that can be of assistance to them upon release. Each inmate at the
MTC is also assigned to a case manager.
The involvement of the CCMs (who provide direct linkages to services in their community, as
well as case management and oversight) and the advocates are unique to the REP program
(with the both the CCMs and advocates working in partnership with parole officers). Other major
differences are that the REP clients were assigned to their parole officer prior to release (and
met with them prior to release) and that the parole officers were actively involved in the
development of the discharge plans.
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
REP participants are selected from the Metropolitan Transition Center (MTC) in Baltimore. In
order to be eligible, individuals must be returning to five specific Baltimore zip codes (21213,
21215, 21216, 21217, and 21218)
1

Law enforcement representatives were initially involved but then stopped participating in the orientations.

Other Eligibility Criteria
The program only serves males. The target population is the general “serious and violent”
offender population between the ages of 18 and 35. The program excludes those convicted for
sex offenses or crimes against children, those with active detainers, and the severely mentally
ill.

Voluntary/Mandatory Nature of Program
Participation in the program is voluntary.

Program Enrollment Period
The program began enrollment in December of 2001. As of March 2006, a total of 411
individuals had been enrolled (from the inception of the program), and the program was
planning to continue enrollment indefinitely.

SVORI and Comparison Populations Selected for the Evaluation
The impact component of the multi-site evaluation of SVORI entailed the identification of a
specific group of SVORI participants and comparable, non-SVORI offenders for which
individual-level data was obtained. Table 1 (below) is intended to convey information about how
these individuals were identified. The table provides details regarding the specific SVORI and
comparison groups included in the two main data collection components of the impact
evaluation: the longitudinal offender interview data and the administrative data.2 For the
interview component, the table presents the selection criteria used to identify the appropriate
group and the specific procedures through which the sample was identified. For the
administrative data component, the table describes the specific groups for which we were able
to obtain data. For both data collection components, it should be noted that our treatment group
assignment protocol specified that individuals are considered to have enrolled in SVORI if they
were in the program for more than one day.
Finally, it is important to understand the relationship between the population identified as being
eligible for inclusion in the interview component (i.e., the individuals who were identified through
the procedures outlined in Table 1 and are included in the “master file” for the program) and the
final group actually enrolled in the baseline interview component (the “baseline sample”), as well
as subsequent interview waves.3 Table 2 provides detailed information about all cases fielded
at each interview wave, listing the number of cases assigned to each final code maintained in
the field management system developed for the interview component of the study. Data are
provided separately by group status (SVORI vs. comparison), gender, and interview wave.

Some administrative data analyses will be limited to only the sample for which interviews were
conducted (the “interview-only” sample). However other analyses will be based entirely on administrative
data and are designed to include an expanded sample of SVORI and comparison individuals.
3
Only individuals in the baseline sample were pursued for the subsequent interview waves.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Maryland Adult)

SVORI Population Selected for
Evaluation
Interview Component

Selection Criteria: Individuals
flagged for REP participation*
who were projected to be (and
actually) released during the
baseline enrollment period.
Sample Identification
Procedures: A DPSCS staff
member sent a monthly file
after getting confirmation of
REP participants from MTC
(the list is confirmed by
Enterprise as well).
*note that not all “treatment”
group members received
treatment

Administrative
Data
Component
All individuals
flagged for REP
participation
included in
sampling frame
for the interview
component

Comparison Population Selected for
Evaluation
Interview Component

Selection Criteria: Individuals
who were housed at MTC, were
not enrolled in any specific
reentry programming, met all
other program eligibility criteria
except zip code (comparison
individuals were those returning
to Baltimore zip codes other
than those targeted by the REP
program), and were projected
to be (and actually) released
during our baseline enrollment
period.
Sample Identification
Procedures: The MIS staff at
DPSCD developed computerand hand-generated screening
procedures to select eligible
comparison cases on a monthly
basis

Administrative
Data
Component
All comparison
individuals
included in
sampling frame
for the interview
component

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Maryland Adult)

Description
Number of Cases identified for
Evaluation
Number of Cases Fielded
Number of Cases Interviewed

Baseline
Males
Females
S
C
S
C
n/a

n/a

n/a
n/a

3 Month
Males
Females
S
C
S
C
n/a

n/a

n/a
n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

n/a
n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

n/a
n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Maine Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The Maine Reentry Network is a collaborative model for community reintegration that begins
before release and involves the work of a Reentry Team (for juveniles, Family Systems Teams
are used) consisting of both DOC staff and community members. Intensive transition planning
(using a case management approach, which results in a case plan) begins six months before
release. Videoconferencing technology is used so that community-based service providers and
family members can participate in pre-release planning. Team members have resources in
place before release, and the team reconvenes after release as needed to implement the case
plan that was developed. In addition, in the later years of the grant, Resource Coordinator
positions were developed for regional probation offices to help coordinate services for adult
offenders once they are released.

Differences Between SVORI and “Business as Usual”
The adult Reentry Network was built based on the juvenile model that has been in place for
years – the FST collaborative, intensive case management approach. For adult males, there
are substantial differences in service receipt between treatment and comparison offenders.
However, many of the services that the Reentry Network has created (e.g., pre-release
MaineCare applications, pre-release Social Security Disability Income applications, and prerelease MOU’s with DOL Career Centers) have begun to be translated to all inmates and not
just the Reentry Network participants.
Volunteers of America has a contract to provide all women (Reentry and non-Reentry) being
released the same service package.
Because the Reentry Network was based on the existing juvenile model, there are not many
differences between treatment and comparison youth (unless they age out of the system, in
which case Reentry is their only resource for help). Having the Reentry Specialist on the FST
for case management only occurs for juveniles returning to counties targeted by the programs.
This feature enhances the work of the juvenile FSTs.
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
Maine’s SVORI program targets prisoners in all six adult facilities (Bolduc Correctional Facility,
Central Maine Pre-Release Center, Charleston Correctional Facility, Downeast Correctional
Center, Maine State Prison, and Maine Correctional Center) and two youth facilities (Long
Creek Youth Development Center and Mountain View Youth Development Center) ranging in
age from 16 to 25 and returning to four counties (Washington, Androscoggin, Penobscot, and

Knox).1

Other Eligibility Criteria
Facility staff refer people to the program based only on the two above criteria, and then the
SVORI program director conducts a second level screening, checking the individuals risk of
recidivism and level of need. Despite the second level of screening, the project takes virtually
all referrals received because if a person is in prison in Maine, they are considered high risk
(based on the very low incarceration rate for the state).

Voluntary/Mandatory Nature of Program
According to the program director, participation in the program was voluntary.

Program Enrollment Period
The program began enrolling participants in July of 2003. As of March 2006, the program had
enrolled 439 individuals over the course of their grant and planned to continue enrollment until
March of 2007.

SVORI and Comparison Populations Selected for the Evaluation
The impact component of the multi-site evaluation of SVORI entailed the identification of a
specific group of SVORI participants and comparable, non-SVORI offenders for which
individual-level data was obtained. Table 1 (below) is intended to convey information about how
these individuals were identified. The table provides details regarding the specific SVORI and
comparison groups included in the two main data collection components of the impact
evaluation: the longitudinal offender interview data and the administrative data.2 For the
interview component, the table presents the selection criteria used to identify the appropriate
group and the specific procedures through which the sample was identified. For the
administrative data component, the table describes the specific groups for which we were able
to obtain data. For both data collection components, it should be noted that our treatment group
assignment protocol specified that individuals are considered to have enrolled in SVORI if they
were in the program for more than one day.
Finally, it is important to understand the relationship between the population identified as being
eligible for inclusion in the interview component (i.e., the individuals who were identified through
the procedures outlined in Table 1 and are included in the “master file” for the program) and the
final group actually enrolled in the baseline interview component (the “baseline sample”), as well
as subsequent interview waves.3 Table 2 provides detailed information about all cases fielded
at each interview wave, listing the number of cases assigned to each final code maintained in
the field management system developed for the interview component of the study. Data are
provided separately by group status (SVORI vs. comparison), gender, and interview wave.

An additional county--Cumberland County--was added as a SVORI county in 2005.
Some administrative data analyses will be limited to only the sample for which interviews were conducted (the
“interview-only” sample). However other analyses will be based entirely on administrative data and are designed to
include an expanded sample of SVORI and comparison individuals.
3
Only individuals in the baseline sample were pursued for the subsequent interview waves.
2

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Maine)

SVORI Population Selected for
Evaluation

Comparison Population Selected for
Evaluation

Interview Component

Administrative
Data
Component

Interview Component

Administrative
Data
Component

Selection Criteria:
All SVORI participants from all
facilities who were projected to
be (and were actually)
released during the baseline
enrollment period and who
were from Washington,
Androscoggin, Penobscot, and
Knox counties. Note: on
3/31/05 we made the decision
to stop recruiting participants
from the juvenile system (the
two juvenile facilities) for
interviews.

All SVORI
participants (male
and female, aged
16-25) released
to Washington,
Androscoggin,
Penobscot, and
Knox counties
from any facilities
during the
baseline
interview window.

Selection Criteria: Four
comparison counties selected
to identify male comparisons:
Piscataquis County (to compare
against SVORI Washington
County), York County (to
compare against SVORI
Androscoggin County),
Kennebec County (to compare
against SVORI Penobscot
County), Lincoln County (to
compare with SVORI Knox
County). There is no
comparison group of women in
Maine because women
returning to comparison
counties receive SVORI-like
services through another
contract.

All males aged
16-25 returning
released to
Piscataquis,
York, Kennebec,
or Lincoln
counties from any
facilities during
the baseline
interview window.

Sample Identification
Procedures: Site data contact
ran a report from the Maine
DOC data system, screening
for client age range (16-25),
release date (within 3-months).
Facility liaisons verified client
data (checking file against the
SVORI enrollment file), file
returned to Maine DOC and
then sent to site liaison.
Additional information (on
guardian, address
clarifications, etc.) provided.
List screened for targeted
counties then uploaded.

Sample Identification
Procedures: (same procedures
used for SVORI and
comparison)

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Maine)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases Interviewed

Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Nevada Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
Once enrolled in the Going Home Prepared program (6 to 12 months prior to release), GHP
participants complete a SVORI-specific curriculum and are assigned to one of two reentry
caseworkers that provide case management services. All male participants are transferred to
Southern Desert Correctional Center (SDCC) for SVORI programming and go through the
program while being housed on the same unit; females receive programming at one of two
facilities. Upon release, program participants are supervised by one of four GHP-dedicated
parole officers who have a split caseload with both GHP and non-GHP parolees on intensive
supervision. They also receive case management services from one of two GHP-dedicated
social workers housed at P&P. The pre-release and post-release case managers work closely
together to facilitate GHP participants’ transition from prison to the community. Participants
receive GHP and community-based services and are tracked by the program for 12 months
following release, at which time the formal program participation ends (however, if an offender
needs more services beyond the 12th month of phase 2, the service providers agree to provide
these services).

Differences Between SVORI and “Business as Usual”
GHP participants receive more intensive case management and more service coordination. The
reentry caseworkers have a different role than other prison caseworkers--they are more
proactive and focus more on release than reincarceration. GHP participants receive
assessments, more referrals, and a greater degree of follow-through on those referrals from the
pre-release caseworkers and P&P social workers working together. The GHP participants are
more prepared upon release. While in prison, GHP participants:
o get an individualized reentry plan
o receive assistance with obtaining birth certificates and identification cards
o attend an orientation session in which community service providers, the P&P social
workers, and the GHP-dedicated parole officers explain what community services are
available and supervision expectations
o receive referrals to community service providers, which results in the ability to
immediately get services when released
GHP participants also have access to services from the subcontracted GHP-specific job
developer/trainer and mental health clinician.
Supervision is the same for GHP offenders as it is for other offenders on intensive supervision;
however, the GHP parole officers may watch more closely if the GHP client flounders because
the client has an advantage and greater support through program participation. In addition, the
P&P social workers have a greater awareness of community resources and can share this
knowledge and extra information with the parole officers (which benefits the participant).
An additional difference in post-release services is that GHP requires participants to attend
reentry court, and participate in groups facilitated by the mental health clinician.
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
Participants from any of the state’s prisons are eligible but must serve their parole sentence in
Clark, Lincoln, Nye, or Esmeralda Counties (southern Nevada). All male participants are
transferred to Southern Desert Correctional Center (SDCC) for SVORI programming, while
female participants receive programming at one of two facilities—Jean Conservation Camp or
Southern Nevada Women's Correctional Facility.

Other Eligibility Criteria
The GHP program targets male and female adult and juvenile offenders between the ages of 14
and 35 incarcerated in the adult NDOC system. Participants are screened for eligibility by
Offender Management and then GHP staff, after which the GHP Selection Committee reviews
offenders’ files to make final selections. Criteria that could keep a person out of the program
include insufficient time left on sentence to participate and inability to transfer to program facility
due to enemy gang on yard

Voluntary/Mandatory Nature of Program
Participation in the program is mandatory following enrollment and is made a condition of the
offender’s parole.

Program Enrollment Period
The program began enrolling participants in July of 2003 and continued enrollment until June of
2005. The program enrolled a total of 321 individuals over the course of the grant.

at each interview wave, listing the number of cases assigned to each final code maintained in
the field management system developed for the interview component of the study. Data are
provided separately by group status (SVORI vs. comparison), gender, and interview wave.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Nevada Adult)

SVORI Population Selected for
Evaluation
Interview Component

Selection Criteria: All
individuals 18+ who were
enrolled in GHP (and therefore
transferred to one of the 3
facilities in which programming
was delivered), who were
projected to be (and actually)
released during the baseline
enrollment period, and who
were returning to a Southern
Nevada county: Clark, Lincoln,
Nye, or Esmeralda.
Sample Identification
Procedures: PD uploaded file
with both comparison and
treatment subjects identified.
SL excluded treatment cases
that were in a facility in which
we were not conducting
interviews.

Administrative
Data
Component
All GHP
participants
included in
sampling frame
for the interview
component

Comparison Population Selected for
Evaluation
Interview Component

Selection Criteria: Individuals
who met all GHP eligibility
criteria except for the
“subjective” exclusion criteria
(e.g., inability to transfer to
program facility, insufficient
time left on sentence) and postrelease geographic
parameters3. Comparison
subjects came from facilities
located in Southern Nevada
(High Desert State Prison,
Indian Springs Conservation
Camp, Jean Conservation
Camp (female), Southern
Desert Correctional Center,
Southern Nevada Women's
Correctional Center)and (until
9/23/04) Northwest Nevada .
Comparison group members
must have been projected to be
(and actually) released during
the baseline enrollment period.

Administrative
Data
Component
All comparison
individuals
included in
sampling frame
for the interview
component

Sample Identification
Procedures: PD uploaded file
with both comparison and
treatment subjects identified.
SL excluded 1) all cases in a
facility in which we were not
conducting interviews), 2) all
cases that were reviewed by
the reentry program and that
had the potential to be reviewed
again, 3) all cases that were
reviewed by the reentry
program and that were denied
participation for a subjective
reason (e.g., disciplinary
history). Cases that were
reviewed for GHP but were
determined to be ineligible for
3

Interviewing was discontinued in northern Nevada due to small numbers of eligible participants in late September
2004; however prior to that time, comparison subjects could have released to a county in northwest Nevada,
including Washoe, Churchill, Carson City, Douglas, and Lyon counties.

SVORI Population Selected for
Evaluation
Interview Component

Administrative
Data
Component

Comparison Population Selected for
Evaluation
Interview Component

other reasons (e.g., from a
Northern NV county, enemy
gang in yard, not enough time
on sentence to participate)
were included as comparison
subjects.

Administrative
Data
Component

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Nevada Adult)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Ohio Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The Ohio Community-Oriented Reentry Program (CORE) entails intensive case management
services and treatment programs in the domain areas that are identified as necessary for
program participants. Prior to release, CORE participants meet frequently with the Reentry
Management Team (RMT) formed for them, which uses the results of various risk/needs
assessments to develop a Reentry Accountability Plan (RAP) and consists of: institutional and
community case managers, community reentry coordinators, institutional treatment and
programming personnel, Adult Parole Authority officers, faith-based organizations, regional
service coordinators, and institutional and community service providers. CORE participants
also receive two programs offered to CORE participants only: a career exploration program and
a CD-ROM-based program that entails the creation of career and financial goals, a resume, a
job search strategy, a personal budget, and a savings and investment plan. For any programs
that the RMT identifies as important or necessary, CORE participants are prioritized in the
institutions. After release, CORE participants meet regularly with their Community Reentry
Management Team (CRMT), which consists of the same stakeholders as the RMT (with the
exception of institutional based providers) and have access to funds for things such as
emergency housing, rental assistance, bus passes, phone cards, and application fees.
The program also provides for intensive wrap around services for offenders designated as
seriously mentally ill (with a disproportionately high rates of co-occurring mental health and
substance abuse problems evident in the CORE population).
Differences Between SVORI and “Business as Usual”
All inmates receive formal risk and needs assessments upon entry into the Ohio prison system.
In addition, all “reentry intensive” cases (individuals considered to be at high risk, based on the
risk assessment) receive a dynamic needs/domain assessment. However, CORE participants
have this assessment reviewed on a much more frequent basis (monthly, as opposed to
quarterly). All “reentry intensive” inmates are assigned an RMT and have an RAP developed for
them. However, CORE participants meet with their RMT at least once a month whereas other
offenders meet only twice a year with their RMT. Another difference pre-release is that CORE
participants are the only inmates who receive the career exploration program and the CD-ROMbased program described above, and that CORE participants are prioritized for existing prerelease services.
Upon release, all reentry intensive inmates meet with a CRMT at least once a year, with CORE
participants meeting more frequently (monthly for the first 6 months and then bimonthly). The
only differences between CORE and non-CORE offenders post-release is that CORE offenders
meet with their CRMT more frequently, are supervised by parole officers that have smaller
average caseloads, and are able to access funds for miscellaneous costs.
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
CORE participants were enrolled from 10 Ohio DRC facilities. In order to be eligible,
participants must be released to one of three Ohio Counties (Allen, Cuyahoga, and Franklin).

Other Eligibility Criteria
Eligible offenders must have been convicted of murder robbery, aggravated assault, burglary,
theft, and arson; convicted of high felony level such as murder, felony 1, 2, or 3 or be
considered a serious offender (a serious offender was defined as anyone with more than two
felony convictions or with a lengthy juvenile history and confined as an adult at an early age); be
between the ages of 18-35 at the time of program enrollment; have a static risk assessment
score of 5 or higher; have more than one dynamic domain at 3 or more need level; have a
sentence of incarceration of 12 months or more; and be on parole or post-release control for at
least one year following release from prison.
Sex offenders were initially excluded from participation, but were later included if they met
criteria as a ‘seriously mentally ill’ offender under supplemental funding guidelines and met
other federal criteria.
After being screened and meeting with program staff to confirm interest in the program, eligible
inmates were randomly assigned to receive either CORE programming or “business as usual”
for offenders classified as “reentry intensive”.

Voluntary/Mandatory Nature of Program
Participation in the program was voluntary.

Program Enrollment Period
The program began enrolling participants in January of 2003. It continued enrollment until
December 2005, enrolling a total of 202 individuals over the course of the grant.

Finally, it is important to understand the relationship between the population identified as being
eligible for inclusion in the interview component (i.e., the individuals who were identified through
the procedures outlined in Table 1 and are included in the “master file” for the program) and the
final group actually enrolled in the baseline interview component (the “baseline sample”), as well
as subsequent interview waves.2 Table 2 provides detailed information about all cases fielded
at each interview wave, listing the number of cases assigned to each final code maintained in
the field management system developed for the interview component of the study. Data are
provided separately by group status (SVORI vs. comparison), gender, and interview wave.

Only individuals in the baseline sample were pursued for the subsequent interview waves.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Ohio Adult)

SVORI Population Selected for
Evaluation

Comparison Population Selected for
Evaluation

Interview Component

Administrative
Data
Component

Interview Component

Administrative
Data
Component

Selection Criteria: Individuals
randomly assigned to the
CORE program and who were
projected to be (and actually)
released during the baseline
enrollment period. Random
assignment entailed the
following: offenders who meet
the eligibility requirements met
with program staff to explain
the program, ascertain interest,
and make a final eligibility
determination; the remaining
inmates were randomly
assigned to the experimental
or control group.

All CORE
participants
enrolled from
program
inception to
December 2005.

Selection Criteria: Individuals
randomly assigned to the
control group and projected to
be (and actually) released
during the baseline enrollment
period. Random assignment
procedures described in the
“SVORI population selected for
Evaluation” column.

All comparison
subjects
identified as part
of random
assignment
process from
program
inception to
December 2005.

Sample Identification
Procedures: SL received files
from site staff listing all
experimental and control
cases, and identified those
with eligible release dates.

Sample Identification
Procedures: SL received files
from site staff listing all
experimental and control cases,
and identified those with eligible
release dates.

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Ohio Adult)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Oklahoma Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
PROTECT was established to fill service gaps and expand existing services for returning
prisoners. The program attempts to provide a broad set of needed services to participants.
Program participants are assigned to transition workers (in addition to the facility case manager,
which is part of “business as usual”) who provide intensive case management due to smaller
caseloads and more specialized skills. The transition workers work with clients prior to release
(developing a release plan and lining up housing, identification, and other needed services for
after the client is released) and follow the clients after release (making contact several times a
week during the early days of release). In addition, PROTECT clients (except for sex
offenders) are assigned to a single probation officer, in order to facilitate communication and
information-sharing between the transition worker and probation officer.
PROTECT participants can receive vocational training after release and receive intensive
employment services by one of the partnering agencies. PROTECT participants also have
access to housing (for women with children) and flexible funds (assistance with utility bills, bus
tokens, gas allowance, etc.) not available to other former prisoners.
Differences Between SVORI and “Business as Usual”
The general population of inmates receives a release plan completed by a case manager and
unit manager 6 months prior to leaving. Post-release, inmates on supervision are supervised by
probation or parole agents who have larger caseloads and only make contact once a month.
PROTECT participants are prioritized for some services and receive longer periods of services
in terms of intensity and length.
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
Inmates must be returning to Oklahoma City after incarceration in order to be eligible for the
program. Offenders from all state prisons, private (contract) prisons, community work centers,
and community corrections centers are eligible.

Other Eligibility Criteria
Inmates must be between the ages of 18-35 and have an LSI score of 29 or greater.

Voluntary/Mandatory Nature of Program
Participation in the program is voluntary.

Program Enrollment Period

The program began enrolling participants in September of 2003. As of March 2006, 139
participants had been enrolled (cumulative enrollment) and the program planned on continuing
enrollment indefinitely.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Oklahoma Adult)

SVORI Population Selected for
Evaluation
Interview Component

Administrative
Data
Component

Selection Criteria: PROTECT
participants from the 19 state
prisons (and, as of 11/23/04,
any of the four private prisons
and six community corrections
enters) who were projected to
be (and actually) released
during the baseline enrollment
period.

All PROTECT
participants ever
enrolled in the
PROTECT
program up until
March 2007

Sample Identification
Procedures: SL received files
(originally from program
director and then from the MIS
contact, who received the lists
from the program director)
listing all PROTECT
participants with expected
release dates within the next 3
months.

Comparison Population Selected for
Evaluation
Interview Component

Selection Criteria: Offenders
who meet the PROTECT age
and LSI criteria but who are
returning to Tulsa county
(based on pre-incarceration
county of residence) and
projected to be (and actually)
released during the baseline
enrollment period.
Sample Identification
Procedures: SL received files
(originally from program
director, who received lists from
MIS contact, and then from the
MIS contact himself) listing all
individuals who met the
comparison group criteria and
had with expected release
dates within the next 3 months.

Administrative
Data
Component
All comparison
individuals
included in
sampling frame
for interview
component.

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Oklahoma Adult)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Pennsylvania Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The Erie PA Reentry Project (EPRP) provide a broad range of services (primarily focusing on
the pre-release phase, which is implemented by a contractor, Greater Area Community Action
Committee) and continuity of care. Under the EPRP, a community support specialist begins
working with the offender and the institutional case manager to oversee pre-release services
and plan for transition to the community. The community support specialist provides continued
services to the offender throughout the phases of the program. All EPRP participants are
paroled to a community corrections facility in Erie as part of their transition back into the
community. The average length of stay for EPRP participants at the community corrections
facilities is 90 days.
Differences Between SVORI and “Business as Usual”
All offenders are offered a curriculum-based pre-release program (including education, job
training, resume development, interviewing skills, etc.) and have an institutional case manager.
However, EPRP participants have the assistance of the community support specialist (which
increases access to and intensity of services) and receive longer follow-up care (up to one year
after release). There are also specific accommodations made for EPRP participants for
substance abuse treatment (EPRP are given priority for group therapy, have residential beds
reserved for them, receive drug and alcohol assessments, and have extended treatment
available) and financial assistance (EPRP has a pool of funding for participants to attend
training and obtain housing)
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
The program serves male and female adults from any of the secure correctional institutions
(SCI) who are returning to Erie County. EPRP participants are transferred to Albion SCI (males)
or Cambridge Springs SCI (females) to begin EPRP programming.

Other Eligibility Criteria
Other eligibility criteria are that the offender is between 18 and 35 years old and is going to be
released on state parole or state re-parole. All EPRP participants are paroled to a community
corrections facility in Erie.

Voluntary/Mandatory Nature of Program
Participation in the program is voluntary.

Program Enrollment Period
The program began enrolling participants in March of 2004 and stopped enrollment in
December of 2006. As of March 2006, the program had enrolled 133 participants (cumulative
enrollment) over the course of the grant.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Pennsylvania Adult)

SVORI Population Selected for
Evaluation
Interview Component

Selection Criteria: Interview
participants identified from
community corrections centers,
not prisons. Male EPRP
participants residing in Erie
CCC and female EPRP
participants at Gaudenzia who
were projected to be (and
actually) released from the
community corrections facility
to the community during the
baseline enrollment period.
Sample Identification
Procedures:

Administrative
Data
Component

Comparison Population Selected for
Evaluation
Interview Component

Selection Criteria: Male state
parolees, state re-parolees,
technical parole violators
(TPV’s) with community parole
center (CPC) placements
("halfway-backs"), and prerelease cases who were
between 18 and 35 years,
returning to Erie, Crawford, or
Warren County, and projected
to be (and actually) released
from Erie CCC and Gateway
Erie (another treatment facility
contracted by DOC) during the
baseline enrollment period. No
female comparison group.
Sample Identification
Procedures: FIs confirm that
the individual is returning to
Erie, Crawford, or Warren
County.

Administrative
Data
Component

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Pennsylvania Adult)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C
n/a
n/a
n/a

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
South Carolina Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The South Carolina adult SVORI program was designed to provide desperately lacking services
and to increase coordination (particularly improving the link from the facility to the community).
The SVORI program entailed a Transition Coordinator (hired by the grant) being assigned to
each of seven medium and maximum security facilities served by SVORI. The Transition
Coordinators served as the case managers for SVORI participants and had the additional
responsibility of identifying community resources in the counties surrounding the facility. While
incarcerated, SVORI participants received intensive case management and services provided
directly by the Transition Coordinators (through curriculum topics, groups, etc.). The program
also entailed SVORI participants having the Transition Coordinators check in with them
approximately 3 months after release. There were no other post-release components of the
program.
Differences Between SVORI and “Business as Usual”
Because of severe budget cuts for SCDC over the past several years, the general prison
population1 receives virtually no services. For the general prison population (including the
specific groups of comparison offenders selected for the multi-site evaluation), the only prerelease case management they could receive was offered by the classification case managers
(who are primarily responsible for classifying offenders and who have caseloads of 250-350
individuals), clinical counselors (for the small number of offenders who are referred to the
clinical counselor), and facility Chaplains. There was no equivalent of the Transition
Coordinator for non-SVORI offenders, and no follow-up after release.
All SVORI and virtually all comparison group members received post-release supervision by a
parole officer. The SVORI grant does not appear to have affected post-release supervision for
individual offenders. The program could have affected service delivery after release, however,
in that SVORI participants may have been more likely to leave the facility with referrals in hand
or post-release services already set up, or received additional help from the post-release
contact made by the facility Transition Coordinator.
Detailed information on specific services available pre- and post-release to both SVORI
participants and the group of comparable, non-SVORI offenders identified in the multi-site
evaluation is included in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
The program served individuals incarcerated in one of the 7 medium or maximum security
institutions (Kershaw, Leath [women’s], Lee, MacDougall, Ridgeland, Tyger River [sex offenders
only], and Wateree River). Post-release community was not an eligibility criterion; the program
served individuals returning to any county in the state.

As mandated by law, youthful offenders do receive some services.

Other Eligibility Criteria
The program served adults between the ages of 17-35 (with no age restrictions for sex
offenders), who had post-release supervision (TIS sentences/splits, CSP, a few who are
paroled, and possibly some that were on supervised furlough), had a prior conviction (this
criterion did not apply to sex offenders), were not sentenced under the Youthful Offender Act,
and who were 12 months from their projected release date.

Voluntary/Mandatory Nature of Program
Eligible prisoners were offered the program and those who agreed to participate were enrolled
in the program.

Program Enrollment Period
The program began enrolling participants in November 2003, with date of program inception
varying across facilities. As of 4/01/04, all facilities except Kershaw and Ridgeland had enrolled
participants. SVORI programming at Tyger River was suspended from June through November
of 2004 (due to staff turnover). As of March 2006, the program had enrolled a total of 649
individuals (cumulatively) and was planning on continuing enrollment indefinitely.

SVORI and Comparison Populations Selected for the Evaluation
The impact component of the multi-site evaluation of SVORI entailed the identification of a
specific group of SVORI participants and comparable, non-SVORI offenders for which
individual-level data was obtained. Table 1 (below) is intended to convey information about how
these individuals were identified. The table provides details regarding the specific SVORI and
comparison groups included in the two main data collection components of the impact
evaluation: the longitudinal offender interview data and the administrative data.2 For the
interview component, the table presents the selection criteria used to identify the appropriate
group and the specific procedures through which the sample was identified. For the
administrative data component, the table describes the specific groups for which we were able
to obtain data. For both data collection components, it should be noted that our treatment group
assignment protocol specified that individuals are considered to have enrolled in SVORI if they
were in the program for more than one day.
Finally, it is important to understand the relationship between the population identified as being
eligible for inclusion in the interview component (i.e., the individuals who were identified through
the procedures outlined in Table 1 and are included in the “master file” for the program) and the
final group actually enrolled in the baseline interview component (the “baseline sample”), as well
as subsequent interview waves.3 Table 2 provides detailed information about all cases fielded
at each interview wave, listing the number of cases assigned to each final code maintained in
the field management system developed for the interview component of the study. Data are
provided separately by group status (SVORI vs. comparison), gender, and interview wave.

Some administrative data analyses will be limited to only the sample for which interviews were
conducted (the “interview-only” sample). However other analyses will be based entirely on administrative
data and are designed to include an expanded sample of SVORI and comparison individuals.
3
Only individuals in the baseline sample were pursued for the subsequent interview waves.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(South Carolina Adult)

SVORI Population Selected for
Evaluation

Comparison Population Selected for
Evaluation

Interview Component

Administrative
Data
Component

Interview Component

Administrative
Data
Component

Selection Criteria: All
individuals enrolled in SVORI
and who were projected to be
(and actually) released during
the baseline enrollment period.

All SVORI
participants
released in 2004
and 2005

Selection Criteria: Three
comparison groups were
identified, as follows: a)
individuals from the 7 SVORI
facilities who met all SVORI
eligibility criteria (including postrelease supervision) b) non-sex
offenders from Tyger River who
met all SVORI eligibility criteria
(including post-release
supervision), and c) individuals
who met all SVORI eligibility
criteria (including post-release
supervision) but who were
incarcerated in three nonSVORI medium or maximum
security facilities (Camille Griffin
Graham [women’s], Allendale,
and Evans). For all 3
comparison groups, individuals
must have been projected to be
(and actually) released during
the baseline enrollment period.

From 2004-2005
release cohort
files, will subset
to individuals
meeting SVORI
eligibility criteria
for an “expanded”
comparison
group

Sample Identification
Procedures: SCDC MIS staff
member wrote a program to
identify eligible SVORI and
comparison group members
and submitted monthly files

Sample Identification
Procedures: SCDC MIS staff
member wrote a program to
identify eligible SVORI and
comparison group members
and submitted a monthly file;
Although we initially planned to
recruit an additional comparison
group that did not have the
post-release supervision
requirement, we had a
sufficiently high number of
comparison group members to
preclude this necessity.

Table 2.

Description
Number of Cases
identified for
Evaluation
Number of Cases
Fielded
Number of Cases
Interviewed

Summary Information about SVORI (S) and Comparison (C) Cases
Fielded at each Interview Wave (South Carolina Adult)
Baseline
Males
Females
S
C
S
C

185

175

3 Month
Males
Females
S
C
S
C

123

104

9 Month
Males
Females
S
C
S
C

119

15 Month
Males
Females
S
C
S
C

107

SVORI Multi-Site Evaluation Sample Selection Procedures:
South Carolina Juvenile Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The South Carolina juvenile program is intended to improve service coordination and was
heavily influenced by the Intensive Aftercare Program (IAP) model. For the SVORI grant, DJJ
hired specialized community caseworkers (called “Reintegration Coordinators”) to serve SVORI
youth. The SVORI participants are assigned to a Reintegration Coordinator in place of the
community caseworker that all DJJ youth who are identified for post-release supervision are
assigned. Each of the five counties served by SVORI also developed a Planning and Review
(P&R) Team that consists of various local service providers and education representatives and
that meets on a monthly basis to review individual cases prior to release. The team assists the
Reintegration Coordinator in developing the youth’s aftercare treatment plan and identifying
needed services for the youth. Following release, the P&R Team reviews the youth’s progress
and aftercare plan, offers services, and helps ensure that services are provided following
release.
Differences Between SVORI and “Business as Usual”
The primary difference between youth identified for the SVORI program and non-SVORI youth
is assignment to a Reintegration Coordinator rather than a community caseworker. This
assignment occurs during incarceration for youth who are identified for post-release supervision.
Theoretically, both the community caseworkers and the Reintegration Coordinators work to
prepare the youth for release. However, the community caseworkers do not have a specialized
caseload (they supervise youth on probation as well as parole, and have much larger caseloads
overall) and typically do not provide services to youth who are committed to a DJJ facility until
after release. The SVORI Reintegration Coordinators have a specialized (and smaller)
caseload, begin the transition/reintegration planning process earlier, provide a more intensive
level of case management and supervision, and receive multidisciplinary assistance from the
P&R Team.
All youth in DJJ facilities (including SVORI and comparison) are also assigned to an institutional
case manager, who provides case management during the period of incarceration. This
component is the same for SVORI and comparison youth.

SVORI Population Served
Geographical Eligibility Criteria
The program served individuals incarcerated in one of the maximum security institutions in the
Broad River Complex (Birchwood Institution, John G. Richards Institution, and Willow Lane
Institution [females]) or any of the medium security community corrections facilities (including
Camp Bennettsville I and II, Camp Sandhills, Camp White Pines I and II, Generations Bridges
Camp, and Aspen Alternative). The only institutions that the program excluded from SVORI
were the 3 evaluation centers and the juvenile detention center. Regarding post-release
eligibility criteria, the program only served youth who were returning to the following counties:
Orangeburg/Calhoun, Dorchester, Spartanburg, and Florence.

Other Eligibility Criteria
The program serves juvenile males and females between the ages of 14 and 17 who (1) have
committed a serious or violent offense, (2) have committed a technical violation related to a
serious or violent offense, or (3) are chronic offenders. All participants must be assigned to a
Department of Juvenile Justice (DJJ) facility with either a determinate commitment of at least 90
days or an indeterminate commitment of at least 3-6 months. Offenders who are committed to
DJJ with a requirement that they transfer to an adult facility at the age of 18 are excluded from
the program. The SVORI case managers are responsible for identifying eligible participants.
Voluntary/Mandatory Nature of Program
Assignment to SVORI was not voluntary (it was the “standard” way that youth were released).

Program Enrollment Period
The program began enrolling participants on October 2003, but the date of program inception
varied across post-release counties. Spartanburg and Florence were rolled in last (October
2004). As of March 2006, the program had enrolled 120 individuals (cumulatively) and planned
to continue enrolling indefinitely.

SVORI Population Selected for
Evaluation
Interview Component

Selection Criteria: All
individuals enrolled in SVORI
and who were projected to be
(and actually) released during
the baseline enrollment period.
Sample Identification
Procedures: SL reviewed the
monthly parole board release
lists, determined SVORI
enrollment status based on
SVORI enrollment rosters
submitted (monthly) from
program director, and
submitted the file to SC DJJ
research staff member. SC
DJJ research staff member
added additional data
elements required for fielding
not contained on the original
lists.

Administrative
Data
Component
All individuals
enrolled in
SVORI from
program
inception and
released at any
point up through
12/31/2006

Comparison Population Selected for
Evaluation
Interview Component

Selection Criteria: The
comparison group primarily
consisted of (a) youth
incarcerated in the same
facilities served by SVORI but
who were committed from
different counties (Greenville,
Aiken, and Sumter) but also
included a small number of (b)
youth incarcerated in the same
facilities as served by SVORI
and committed from the same
counties (these comparison
group members were likely not
enrolled in SVORI because of
case flow caps established for
the reentry coordinators)
Sample Identification
Procedures: SL screened the
monthly parole board release
lists (screened out non-study
facilities, non-study counties,
and youth who had not met
their minimum time served),
determined SVORI/comparison
status based on SVORI
enrollment rosters submitted
(monthly) from program
director, and submitted the file
to SC DJJ research staff
member. SC DJJ research
staff member added additional
data elements required for
fielding not contained on the
original lists.

Administrative
Data
Component
All comparison
individuals
included in
sampling frame
for interview
component

Table 2.

Description
Number of Cases
identified for
Evaluation
Number of Cases
Fielded
Number of Cases
Interviewed

Summary Information about SVORI (S) and Comparison (C) Cases
Fielded at each Interview Wave (South Carolina Juvenile)
Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

SVORI Multi-Site Evaluation Sample Selection Procedures:
Washington Adult Program
Description of the SVORI Program and “Business as Usual”
Overview of SVORI
The Washington Going Home (GH) Project provided reentry services to the highest risk
offenders under the supervision of the Washington Department of Correction (DOC), using a
team approach that emphasized community involvement. The GH program built on an existing
Risk Management Intensive Transition (RMIT) process—the standard post-release planning
process for high risk/high needs offenders in the WA DOC system—which includes developing
an Offender Accountability Plan (described below). In addition, GH program participants were
connected with a Community Advisor (CA) and Neighborhood Readiness Team (NRT) prior to
release. The Community Advisor was an individual in each county of release who marshaled
community resources for returning offenders; essentially serving as a conduit between the
community and the Going Home program. The NRT in each county was made up of individuals
from the offender’s community of return and supports the offender during release. The NRT was
expected to meet with returning offenders via video-conference prior to release, as well as
regularly post-release, in order to provide continuity between the pre- and post-release phases
and increase offender “buy-in” to the program.
In addition to the CA and NRT, the GH program was intended to increase reentry coordination
service availability (e.g., use of video-conferencing, funding for community mental health
services). Enhanced coordination began 18 months prior to release when Going Home
participants were identified for inclusion in the program and intensified 9 months prior to release
when the post-release components (e.g., NRT, CA) began meeting with the offender to plan for
release. After release, GH participants were assigned to a dedicated (i.e., assigned only to GH
participants) Risk Management Specialist (RMS), who provided intensive case management
services and worked with the CA to identify community resources that would meet each
participant’s reentry needs.
Although the program also targeted juvenile offenders, using a different model, the juvenile
component was not included in the multi-site evaluation (due to low estimated enrollment) and is
therefore not described further in this document.
Differences Between SVORI and “Business as Usual”
In Washington, all high risk/high needs offenders go through an existing RMIT process, which
begins 28 months prior to release. The RMIT process consists of developing a Risk
Management Team (RMT) which identifies resources based on an offender’s risk factors. This
process culminates in an Offender Accountability Plan (OAP) which guides the post-release
phase of reentry. The membership of the offender-specific RMT is often modified between the
pre-release and post-release phases to best target the needs of the specific offender. The RMT
and RMS work with the offender throughout the mandatory community custody period. Prior to
release, all prisoners receive two pre-release curricula (a life skills course and employment
course) and have access to numerous services.
GH participants were prioritized for several pre-release services (including substance abuse,
tutoring, employment seminars, housing assistance, and videoconferencing). The primary
difference between GH and “business as usual” was the addition of several post-release
components, all of which were involved with the offender prior to release. These include the use

of GH-only RMSs, who had greater access to community resources and more concentrated
time with each offender, and the involvement of the NRT and the CA, both of which were
available only to GH participants and provided a unique experience to returning offenders
because they were outside the “system”.
Detailed information on specific services available pre- and post-release to both GHparticipants
and the group of comparable, non-GH offenders identified in the multi-site evaluation is included
in Appendix ?.

SVORI Population Served
Geographical Eligibility Criteria
The program served offenders releasing to King, Pierce, or Spokane counties and recruited
participants from Ithe Washington State Penitentiary (Walla Walla), Monroe Correctional
Complex-Special Offender Unit (Monroe), Monroe Correctional Complex -Twin Rivers
Corrections Center (Monroe) for male enrollees and Washington Corrections Center for Women
(Gig Harbor) for female enrollees.

Other Eligibility Criteria
The program served offenders under the age of 35, who met one of the following categories: 1)
“high risk” (defined as having an LSI score of at least 41 or having at least one violent
conviction, (2) “high needs” (disabled or have other health problems), 3) sex offenders, 4)
dangerous mentally ill offenders, or 5) offenders who are an imminent risk or threat.

Voluntary/Mandatory Nature of Program
Participation in the SVORI program was mandatory.

Program Enrollment Period
The program began enrolling participants in August 2005. Enrollment ceased in February 2006,
with 150 total individuals enrolled over the course of the grant.

Only individuals in the baseline sample were pursued for the subsequent interview waves.

Table 1.

Selection Criteria for Interview and Administrative Data Components
(Washington Adult)

SVORI Population Selected for
Evaluation

Comparison Population Selected for
Evaluation

Interview Component

Administrative
Data
Component

Interview Component

Administrative
Data
Component

Selection Criteria: GH
participants returning to King
and Pierce Counties (Spokane
was excluded) who were
projected to be (and actually)
released during the baseline
enrollment period.

All individuals
ever enrolled in
GH (including
those released to
King, Pierce, and
Spokane)

Selection Criteria: Individuals
who met program criteria, were
incarcerated in 6 correctional
facilities and 11 work release
facilities (near Seattle and in
and around Walla Walla) in
which GH programming was
not offered, were returning to
Pierce and King Counties, and
were projected to be (and
actually) released during the
baseline enrollment period.

All individuals
who met program
criteria, were
incarcerated in 6
correctional
facilities and 11
work release
facilities (near
Seattle and in
and around Walla
Walla) in which
GH programming
was not offered,
were returning to
Pierce, King, and
Spokane
Counties, and
were released
from 2004-2006

Sample Identification
Procedures: SL received lists
of all adult GH participants
(excluding those being
released to Spokane County).

Sample Identification
Procedures: SL received lists of
comparison subjects meeting
the selection criteria (excluding
those being released to
Spokane County)

Table 2.

Summary Information about SVORI (S) and Comparison (C) Cases at
each Interview Wave (Washington Adult)

Description
Number of Cases identified
for Evaluation
Number of Cases Fielded
Number of Cases
Interviewed

Baseline
Males
Females
S
C
S
C

3 Month
Males
Females
S
C
S
C
n/a

n/a

9 Month
Males
Females
S
C
S
C
n/a

n/a

15 Month
Males
Females
S
C
S
C
n/a

n/a

Appendix C. Comparisons of
Impact and Non–impact Sites
Exhibit C-1. Program director turnover among impact and non–impact sites

Impact Sites
Non–impact Sites
Programs
Programs
Adult Programs
Mean
(N)
Mean
(N)
Number of individuals serving as program directora
Adult sites
1.75
12
1.81
38
Juvenile sites
2.00
4
1.39
31
a

All Sites
Programs
Mean
(N)
1.80
1.45

50
35

Source: 2006 program director survey.

C-1

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit C-2a. Program characteristics of adult impact and non–impact sitesa

Impact Sites
Adult Programs
%
N
Primary program phase focus (52 programs reporting)
Pre-release
8.3%
1
Post-release
25.0%
3
Both pre- and post-release
66.7%
8
Population focus (52 programs reporting)
General serious and violent offenders
91.7%
11
Offenders with specific needs
8.3%
1
Service focus (51 programs reporting)
General service focus
91.7%
11
Targeted service provision
8.3%
1
Program participation (52 programs reporting)
Voluntary
75.0%
9
Mandatory
25.0%
3
How SVORI funds used (51 programs reporting)
Fill service gaps
27.3%
3
Expand existing services
18.2%
2
Start a new program
54.5%
6
a

Non–Impact
Sites
%
N

All Sites
%
N

—
32.5%
67.5%

0
13
27

1.9%
30.8%
67.3%

1
16
35

82.5%
17.5%

33
7

84.6%
15.4%

44
8

89.7%
10.3%

35
4

90.2%
9.8%

46
5

82.5%
17.5%

33
7

80.8%
19.2%

42
10

45.0%
37.5%
17.5%

18
15
7

41.2%
33.3%
25.5%

21
17
13

Source: 2005 program director survey.

Exhibit C-2b. Program characteristics of juvenile impact and non–impact sitesa

Juvenile Programs
Primary program phase focus
Pre-release
Post-release
Both pre- and post-release
Population focus
General serious and violent offenders
Offenders with specific needs
Service focus
General service focus
Targeted service provision
Program participation
Voluntary
Mandatory
How SVORI funds used
Fill service gaps
Expand existing services
Start a new program
a

Impact Sites
%
N

Non–impact
Sites
%
N

All Sites
%
N

—
—
100.0%

0
0
4

6.1%
18.2%
75.7%

2
6
25

5.4%
16.2%
78.4%

2
6
29

100.0%
—

4
0

84.8%
15.2%

28
5

86.5%
13.5%

32
5

75.0%
25.0%

3
1

81.8%
18.2%

27
6

81.1%
18.9%

30
7

50.0%
50.0%

2
2

48.5%
51.5%

16
17

48.6%
51.4%

18
19

75.0%
25.0%
—

3
1
0

42.4%
33.3%
24.2%

14
11
8

46.0%
32.4%
21.6%

17
12
8

Source: 2005 Program Director Survey, 37 programs responding: 4 impact and 33 non–impact sites.

C-2

Appendix C — Comparisons of Impact and Non–impact Sites

Exhibit C-3a. Outcome foci among adult impact and non–impact sitesa

Impact Sites
Adult Programs
%
N
Outcomes targeted (program director ranked in top 3)
Community integration
66.70%
8
Employment
58.30%
7
Improved decision-making or self-sufficiency
58.30%
7
Reduced substance use
50.00%
6
Housing
33.30%
4
Educational attainment
16.70%
2
Improved physical and/or mental health
8.30%
1
Family reunification/functioning
8.30%
1
Outcomes fair to determine program effectiveness
Community integration/connectedness
100.00% 12
Employment
100.00% 12
Reduced recidivism
91.70% 11
Reduced substance use
75.00%
9
Family reunification/functioning
75.00%
9
Educational attainment
66.70%
8
Housing
66.70%
8
Improved physical and/or mental health
50.00%
6
a

Non–impact
Sites
%
N

All Sites
%
N

48.72%
66.67%
43.59%
43.59%
46.15%
12.82%
12.82%
12.82%

19
26
17
17
18
5
5
5

52.94%
64.71%
47.06%
45.10%
43.14%
13.73%
11.76%
11.76%

27
33
24
23
22
7
6
6

69.23%
84.62%
79.49%
71.79%
38.46%
38.46%
66.67%
46.15%

27
33
31
28
15
15
26
18

76.47%
88.24%
82.35%
72.55%
47.06%
45.10%
66.67%
47.06%

39
45
42
37
24
23
34
24

Source: 2006 program director survey, 51 programs reporting: 12 impact and 39 non–impact sites.

Exhibit C-3b. Outcome foci among juvenile impact and non–impact sitesa

Impact Sites
%
N

Juvenile Programs
Outcomes targeted (ranked in top 3)
Community integration
100.0%
Educational attainment
75.0%
Employment
50.0%
Family reunification/functioning
25.0%
Improved physical and/or mental health
0.0%
Improved decision-making or self-sufficiency
25.0%
Reduced substance use
25.0%
Housing
0.0%
Outcomes fair to determine program effectiveness
Community integration/connectedness
100.0%
Educational attainment
100.0%
Employment
100.0%
Family reunification/functioning
75.0%
Improved physical and/or mental health
50.0%
Reduced recidivism
100.0%
Reduced substance use
50.0%
Housing
50.0%
a

Non–impact
Sites
%
N

All Sites
%
N

4
3
2
1
0
1
1
0

54.8%
58.1%
25.8%
45.2%
25.8%
41.9%
35.5%
3.2%

17
18
8
14
8
13
11
1

60.0%
60.0%
28.6%
42.9%
22.9%
40.0%
34.3%
2.9%

21
21
10
15
8
14
12
1

4
4
4
3
2
4
2
2

71.0%
80.6%
64.5%
58.1%
48.4%
93.5%
64.5%
29.0%

22
25
20
18
15
29
20
9

74.3%
82.9%
68.6%
60.0%
48.6%
94.3%
62.9%
31.4%

26
29
24
21
17
33
22
11

Source: 2006 program director survey, 35 programs reporting: 4 impact and 31 non–impact sites.

C-3

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit C-4a. Mean proportion of SVORI offenders receiving pre-release services in adult
sites, by site type (as reported by program directors)

Impact
Pre-release Services in Adult Sites
Sites
Bundle 1: Coordination Services
Risk assessment
0.88 (0.30)
Needs assessment
0.88 (0.30)
Treatment/release plan development
0.91 (0.29)
Bundle 2: Transitional Services
Legal assistance
0.24 (0.25)
0.60 (0.42)
Assistance obtaining identification (e.g., driver’s license,
Social Security card)
0.44 (0.35)
Assistance obtaining benefits and completing applications
(e.g., Medicaid, disability)
Financial support/emergency assistance
0.45 (0.43)
Peer support groups
0.42 (0.38)
One-on-one mentoring
0.50 (0.42)
Housing placements or referrals
0.54 (0.39)
Bundle 3: Health Services
Comprehensive drug treatment programs
0.31 (0.20)
AA/NA
0.39 (0.33)
Counseling sessions
0.60 (0.37)
Mental health services
0.35 (0.31)
Anger management/violence counseling
0.52 (0.34)
Medical services
0.70 (0.41)
Dental services
0.70 (0.41)
Bundle 4: Employment, Education, and Skills Development Services
Education/GED/tutoring/literacy
0.67 (0.29)
Vocational training
0.33 (0.29)
Employment referrals/job placement
0.51 (0.35)
Résumé and interviewing skills development
0.68 (0.33)
Work-release program
0.17 (0.30)
Cognitive skills development/behavioral programming
0.49 (0.35)
Life skills training
0.71 (0.39)
Bundle 5: Family Services
Domestic violence services
0.38 (0.35)
Parenting skills development
0.40 (0.37)
Family reunification
0.39 (0.34)
Family counseling
0.18 (0.26)

Mean (SD)
Non–
impact
Sites

All Sites

0.93 (0.21)
0.93 (0.21)
0.92 (0.24)

0.92 (0.23)
0.92 (0.23)
0.92 (0.25)

0.40 (0.45)
0.63 (0.41)

0.37 (0.41)
0.62 (0.41)

0.47 (0.44)

0.46 (0.42)

0.27 (0.40)
0.47 (0.42)
0.37 (0.40)
0.56 (0.38)

0.31 (0.41)
0.46 (0.41)
0.40 (0.40)
0.56 (0.38)

0.37 (0.36)
0.46 (0.37)
0.71 (0.39)
0.51 (0.38)
0.64 (0.38)
0.82 (0.33 )
0.79 (0.36 )

0.36 (0.33)
0.44 (0.36)
0.69 (0.38)
0.47 (0.37)
0.61 (0.37)
0.79 (0.35)
0.77 (0.37)

0.59 (0.34)
0.39 (0.33)
0.51 (0.45)
0.66 (0.41 )
0.23 (0.33)
0.70 (0.37)
0.74 (0.34)

0.61 (0.33)
0.38 (0.32)
0.51 (0.43)
0.67 (0.39)
0.22 (0.32)
0.65 (0.37)
0.74 (0.35)

0.32 (0.39)
0.52 (0.40)
0.41 (0.40)
0.13 (0.27)

0.33 (0.38)
0.49 (0.39)
0.41 (0.38)
0.14 (0.27)

Note: Source: 2005 program director survey. Values were calculated by taking the midpoint of the response
categories (0%, 1%–25%, 26%–50%, 51%–75%, 76%–99%, and 100%) reported by the SVORI adult program
directors for each of the services. AA = Alcoholics Anonymous, GED = general educational development, NA =
Narcotics Anonymous.

C-4

Appendix C — Comparisons of Impact and Non–impact Sites

Exhibit C-4b. Mean proportion of SVORI offenders receiving post-release services in adult
sites, by site type (as reported by program directors)

Impact
Post-release Services in Adult Sites
Sites
Bundle 1: Coordination Services
Risk assessment
0.86 (0.29)
Needs assessment
0.81 (0.35)
Treatment/release plan development
0.91 (0.29)
Formal post-release supervision
0.82 (0.27)
Bundle 2: Transitional Services
Legal assistance
0.21 (0.24)
0.46 (0.35)
Assistance obtaining identification (e.g., driver’s license,
Social Security card)
0.47 (0.35)
Assistance obtaining benefits and completing applications
(e.g., Medicaid, disability)
Financial support/emergency assistance
0.63 (0.29)
Peer support groups
0.37 (0.34)
One-on-one mentoring
0.45 (0.38)
Housing placements or referrals
0.66 (0.29)
Transportation
0.60 (0.34)
Bundle 3: Health Services
Comprehensive drug treatment programs
0.32 (0.28)
AA/NA
0.44 (0.30)
Counseling sessions
0.48 (0.31)
Mental health services
0.35 (0.30)
Anger management/violence counseling
0.35 (0.25)
Medical services
0.30 (0.32)
Dental services
0.24 (0.23)
Bundle 4: Employment, Education, and Skills Development Services
Education/GED/tutoring/literacy
0.37 (0.29)
Vocational training
0.34 (0.30)
Employment referrals/job placement
0.65 (0.33)
Resume and interviewing skills development
0.56 (0.37)
Work-release program
0.24 (0.39)
Cognitive skills development/behavioral programming
0.46 (0.35)
Life skills training
0.62 (0.37)
Bundle 5: Family Services
Domestic violence services
0.32 (0.32)
Parenting skills development
0.48 (0.34)
Family reunification
0.50 (0.34)
Family counseling
0.24 (0.25)

Mean (SD)
Non–
impact
Sites

All Sites

0.91 (0.27)
0.91 (0.27)
0.92 (0.24)
0.96 (0.08)

0.89 (0.27)
0.88 (0.29)
0.92 (0.25)
0.92 (0.16)

0.16 (0.27)
0.62 (0.41)

0.17 (0.26)
0.58 (0.40)

0.58 (0.42)

0.55 (0.40)

0.56 (0.40)
0.39 (0.40)
0.27 (0.31)
0.55 (0.38)
0.53 (0.37)

0.57 (0.37)
0.39 (0.39)
0.31 (0.33)
0.58 (0.36)
0.55 (0.36)

0.31 (0.31)
0.45 (0.30)
0.69 (0.36)
0.43 (0.37)
0.45 (0.36)
0.37 (0.39)
0.27 (0.36)

0.31 (0.30)
0.45 (0.30)
0.64 (0.36)
0.41 (0.35)
0.42 (0.34)
0.35 (0.37)
0.26 (0.33)

0.38 (0.37)
0.36 (0.36)
0.76 (0.31)
0.70 (0.38)
0.09 (0.20)
0.54 (0.40)
0.53 (0.42)

0.38 (0.35)
0.35 (0.34)
0.73 (0.31)
0.67 (0.38)
0.13 (0.26)
0.52 (0.39)
0.55 (0.40)

0.30 (0.35)
0.33 (0.34)
0.33 (0.35)
0.19 (0.27)

0.30 (0.34)
0.37 (0.34)
0.37 (0.35)
0.20 (0.26)

C-5

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit C-4c. Mean proportion of SVORI offenders receiving pre-release services in juvenile
sites, by site type (as reported by program directors)

Impact
Pre-release Services in Juvenile Sites
Sites
Bundle 1: Coordination Services
Risk assessment
1.0 (0.00)
Needs assessment
1.0 (0.00)
Treatment/release plan development
1.0 (0.00)
Bundle 2: Transitional Services
Legal assistance
0.17 (0.19)
0.97 (0.06)
Assistance obtaining identification (e.g., driver’s license,
Social Security card)
0.78 (0.44)
Assistance obtaining benefits and completing applications
(e.g., Medicaid, disability)
Financial support/emergency assistance
0.35 (0.33)
Peer support groups
0.63 (0.44)
One-on-one mentoring
0.26 (0.14)
Housing placements or referrals
0.28 (0.48)
Bundle 3: Health Services
Comprehensive drug treatment programs
0.66 (0.38)
AA/NA
0.44 (0.41)
Counseling sessions
0.72 (0.27)
Mental health services
0.54 (0.37)
Anger management/violence counseling
0.60 (0.29)
Medical services
1.0 (0.00)
Dental services
1.0 (0.00)
Bundle 4: Employment, Education, and Skills Development Services
Education/GED/tutoring/literacy
1.0 (0.0)
Vocational training
0.81 (0.30)
Employment referrals/job placement
0.75 (0.14)
Resume and interviewing skills development
0.93 (0.07)
Work-release program
0.25 (0.31)
Cognitive skills development/behavioral programming
0.94 (0.07)
Life skills training
0.94 (0.07)
Bundle 5: Family Services
Domestic violence services
0.38 (0.44)
Parenting skills development
0.44 (0.46)
Family reunification
0.66 (0.38)
Family counseling
0.69 (0.24)

Mean (SD)
Non–
impact
Sites

All Sites

0.96 (0.18)
0.95 (0.18)
1.0 (0.02)

0.96 0.17
0.96 0.17
1.0 0.02

0.42 (0.43)
0.46 (0.41)

0.40 (0.42)
0.52 (0.42)

0.45 (0.44)

0.49 (0.44)

0.30 (0.38)
0.72 (0.41)
0.54 (0.40)
0.48 (0.43)

0.31 (0.37)
0.71 (0.40)
0.51 (0.39)
0.46 (0.43)

0.48 (0.38)
0.38 (0.33)
0.94 (0.14)
0.63 (0.36)
0.82 (0.30)
0.94 (0.21)
0.87 (0.31)

0.50 (0.38)
0.39 (0.33)
0.91 (0.17)
0.62 (0.36)
0.80 (0.30)
0.94 (0.20)
0.89 (0.30)

0.97 (0.11)
0.54 (0.35)
0.45 (0.38)
0.60 (0.35)
0.07 (0.14)
0.88 (0.22)
0.82 (0.27)

0.97 (0.11)
0.57 (0.35)
0.48 (0.37)
0.63 (0.35)
0.09 (0.17)
0.89 (0.21)
0.83 (0.26)

0.28 (0.36)
0.30 (0.31)
0.67 (0.38)
0.57 (0.40)

0.29 (0.36)
0.31 (0.32)
0.67 (0.37)
0.58 (0.37)

Note: Source: 2005 program director survey. Values were calculated by taking the midpoint of the response
categories (0%, 1%–25%, 26%–50%, 51%–75%, 76%–99%, and 100%) reported by the SVORI juvenile program
directors for each of the services. AA = Alcoholics Anonymous, GED = general educational development, NA =
Narcotics Anonymous..

C-6

Appendix C — Comparisons of Impact and Non–impact Sites

Exhibit C-4d. Mean proportion of SVORI offenders receiving post-release services in
juvenile sites, by site type (as reported by program directors)

Impact
Post-release Services in Juvenile Sites
Sites
Bundle 1: Coordination Services
Risk assessment
0.25 (0.50)
Needs assessment
0.5 (0.58)
Treatment/release plan development
0.75 (0.50)
Formal post-release supervision
0.53 (0.54)
Bundle 2: Transitional Services
Legal assistance
0.13 (0.00)
0.75 (0.42)
Assistance obtaining identification (e.g., driver’s license,
Social Security card)
0.57 (0.50)
Assistance obtaining benefits and completing applications
(e.g., Medicaid, disability)
Financial support/emergency assistance
0.47 (0.42)
Peer support groups
0.35 (0.26)
One-on-one mentoring
0.19 (0.13)
Housing placements or referrals
0.35 (0.44)
Transportation
0.50 (0.32)
Bundle 3: Health Services
Comprehensive drug treatment programs
0.44 (0.41)
AA/NA
0.29 (0.28)
Counseling sessions
0.60 (0.29)
Mental health services
0.47 (0.37)
Anger management/violence counseling
0.38 (0.20)
Medical services
0.57 (0.50)
Dental services
0.57 (0.50 )
Bundle 4: Employment, Education, and Skills Development Services
Education/GED/tutoring/literacy
0.85 (0.15)
Vocational training
0.57 (0.24)
Employment referrals/job placement
0.69 (0.24)
Resume and interviewing skills development
0.57 (0.24)
Work-release program
0.03 (0.07)
Cognitive skills development/behavioral programming
0.60 (0.41)
Life skills training
0.63 (0.35)
Bundle 5: Family Services
Domestic violence services
0.26 (0.25)
Parenting skills development
0.35 (0.44)
Family reunification
0.66 (0.38)
Family counseling
0.50 (0.25)

Mean (SD)
Non–
impact
Sites

All Sites

0.77 (0.40)
0.81 (0.34)
0.96 (0.15)
0.90 (0.19)

0.71 (0.43)
0.78 (0.37)
0.94 (0.22)
0.86 (0.27)

0.40 (0.41)
0.61 (0.39)

0.37 (0.40)
0.62 (0.39)

0.55 (0.39)

0.56 (0.40)

0.47 (0.41)
0.52 (0.41)
0.47 (0.35)
0.51 (0.42)
0.67 (0.37)

0.47 (0.40)
0.51 (0.40)
0.44 (0.34)
0.49 (0.42)
0.65 (0.36)

0.44 (0.36)
0.35 (0.31)
0.77 (0.26)
0.53 (0.34)
0.61 (0.37)
0.64 (0.35)
0.51 (0.41)

0.44 (0.36)
0.34 (0.31)
0.75 (0.27)
0.52 (0.34)
0.59 (0.36)
0.63 (0.36)
0.52 (0.41)

0.86 (0.16)
0.43 (0.32)
0.67 (0.36)
0.64 (0.37)
0.12 (0.28)
0.65 (0.37)
0.65 (0.37)

0.86 (0.16)
0.44 (0.31)
0.67 (0.35)
0.63 (0.36)
0.11 (0.26)
0.64 (0.37)
0.65 (0.36)

0.20 (0.27)
0.36 (0.31)
0.76 (0.32)
0.68 (0.33)

0.21 (0.26)
0.36 (0.32)
0.75 (0.32)
0.66 (0.32)

Note: Source: 2005 program director survey. Values were calculated by taking the midpoint of the response
categories (0%, 1%–25%, 26%–50%, 51%–75%, 76%–99%, and 100%) reported by the SVORI juvenile program
directors for each of the services. AA = Alcoholics Anonymous, GED = general educational development, NA =
Narcotics Anonymous.

C-7

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit C-5a. Involvement and contributions of agencies and community-based
organizations (CBOs) to SVORI programs in adult sites, by site type (as reported by
program directors)

Non–impact
Impact Sites
Sites
SVORI Agency Involvement and
Contributions in Adult Sites
%
N
%
N
Percentage of PDs in adult sites reporting agency involvement with SVORI
Pre-release supervision agency
100.0%
12
94.9%
37
Post-release supervision agency
100.0%
12
100.0%
39
Faith-based organizations
100.0%
12
82.1%
32
Substance abuse agencies or CBOs
91.7%
11
97.4%
38
Mental health agency or CBOs
100.0%
12
97.4%
38
91.7%
11
82.1%
32
Family/social service agencies or
CBOs
Law enforcement agency
75.0%
9
76.9%
30
Housing agencies or CBOs
83.3%
10
79.5%
31
Employment agencies or CBOs
100.0%
12
94.9%
37
Vocational training agencies or CBOs
100.0%
12
87.2%
34
91.7%
11
66.7%
26
Technical institutions, community
colleges or universities
Local school systems
8.3%
1
15.4%
6
Percentage of PDs in adult sites reporting agency contributions to SVORI
Pre-release supervision agency
91.7%
11
69.2%
27
Post-release supervision agency
75.0%
9
76.9%
30
Faith-based organizations
66.7%
8
38.5%
15
Substance abuse agencies or CBOs
75.0%
9
66.7%
26
Mental health agency or CBOs
66.7%
8
74.4%
29
75.0%
9
46.2%
18
Family/social service agencies or
CBOs
Law enforcement agency
25.0%
3
35.9%
14
Housing agencies or CBOs
58.3%
7
48.7%
19
Employment agencies or CBOs
75.0%
9
64.1%
25
Vocational training agencies or CBOs
66.7%
8
59.0%
23
58.3%
7
30.8%
12
Technical institutions, community
colleges or universities
Local school systems
8.3%
1
10.3%
4

All Sites
%
N
96.1%
100.0%
86.3%
96.1%
98.0%
84.3%

49
51
44
49
50
43

76.5%
80.4%
96.1%
90.2%
72.5%

39
41
49
46
37

13.7%

74.5%
76.5%
45.1%
68.6%
72.5%
52.9%

38
39
23
35
37
27

33.3%
51.0%
66.7%
60.8%
37.3%

17
26
34
31
19

9.8%

Source: 2006 program director survey, 51 programs reporting: 12 impact and 39 non–impact programs.

C-8

Appendix C — Comparisons of Impact and Non–impact Sites

Exhibit C-5b. Involvement and contributions of agencies and community-based
organizations (CBOs) to SVORI programs in juvenile sites, by site type (as reported by
program directors)

Non–impact
Impact Sites
Sites
SVORI Agency Involvement and
Contributions in Juvenile Sites
%
N
%
N
Percentage of PDs in juvenile sites reporting agency involvement with SVORI
Pre-release supervision agency
75.0%
3
93.5%
29
Post-release supervision agency
100.0%
4
93.5%
29
Faith-based organizations
100.0%
4
87.1%
27
Substance abuse agencies or CBOs
50.0%
2
93.5%
29
Mental health agency or CBOs
75.0%
3
96.8%
30
Family/social service agencies or CBOs
100.0%
4
90.3%
28
Law enforcement agency
75.0%
3
58.1%
18
Housing agencies or CBOs
75.0%
3
61.3%
19
Employment agencies or CBOs
100.0%
4
80.6%
25
Vocational training agencies or CBOs
100.0%
4
58.1%
18
4
54.8%
17
Technical institutions, community colleges 100.0%
or universities
Local school systems
100.0%
4
90.3%
28
Percentage of PDs in juvenile sites reporting agency contributions to SVORI
Pre-release supervision agency
75.0%
3
71.0%
22
Post-release supervision agency
75.0%
3
77.4%
24
Faith-based organizations
75.0%
3
48.4%
15
Substance abuse agencies or CBOs
25.0%
1
67.7%
21
Mental health agency or CBOs
25.0%
1
67.7%
21
Family/social service agencies or CBOs
75.0%
3
58.1%
18
Law enforcement agency
50.0%
2
32.3%
10
Housing agencies or CBOs
25.0%
1
35.5%
11
Employment agencies or CBOs
50.0%
2
61.3%
19
Vocational training agencies or CBOs
50.0%
2
38.7%
12
75.0%
3
45.2%
14
Technical institutions, community colleges
or universities
Local school systems
50.0%
2
64.5%
20

All Sites
%
N
91.4%
94.3%
88.6%
88.6%
94.3%
91.4%
60.0%
62.9%
82.9%
62.9%
60.0%

32
33
31
31
33
32
21
22
29
22
21

91.4%

71.4%
77.1%
51.4%
62.9%
62.9%
60.0%
34.3%
34.3%
60.0%
40.0%
48.6%

25
27
18
22
22
21
12
12
21
14
17

62.9%

Source: 2006 program director survey, 35 programs reporting: 4 impact and 31 non–impact programs.

C-9

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit C-6a. Support and resistance by individual stakeholders to SVORI programs in adult
sites, by site type (as reported by program directors)

Non–impact
Impact
Sites
Sites
All Sites
SVORI Support and Resistance in
Adult Sites
%
N
%
N
%
N
Percentage of PDs in adult sites reporting support for SVORI from the following individuals
83.3%
10
69.2%
27
72.5%
37
Top administrators at the pre-release
facilities
Supervisors at the pre-release facilities
66.7%
8
69.2%
27
68.6%
35
Line staff at the pre-release facilities
58.3%
7
51.3%
20
52.9%
27
83.3%
10
76.9%
30
78.4%
40
Top administrators at the post-release
supervision agency
83.3%
10
71.8%
28
74.5%
38
Supervisors at the post-release
supervision agency
66.7%
8
71.8%
28
70.6%
36
Line staff at the post-release
supervision agency
Members of the community
41.7%
5
38.5%
15
39.2%
20
Percentage of PDs in adult sites reporting resistance to SVORI from the following individuals
16.7%
2
5.1%
2
7.8%
4
Top administrators at the pre-release
facilities
Supervisors at the pre-release facilities
25.0%
3
12.8%
5
15.7%
8
Line staff at the pre-release facilities
33.3%
4
17.9%
7
21.6%
11
8.3%
1
10.3%
4
9.8%
5
Top administrators at the post-release
supervision agency
8.3%
1
12.8%
5
11.8%
6
Supervisors at the post-release
supervision agency
33.3%
4
10.3%
4
15.7%
8
Line staff at the post-release
supervision agency
Members of the community
33.3%
4
48.7%
19
45.1%
23
Source: 2006 program director survey, 51 programs reporting: 12 impact and 39 non–impact programs.

C-10

Appendix C — Comparisons of Impact and Non–impact Sites

Exhibit C-6b. Support and resistance by individual stakeholders to SVORI programs in
juvenile sites, by site type (as reported by program directors)

Non–impact
Impact
Sites
Sites
All Sites
SVORI Support and Resistance in
Juvenile Sites
%
N
%
N
%
N
Percentage of PDs in juvenile sites reporting support for SVORI from the following individuals
75.0%
3
71.0%
22
71.4%
25
Top administrators at the pre-release
facilities
Supervisors at the pre-release facilities
75.0%
3
74.2%
23
74.3%
26
Line staff at the pre-release facilities
75.0%
3
51.6%
16
54.3%
19
75.0%
3
71.0%
22
71.4%
25
Top administrators at the post-release
supervision agency
75.0%
3
67.7%
21
68.6%
24
Supervisors at the post-release
supervision agency
25.0%
1
64.5%
20
60.0%
21
Line staff at the post-release
supervision agency
Members of the community
50.0%
2
48.4%
15
48.6%
17
Percentage of PDs in juvenile sites reporting resistance to SVORI from the following individuals
25.0%
1
9.7%
3
11.4%
4
Top administrators at the pre-release
facilities
Supervisors at the pre-release facilities
25.0%
1
29.0%
9
28.6%
10
Line staff at the pre-release facilities
25.0%
1
9.7%
3
11.4%
4
0.0%
0
6.5%
2
5.7%
2
Top administrators at the post-release
supervision agency
0.0%
0
16.1%
5
14.3%
5
Supervisors at the post-release
supervision agency
50.0%
2
35.5%
11
37.1%
13
Line staff at the post-release
supervision agency
Members of the community
25.0%
1
9.7%
3
11.4%
4
Source: 2006 program director survey, 35 programs reporting: 4 impact and 31 non–impact programs.

C-11

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit C-7a. Geographic targeting among adult impact and non–impact sitesa

Adult Programs
Pre-release geographic targeting
One facility
Multiple facilities
All facilities
Post-release geographic targeting
One community
Multiple communities
All communities (statewide)
a

Impact Sites
%
N

Non–impact
Sites
%
N

—
50.0%
50.0%

0
6
6

10.0%
42.5%
47.5%

4
17
19

7.7%
44.2%
48.1%

4
23
25

50.0%
41.7%
8.3%

6
5
1

45.0%
45.0%
10.0%

18
18
4

46.2%
44.2%
9.6%

24
23
5

All Sites
%
N

Source: 2003 program workplan review of 52 programs: 12 impact and 40 non–impact sites.

Exhibit C-7b. Geographic targeting among juvenile impact and non–impact sitesa

Juvenile Programs
Pre-release geographic targeting
One facility
Multiple facilities
All facilities
Post-release geographic targeting
One community
Multiple communities
All communities (statewide)
a

Impact Sites
%
N

Non–impact
Sites
%
N

—
50.0%
50.0%

0
2
2

9.1%
51.5%
39.4%

3
17
13

8.1%
51.4%
40.5%

3
19
15

—
100.0%
—

0
4
0

42.4%
48.5%
9.1%

14
16
3

37.8%
54.1%
8.1%

14
20
3

Source: 2003 program workplan review of 37 programs: 4 impact and 33 non–impact sites.

C-12

All Sites
%
N

Appendix C — Comparisons of Impact and Non–impact Sites

Exhibit C-8a. Enhancements to pre-release services in adult sites, by site type (as reported
by program directors)
Percentage of PDs in adult sites reporting newly
implemented services or enhancements to pre-release
services through SVORI

Case management
Risk assessment
Needs assessment
Treatment/release plan development
In-person contact from post-release case manager
while offender is still incarcerated
Reentry courts
Video-conferencing
Offender-specific reentry teams
AA/NA
Counseling sessions for drug or alcohol use
Comprehensive drug treatment programs
Mental health services
Anger management/violence counseling
Education/GED/tutoring/literacy
Employment referrals/job placement
Resume and interviewing skills development
Cognitive skills development/behavioral
programming
Life skills training
Pre-release curriculum
Assistance obtaining identification (e.g., driver’s
license, Social Security card)
Assistance obtaining benefits and completing
applications (e.g., Medicaid, disability)
Financial support/emergency assistance
Parenting skills development
Family reunification
Peer support groups
One-on-one mentoring
Housing placements or referrals

Impact Sites
%
N
83.3% 10
50.0%
6
66.7%
8
75.0%
9
91.7% 11

Non–Impact
Sites
%
N
69.2% 27
51.3% 20
69.2% 27
69.2% 27
71.8% 28

All Sites
%
N
72.5% 37
51.0% 26
68.6% 35
70.6% 36
76.5% 39

0.0%
33.3%
75.0%
0.0%
33.3%
0.0%
25.0%
25.0%
25.0%
66.7%
66.7%
58.3%

0
4
9
0
4
0
3
3
3
8
8
7

10.3%
20.5%
66.7%
5.1%
20.5%
7.7%
28.2%
25.6%
23.1%
53.8%
51.3%
33.3%

4
8
26
2
8
3
11
10
9
21
20
13

7.8%
23.5%
68.6%
3.9%
23.5%
5.9%
27.5%
25.5%
23.5%
56.9%
54.9%
39.2%

4
12
35
2
12
3
14
13
12
29
28
20

58.3%
75.0%
75.0%

7
9
9

43.6%
51.3%
53.8%

17
20
21

47.1%
56.9%
58.8%

24
29
30

83.3%

48.7%

56.9%

50.0%
41.7%
50.0%
16.7%
108.3%
75.0%

6
5
6
2
13
9

35.9%
15.4%
28.2%
20.5%
10.3%
51.3%

14
6
11
8
4
20

39.2%
21.6%
33.3%
19.6%
33.3%
56.9%

20
11
17
10
17
29

Note: Source: 2006 program director survey, 51 programs reporting: 12 impact, 39 non–impact sites. AA = Alcoholics
Anonymous, GED = general educational development, NA = Narcotics Anonymous.

C-13

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit C-8b. Enhancements to post-release services in adult sites, by site type (as reported
by program directors)
Percentage of PDs in adult sites reporting newly
implemented services or enhancements to postrelease services through SVORI

Case management
Risk assessment
Needs assessment
Treatment/release plan development
Formal post-release supervision
Reentry courts
Video-conferencing
Offender-specific reentry teams
AA/NA
Counseling sessions for drug or alcohol use
Comprehensive drug treatment programs
Mental health services
Anger management/violence counseling
Education/GED/tutoring/literacy
Employment referrals/job placement
Resume and interviewing skills development
Cognitive skills development/behavioral
programming
Life skills training
Assistance obtaining identification (e.g., driver’s
license, Social Security card)
Assistance obtaining benefits and completing
applications (e.g., Medicaid, disability)
Financial support/emergency assistance
Parenting skills development
Family reunification
Peer support groups
One-on-one mentoring
Housing placements or referrals
Transportation

Impact Sites
%
N
100.0% 12
58.3%
7
66.7%
8
66.7%
8
91.7% 11
25.0%
3
50.0%
6
83.3% 10
100.0% 12
66.7%
8
100.0% 12
83.3% 10
25.0%
3
66.7%
8
100.0% 12
83.3% 10
58.3%
7

Non–Impact
Sites
%
N
64.1% 25
51.3% 20
59.0% 23
64.1% 25
53.8% 21
12.8%
5
5.1%
2
69.2% 27
82.1% 32
41.0% 16
87.2% 34
51.3% 20
28.2% 11
38.5% 15
71.8% 28
66.7% 26
28.2% 11

All Sites
%
N
72.5% 37
52.9% 27
60.8% 31
64.7% 33
62.7% 32
15.7%
8
15.7%
8
72.5% 37
86.3% 44
47.1% 24
90.2% 46
58.8% 30
27.5% 14
45.1% 23
78.4% 40
70.6% 36
35.3% 18

66.7%
91.7%

8
11

35.9%
53.8%

14
21

43.1%
62.7%

22
32

100.0%

59.0%

68.6%

83.3%
41.7%
75.0%
41.7%
75.0%
100.0%
66.7%

10
5
9
5
9
12
8

61.5%
15.4%
33.3%
33.3%
41.0%
59.0%
51.3%

24
6
13
13
16
23
20

66.7%
21.6%
43.1%
35.3%
49.0%
68.6%
54.9%

34
11
22
18
25
35
28

Note: Source: 2006 program director survey, 51 programs reporting: 12 impact, 39 non–impact sites. AA = Alcoholics
Anonymous, GED = general educational development, NA = Narcotics Anonymous,.

C-14

Appendix C — Comparisons of Impact and Non–impact Sites

Exhibit C-8c. Enhancements to pre-release services in juvenile sites, by site type (as
reported by program directors)
Percentage of PDs in juvenile sites reporting newly
implemented services or enhancements to prerelease services through SVORI

Case management
Risk assessment
Needs assessment
Treatment/release plan development
In-person contact from post-release case
manager while offender is still incarcerated
Reentry courts
Video-conferencing
Offender-specific reentry teams
AA/NA
Counseling sessions for drug or alcohol use
Comprehensive drug treatment programs
Mental health services
Anger management/violence counseling
Education/GED/tutoring/literacy
Employment referrals/job placement
Resume and interviewing skills development
Cognitive skills development/behavioral
programming
Life skills training
Pre-release curriculum
Assistance obtaining identification (e.g., driver’s
license, Social Security card)
Assistance obtaining benefits and completing
applications (e.g., Medicaid, disability)
Financial support/emergency assistance
Parenting skills development
Family reunification
Peer support groups
One-on-one mentoring
Housing placements or referrals

Impact Sites
%
N
100.0%
4
50.0%
2
75.0%
3
100.0%
4
75.0%
3

Non–Impact
Sites
%
N
61.3% 19
48.4% 15
54.8% 17
71.0% 22
64.5% 20

0.0%
25.0%
75.0%
0.0%
25.0%
100.0%
0.0%
100.0%
75.0%
75.0%
75.0%
75.0%

0
1
3
0
1
4
0
4
3
3
3
3

16.1%
29.0%
61.3%
22.6%
38.7%
74.2%
32.3%
100.0%
35.5%
41.9%
35.5%
41.9%

5
9
19
7
12
23
10
31
11
13
11
13

14.3%
28.6%
62.9%
20.0%
37.1%
77.1%
28.6%
100.0%
40.0%
45.7%
40.0%
45.7%

5
10
22
7
13
27
10
35
14
16
14
16

75.0%
100.0%
50.0%

3
4
2

51.6%
32.3%
96.8%

16
10
30

54.3%
40.0%
91.4%

19
14
32

50.0%

25.8%

28.6%

75.0%
100.0%
50.0%
0.0%
75.0%
50.0%

3
4
2
0
3
2

35.5%
83.9%
38.7%
22.6%
45.2%
19.4%

11
26
12
7
14
6

40.0%
85.7%
40.0%
20.0%
48.6%
22.9%

14
30
14
7
17
8

All Sites
%
N
65.7% 23
48.6% 17
57.1% 20
74.3% 26
65.7% 23

Note: Source: 2006 program director survey, 35 programs reporting: 4 impact, 31 non–impact sites. AA = Alcoholics
Anonymous, GED = general educational development, NA = Narcotics Anonymous.

C-15

The Multi-site Evaluation of SVORI: Methodology & Analytic Approach

Exhibit C-8d. Enhancements to post-release services in juvenile sites, by site type (as
reported by program directors)
Percentage of PDs in juvenile sites reporting newly
implemented services or enhancements to postrelease services through SVORI

Impact Sites
%
N
100.0%
4
75.0%
3
100.0%
4
75.0%
3
75.0%
3
0.0%
0
25.0%
1
100.0%
4
25.0%
1
50.0%
2
50.0%
2
75.0%
3
100.0%
4
100.0%
4
100.0%
4
75.0%
3
75.0%
3

Non–Impact
Sites
%
N
74.2% 23
45.2% 14
45.2% 14
74.2% 23
67.7% 21
19.4% 6
22.6% 7
67.7% 21
29.0% 9
48.4% 15
29.0% 9
41.9% 13
100.0% 31
41.9% 13
48.4% 15
54.8% 17
48.4% 15

All Sites
%
N
77.1% 27
48.6% 17
51.4% 18
74.3% 26
68.6% 24
17.1% 6
22.9% 8
71.4% 25
28.6% 10
48.6% 17
31.4% 11
45.7% 16
100.0% 35
48.6% 17
54.3% 19
57.1% 20
51.4% 18

75.0%
75.0%

3
3

58.1%
51.6%

18
16

60.0%
54.3%

21
19

50.0%

41.9%

42.9%

100.0%
100.0%
100.0%
50.0%
100.0%
75.0%
100.0%

4
4
4
2
4
3
4

93.5%
83.9%
48.4%
38.7%
48.4%
38.7%
61.3%

29
26
15
12
15
12
19

94.3%
85.7%
54.3%
40.0%
54.3%
42.9%
65.7%

33
30
19
14
19
15
23

Note: Source: 2006 program director survey. 35 programs reporting: 4 impact, 31 non–impact sites. AA = Alcoholics
Anonymous, GED = general educational development, NA = Narcotics Anonymous.

C-16

Appendix D. MOU Template

D-1

Memorandum of Understanding between [Facility/Agency] and RTI International
(RTI) and the Urban Institute (UI) Regarding Data Collection for the Multi-Site
Evaluation of the Serious and Violent Offender Reentry Initiative
The multi-site evaluation of the Serious and Violent Offender Reentry Initiative (SVORI) is a
research project that will help determine whether federal SVORI funding received by local and
state correctional agencies is having an impact on public safety and the social functioning of
returning offenders in [STATE].
The protocols of this study have been approved by [STATE CORRECTIONS AGENCY] and
RTI’s Institutional Review Board. By signing this Memorandum of Understanding, we agree to
ensure that [STATE CORRECTIONS AGENCY], RTI, and UI staff follow the data collection,
confidentiality, voluntary participation, and data security protocols and guidelines outlined below
and those included in the attached Data Security Plan and Privacy Certification, and Staff
Confidentiality Pledge.
Facility Staff
[AGENCY/FACILITY] staff will assist RTI and UI research staff by providing the data and
information necessary to identify, access, and track eligible study participants. Such information
includes but is not limited to current lists of soon-to-be-released offenders who meet the
sampling requirements of the study, parental/guardian addresses for soon-to-be-released juvenile
offenders, updated release dates of study participants, contact information, and other information
as necessary that will assist in the post-release locating of study participants.
By signing this agreement, [STATE CORRECTIONS AGENCY] staff further acknowledge that
offender participation in the pre-release interviews must be voluntary and that offenders who
choose not to participate in the study will not be subjected to any adverse punishment or
consequences as a result of that decision.
Trained Field Interviewers
Pre-release interviews will be conducted by trained field interviewers from RTI and UI, or their
subcontractors and consultants. RTI and UI certify that field interviewers will comply with
[STATE CORRECTIONS AGENCY] and [FACILITY] rules and regulations regarding entering
the facility and conducting interviews with prisoners in the facility.
For security clearance purposes, RTI and UI agree to provide the appropriate facility staff person
with the names of field interviewers who will be entering the facility to conduct interviews, at
least 48 hours in advance.
Field interviewers will be trained to recognize signs of serious emotional distress. If a study
participant becomes distressed during the course of an interview, the field interviewer will notify
the appropriate facility staff person, as designated by the facility director.
Pre-release Interviews

Page 1 of 7

The inmates to be included in the study sample will be selected and pre-release interviews
conducted from [MONTH 2004] through [MONTH 2005]*.
This agreement authorizes trained research staff from RTI and UI, or their subcontractors and
consultants, to interview soon-to-be-released offenders who consent to participate in the study.
Interviews will be conducted using laptop computers, which field interviewers will be permitted
to bring into the facility. RTI staff will work with facility leadership to ensure that facility
security is not breached by the possession and use of the project laptops.
The interviews must be conducted in a private setting in [FACILITY] that ensures the
confidentiality of the information discussed by the interviewers and the study participants but for
safety reasons, still permits visual observation by facility staff.
*The end dates may be extended for up to 3 months, if necessary to achieve optimal sample size.
Participant Confidentiality & Data Security
Participation in this study is voluntary and the information collected is confidential. Facility staff
will not attempt to determine who agrees to participate and who does not. Facility staff will not
have access to any information provided in the interviews, with the exception of a study
participant expressing his or her intention to escape or seriously hurt him/herself or someone
else. If a study participant discloses such information to an interviewer, the interviewer will
notify the appropriate facility staff person, as designated by the facility director.
As described in the attached Data Security Plan, subjects’ names will appear only on a locator
form identifying the study as a national reentry study, and subjects will be identified in the
laptop’s case management system only by a case ID number. Access to both the individual
laptops and the case management system on the laptops will be password protected at all levels.
Reports and other products generated from this study will present aggregate data and individual
facilities will not be identified.
RTI and UI agree to use the data only for research purposes and to maintain the confidentiality of
the data. All information will be kept strictly confidential and is protected under Federal law by
28 CFR Part 22 -Confidentiality of Identifiable Research and Statistical Information, which
regulates the disclosure of identifiable information and requires that RTI protect the
confidentiality information provided. RTI, UI, and subcontracted project staff will sign the
attached Certificate of Confidentiality. All parties agree that electronic and faxed signatures are
both acceptable forms of signature.
_________________________
______________________________
NAME
Pamela K. Lattimore, Ph.D.
Agency Director/Facility Director
Co-Principal Investigator
Agency/Facility
Research Triangle Institute
_______________________________
Christy Visher, Ph.D.
Co-Principal Investigator
The Urban Institute

Page 2 of 7

RESEARCH TRIANGLE INSTITUTE & THE URBAN INSTITUTE’S
DATA SECURITY PLAN

While interviews conducted in correctional facilities will take place within the line of sight of
prison staff (for safety reasons), they will be conducted at enough distance that staff cannot
overhear or see the interview and with the subject facing away from staff so that what is being
said cannot be discerned. In addition, those electing not to participate will spend enough time
with the interviewer to ensure that staff cannot determine who did or did not participate.
To protect confidentiality, locator information will be gathered early in the interview, and
automatically secured in a separate file so that it cannot be accessed from the laptop computer
during the remainder of the interview nor linked with information contained on the laptop.
Likewise, near the end of the interviewing process, after all interview response data is collected
but prior to finalizing the case, all data collected thus far will again be automatically secured and
made inaccessible from the laptop. Subjects’ names will appear only on a locator form
identifying the study as a national reentry study, and subjects will be identified in the laptop’s
case management system solely by a case ID number. Access to both the individual laptops and
the case management system on the laptops will be password protected at each level.
As an added precaution, all interviewers will be required to sign a pledge of confidentiality and
the interviewer training will strongly emphasize the privacy and confidentiality aspects of the
study.
We will treat all data as confidential and it will be released to the public only as aggregate data in
statistical tables that protect respondents’ identities.

Page 3 of 7

CERTIFICATION OF PRIVACY
Multi-Site Evaluation of the Serious and Violent Offender Reentry Initiative

Pamela K. Lattimore and Christy Visher, as Co-Principal Investigators (hereinafter
collectively referred to as Grantees) certify on behalf of Research Triangle Institute and the
Urban Institute, that data identifiable to a private person1 will not be used or revealed, except as
authorized in 28 CFR Part 22, Sections 22.21 & 22.22.
Grantee certifies that access to the data will be limited to those employees having a need for such
data and that such employees shall be advised of and agree in writing to comply with the
regulations in 28 CFR Part 22.
Grantee certifies that all contractors, subcontractors, and consultants requiring access to
identifiable data will agree, through conditions in their subcontract or consultant agreement, to
comply with the requirements of 28 CFR §22.24, regarding information transfer agreements.
Grantee also certifies that NIJ will be provided with copies of any and all transfer agreements
before they are executed as well as the name and title of the individual(s) with the authority to
transfer data.
Grantee certifies that any private person from whom identifiable information is collected or
obtained shall be notified, in accordance with 28 CFR §22.27, that such data will only be used or
revealed for research or statistical purposes and that compliance with the request for information
is not mandatory and participation in the project maybe terminated at any time. In addition,
grantee certifies that where findings in a project cannot, by virtue of sample size or uniqueness of
subject, be expected to totally conceal the identity of an individual, such individual shall be so
advised.
Grantee certifies that project plans will be designed to preserve the confidentiality of private
persons to whom information relates, including where appropriate, name-stripping, coding of
data, or other similar procedures.
___________________________________
1
Information identifiable to a private person is defined in 28 CFR §22.2(e) as “information
which either--(1) Is labeled by name or other personal identifiers, or (2) Can, by virtue of sample
size or other factors, be reasonably interpreted as referring to a particular person.”

Page 4 of 7

Grantee certifies that copies of all questionnaires that have already been designed for use in the
project are attached to this Privacy Certificate. Grantee also certifies that any questionnaires
developed during the project period will be provided to NIJ at the end of the project.
Grantee certifies that project findings and reports prepared for dissemination will not contain
information which can reasonably be expected to be identifiable to a private person, except as
authorized by 28 CFR §22.22.
Grantee certifies that adequate precautions will be taken to ensure administrative and physical
security of identifiable data and to preserve the confidentiality of the personally identifiable
information.
Grantee certifies that all project personnel, including subcontractors, have been advised of and
have agreed, in writing, to comply with all procedures to protect privacy and the confidentiality
of personally identifiable information.

Page 5 of 7

STAFF CONFIDENTIALITY PLEDGE
Pursuant to Title 28 of the Code of Federal Regulations, Part 22, project staff have an
obligation to those we interview to protect their identities and the information they
provide to the Multi-Site Evaluation of the Serious and Violent Offender Reentry
Initiative project. The identity of persons interviewed and the related data are to remain
confidential. Removal of names or disclosure of identities and related information is
strictly forbidden. Contents of interviews are not to be discussed with anyone except
project staff, and only as it is necessary to complete the assigned work. Additionally,
sensitive interview information should not be discussed anywhere it could be overheard
by persons who are not authorized to know this information. There are exceptions to the
mandate of confidentiality. If I reasonably suspect that a participant is planning an
escape, in immediate danger, intends to harm someone else, or is a victim or
perpetrator of child abuse, I may be required to inform the appropriate authorities
according to project protocol and in compliance with state and local law.
As a member of the site personnel, contractor, or subcontractor staff of the Multi-Site
Evaluation of the Serious and Violent Offender Reentry Initiative project, I,
_____________________________, agree that I will protect the confidentiality of all
information identifiable to a private person that is collected in the conduct of my work for
the project.
I agree that I shall not discuss any identifiable information that I may learn of during the
course of my employment as part of the site personnel, contractor, or subcontractor
staff with anyone other than project staff members who have a need-to-know this
information.
I agree to follow the procedures established by the project to prevent unauthorized
access to information identifiable to a private person.
I certify that I have been informed that, the Multi-Site Evaluation of the Serious and
Violent Offender Reentry Initiative project, which is being funded in whole or in part
by the National Institute of Justice, is governed by the Department of Justice
Regulations in 28 CFR Part 22 & Part 46, which govern the use and revelation of
research and statistical information identifiable to a private person, and that I, as a
member of the project's site personnel, contractor, or subcontractor staff am governed
by these regulations as well.
I certify that I have been given copies of the regulations at 28 CFR Part 22 & Part 46
and that I understand the obligations imposed by them.
I understand that my signing this agreement is a condition of my employment as part of
the Multi-Site Evaluation of the Serious and Violent Offender Reentry Initiative
project's site personnel, contractor, or subcontractor staff.

Page 6 of 7

By signing this statement, I am acknowledging that I understand the rules surrounding
the protection of confidential information and, if I am found to be in violation of these
provisions, I can be fined not to exceed $10,000 in addition to any other penalty
imposed by law.

Full Legal Name (please print): ___________________________________

_________________________________
Signature

______________________
Date

Page 7 of 7

Appendix E. Consent Forms
and Interview Materials

E-1

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR ADULT PARTICIPANTS
Baseline Interview
National Re-Entry Study
About the Study
The National Re-Entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We will be interviewing approximately 4,000 individuals selected from
15 states nationwide. You are invited to participate in the study.
Participation in the Study
If you decide to be in the study, we will ask you to complete an interview. This interview will take about 1 and ½ hours. The
interviewer will read questions from a computer screen and type your answers into the computer. The interviewer will ask you
questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other illegal
behavior, and services or treatment you have or may receive. Once you have been released, we would like to get additional data about
your period of criminal justice supervision and employment by contacting the agency in charge of your supervision and state or
national agencies responsible for keeping this data. We have a separate form we will ask you to sign to give your permission. This
form will have more detail about who we would like to contact and what we will ask them. You can decline to allow us to obtain this
data and still participate in the interview.
Voluntary Participation
The choice of whether to participate in this study is completely up to you. No one will be upset or angry if you decide not to
participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse to
answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go on
to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. In order to avoid that, we
will do the interview in a private setting where no one can overhear the answers. Also, we will replace your name with an ID number
to identify your interview in the computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, correctional facility staff, will be able to find out what you said in the interview. We will not use
your name in connection with any of your answers to the interview questions. Instead, we will enter an ID number into the computer.
This ID number replaces your name and only we will be able to link it to you. Once the study is over, we will delete the answers you
gave us and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are planning an escape, that you are in immediate
danger, or that you intend to harm someone else, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will contact you again for three more interviews. We would like you to participate in these interviews at three, nine, and fifteen
months from when you are released. Each time we do one of these interviews, you will have another chance to read a consent form
like this and decide if you want to do the interview. For those interviews conducted outside of a correctional facility, you will receive
a monetary payment of $35 for the three month interview, and $50 each for the nine and fifteen month interviews. You will also
receive an additional $5 each time you call us to schedule an appointment for one of these follow-up interviews. Finally, if you
participate in all 4 interviews, you will receive an additional $50.
Original (white) to RTI, Copy to Respondent

Page 2 of 2
Case ID Number: __ __ __ __ __ __ __ __
Questions
If you have any questions about the study, you may call 1-877-475-7056 (toll free) and leave a message so that someone from the
project can contact you; or you can write to us at National Re-Entry Study, 3040 Cornwallis Road, PO Box 12194, Research Triangle
Park, North Carolina, 27709-2194. If you have any questions about your rights as a research study participant, you may call the RTI
International Office of Human Subject Protections toll-free at 1-866-214-2043; or you can write to them at RTI International Office of
Human Subject Protections, 3040 Cornwallis Road, PO Box 12194, Research Triangle Park, North Carolina, 27709-2194.
Do you have any questions that might help you decide if you do or do not want to participate in the study?
By checking the box below, you are agreeing to participate in the study. Please check the box only if:
9
9
9

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to participate in the study.

□Check here if the above statements are true.
Interviewer’s Signature

Date

Original (white) to RTI, Copy to Respondent

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR YOUTH PARTICIPANTS
Baseline Interview
National Re-Entry Study
About the Study
The National Re-entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We will be interviewing approximately 4,000 individuals selected from
15 states nationwide. You are invited to participate in the study.
Participation in the Study
If you decide to be in the study, we will ask you to complete an interview. This interview will take about 1 and ½ hours. The
interviewer will read questions from a computer screen and type your answers into the computer. The interviewer will ask you
questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other illegal
behavior, and services or treatment you have or may receive. Once you have been released to the community, we would like to get
additional data about your period of criminal justice supervision and employment by contacting the agency in charge of your
supervision and state or national agencies responsible for keeping this data. We have a separate form we will ask you to sign to give
your permission. This form will have more detail about who we would like to contact and what we will ask them. You can decline to
allow us to obtain this data and still participate in the interview.
Voluntary Participation
The choice of whether to participate in this study is completely up to you. No one will be upset or angry if you decide not to
participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse to
answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go on
to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. For example, if you tell us
that you used drugs or committed a crime, someone might overhear this. In order to avoid that, we will do the interview in a private
setting where no one can overhear the answers. Also, we will replace your name with an ID number to identify your interview in the
computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, your parents or correctional facility staff, will be able to find out what you said in the interview.
We will not ask your parents’ permission for you to participate; however we have sent them a letter to let them know that you have
been invited to be in the study. We will not use your name in connection with any of your answers to the interview questions.
Instead, we will enter an ID number into the computer. This ID number replaces your name and only we will be able to link it to you.
Once the study is over, we will delete the answers you gave us and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are planning an escape, that you are in immediate
danger, that you intend to harm someone else, or that you are being mistreated, we may need to inform the appropriate authorities
according to state and local law.
Future Contacts
We will contact you again for three more interviews. We would like you to participate in these interviews at three, nine, and fifteen
months from when you are released. Each time we do one of these interviews, you will have another chance to read a consent form
like this and decide if you want to do the interview. For those interviews conducted outside of a correctional facility, you will receive
a monetary payment of $35 for the three months interview, and $50 each for the nine and fifteen month interviews. You will also
Original (white) to RTI, Copy to Respondent

Page 2 of 2
Case ID Number: __ __ __ __ __ __ __ __
receive an additional $5 each time you call us to schedule an appointment for one of these follow-up interviews. Finally, if you
participate in all 4 interviews, you will receive an additional $50.
Questions
If you have any questions about the study, you may call 1-877-475-7056 (toll free) and leave a message so that someone from the
project can contact you; or you can write to us at National Re-Entry Study, 3040 Cornwallis Road, PO Box 12194, Research Triangle
Park, North Carolina, 27709-2194. If you have any questions about your rights as a research study participant, you may call the RTI
International Office of Human Subject Protections toll-free at 1-866-214-2043; or you can write to them at RTI International Office of
Human Subject Protections, 3040 Cornwallis Road, PO Box 12194, Research Triangle Park, North Carolina, 27709-2194.
Do you have any questions that might help you decide if you do or do not want to participate in the study?
By checking the box below, you are agreeing to participate in the study. Please check the box only if:
9
9
9

□

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to participate in the study.

Check here if the above statements are true.

Interviewer’s Signature

Date

Original (white) to RTI, Copy to Respondent

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR ADULT PARTICIPANTS
Three-Month Follow Up Interview - Community
National Re-Entry Study
About the Study
The National Re-Entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, you
completed a computer interview with one of our interviewers. If you now agree to continue to be in the study, we will ask you to
complete a follow up interview. This interview will take about 1 and ½ hours. Like last time, the interviewer will read questions from
a computer screen and type your answers into the computer. The interviewer will ask you questions about your housing situation,
mental and physical health, family and friends, job, alcohol and drug use, other illegal behavior, and services or treatment you have or
may receive. In an effort to thank you for taking the time for the interview, we will give you $35.00 upon completion of the interview.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. In order to avoid that, we
will do the interview in a private setting where no one can overhear the answers. Also, we will replace your name with an ID number
instead of a name to identify your interview in the computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else will be able to find out what you said in the interview. We will not use your name in connection with any of your answers
to the interview questions. Instead, we will enter an ID number into the computer. This ID number replaces your name and only we
will be able to link it to you. Once the study is over, we will delete the answers you gave us and the list that links your name and ID
number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are in immediate danger or that you intend to harm
someone else, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will contact you again for two more interviews. We would like you to participate in these interviews at nine and fifteen months
from when you were released. You will have another chance to read a consent form like this and decide if you want to do each of the
interviews. For interviews conducted outside of a correctional facility, you will receive a monetary payment of $50 each for the nine
and fifteen month interviews. You will also receive an additional $5 each time you call to schedule an appointment for one of the
follow-up interviews. Finally, if you participate in all 4 interviews, you will receive an additional $50.

Original (white) to RTI, Copy to Respondent

Three-month Follow-up Interview – Community ADULT

Page 2 of 2

Case ID Number: __ __ __ __ __ __ __ __
Questions
If you have any questions about the study, you may call 1-877-475-7056 (toll free) and leave a message so that someone from the
project can contact you; or you can write to us at National Re-Entry Study, 3040 Cornwallis Road, PO Box 12194, Research Triangle
Park, North Carolina, 27709-2194. If you have any questions about your rights as a research study participant, you may call the RTI
International Office of Human Subject Protections toll-free at 1-866-214-2043; or you can write to them at RTI International Office of
Human Subject Protections, 3040 Cornwallis Road, PO Box 12194, Research Triangle Park, North Carolina, 27709-2194.
Do you have any questions that might help you decide if you do or do not want to continue to participate in the study?
By checking the box below, you are agreeing to continue to participate in the study. Please check the box only if:
9
9
9

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

□Check here if the above statements are true.
Interviewer’s Signature

Date

Original (white) to RTI, Copy to Respondent

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR YOUTH PARTICIPANTS
Three-month Follow Up Interview - Community
National Re-Entry Study
About the Study
The National Re-entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, you
completed a computer interview with one of our interviewers. If you now agree to continue to be in the study, we will ask you to
complete a follow up interview. This interview will take about 1 and ½ hours. Like last time, the interviewer will read questions from
a computer screen and type your answers into the computer. The interviewer will ask you questions about your housing situation,
mental and physical health, family and friends, job, alcohol and drug use, other illegal behavior, and services or treatment you have or
may receive. In an effort to thank you for taking the time for the interview, we will give you $35.00 upon completion of the interview.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. For example, if you tell us
that you used drugs or committed a crime, someone might overhear this. In order to avoid that, we will do the interview in a private
setting where no one can overhear the answers. Also, we will replace your name with an ID number to identify your interview in the
computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, your parents, will be able to find out what you said in the interview. We will not use your name
in connection with any of your answers to the interview questions. Instead, we will enter an ID number into the computer. This ID
number replaces your name and only we will be able to link it to you. Once the study is over, we will delete the answers you gave us
and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are in immediate danger, that you intend to harm
someone else, or that you are being mistreated, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will contact you again for two more interviews. We would like you to participate in these interviews at nine and fifteen months
from when you were released. You will have another chance to read a consent form like this and decide if you want to do each of the
interviews. For interviews conducted outside of a correctional facility, you will receive a monetary payment of $50 each for the nine
and fifteen month interviews. You will also receive an additional $5 each time you call us to schedule an appointment for one of the
follow-up interviews. Finally, if you participate in all 4 interviews, you will receive an additional $50.

Original (white) to RTI, Copy to Respondent

Three-month Follow-up Interview – Community YOUTH

Page 2 of 2

□

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

Check here if the above statements are true.

Interviewer’s Signature

Date

Original (white) to RTI, Copy to Respondent

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR ADULT PARTICIPANTS
Three-Month Follow Up Interview - Facility
National Re-Entry Study
About the Study
The National Re-Entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility about three months ago.
At that time, you completed a computer interview with one of our interviewers. If you now agree to continue to be in the study, we
will ask you to complete a follow up interview. This interview will take about 1 and ½ hours. Like last time, the interviewer will read
questions from a computer screen and type your answers into the computer. The interviewer will ask you questions about your
housing situation, mental and physical health, family and friends, job, alcohol and drug use, other illegal behavior, and services or
treatment you have or may receive.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. In order to avoid that, we
will do the interview in a private setting where no one can overhear the answers. Also, we will replace your name with an ID number
to identify your interview in the computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, correctional facility staff, will be able to find out what you said in the interview. We will not use
your name in connection with any of your answers to the interview questions. Instead, we will enter an ID number into the computer.
This ID number replaces your name and only we will be able to link it to you. Once the study is over, we will delete the answers you
gave us and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are planning an escape, that you are in immediate
danger, or that you intend to harm someone else, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will contact you again for two more interviews. We would like you to participate in these interviews at nine and fifteen months
from when you were released. You will have another chance to read a consent form like this and decide if you want to do each of the
interviews. For interviews conducted outside of a correctional facility, you will receive a monetary payment of $50 each for the nine
and fifteen month interviews. You will also receive an additional $5 each time you call to schedule an appointment for one of the
follow-up interviews.
Finally, if you participate in all 4 interviews, you will receive an additional $50.

Original (white) to RTI, Copy to Respondent

Three-month Follow-up Interview – Facility ADULT

Page 2 of 2

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

□Check here if the above statements are true.
Interviewer’s Signature

Date

Original (white) to RTI, Copy to Respondent

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR YOUTH PARTICIPANTS
Three-month Follow Up Interview-Facility
National Re-Entry Study
About the Study
The National Re-entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, you
completed a computer interview with one of our interviewers. If you now agree to continue to be in the study, we will ask you to
complete a follow up interview. This interview will take about 1 and ½ hours. Like last time, the interviewer will read questions from
a computer screen and type your answers into the computer. The interviewer will ask you questions about your housing situation,
mental and physical health, family and friends, job, alcohol and drug use, other illegal behavior, and services or treatment you have or
may receive.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. For example, if you tell us
that you used drugs or committed a crime, someone might overhear this. In order to avoid that, we will do the interview in a private
setting where no one can overhear the answers. Also, we will replace your name with an ID number to identify your interview in the
computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, your parents or correctional facility staff, will be able to find out what you said in the interview.
We will not use your name in connection with any of your answers to the interview questions. Instead, we will enter an ID number
into the computer. This ID number replaces your name and only we will be able to link it to you. Once the study is over, we will
delete the answers you gave us and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are planning an escape, that you are in immediate
danger, that you intend to harm someone else, or that you are being mistreated, we may need to inform the appropriate authorities
according to state and local law.
Future Contacts
We will contact you again for two more interviews. We would like you to participate in these interviews at nine and fifteen months
from when you were released. You will have another chance to read a consent form like this and decide if you want to do each of the
interviews. For interviews conducted outside of a correctional facility, you will receive a monetary payment of $50 each for the nine
and fifteen month interviews. You will also receive an additional $5 each time you call us to schedule an appointment for one of the
follow-up interviews. Finally, if you participate in all 4 interviews, you will receive an additional $50.

Original (white) to RTI, Copy to Respondent

Three-month Follow-up Interview – Facility YOUTH

Page 2 of 2

□

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

Check here if the above statements are true.

Interviewer’s Signature

Date

Original (white) to RTI, Copy to Respondent

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR ADULT PARTICIPANTS
Nine-Month Follow Up Interview - Community
National Re-Entry Study
About the Study
The National Re-Entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, and
probably 3 months after you were released, you completed computer interviews with one of our interviewers. If you now agree to
continue to be in the study, we will ask you to complete a follow up interview. This interview will take about 1 and ½ hours. Like
last time, the interviewer will read questions from a computer screen and type your answers into the computer. The interviewer will
ask you questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other illegal
behavior, and services or treatment you have or may receive. In an effort to thank you for taking the time for the interview, we will
give you $50.00 upon completion of the interview.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. In order to avoid that, we
will do the interview in a private setting where no one can overhear the answers. Also, we will replace your name with an ID number
to identify your interview in the computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else will be able to find out what you said in the interview. We will not use your name in connection with any of your answers
to the interview questions. Instead, we will enter an ID number into the computer. This ID number replaces your name and only we
will be able to link it to you. Once the study is over, we will delete the answers you gave us and the list that links your name and ID
number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are in immediate danger or that you intend to harm
someone else, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will contact you again for one more interview. We would like you to participate in this interview at approximately fifteen months
from when you were released. You will have another chance to read a consent form like this and decide if you want to do the
interview. If the interview is conducted outside of a correctional facility, you will receive a monetary payment of $50 for the fifteen
month interview. You will also receive an additional $5 if you call to schedule an appointment for the final follow-up interview. If
you participate in all 4 interviews, you will receive an additional $50.

Nine-month Follow-up Interview – Community ADULT

Page 2 of 2

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

□Check here if the above statements are true.
Interviewer’s Signature

Date

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR YOUTH PARTICIPANTS
Nine-month Follow Up Interview - Community
National Re-Entry Study
About the Study
The National Re-entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, and
probably 3 months after you were released, you completed computer interviews with one of our interviewers. If you now agree to
continue to be in the study, we will ask you to complete another follow up interview. This interview will take about 1 and ½ hours.
Like last time, the interviewer will read questions from a computer screen and type your answers into the computer. The interviewer
will ask you questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other
illegal behavior, and services or treatment you have or may receive. In an effort to thank you for taking the time for the interview, we
will give you $50.00 upon completion of the interview.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. For example, if you tell us
that you used drugs or committed a crime, someone might overhear this. In order to avoid that, we will do the interview in a private
setting where no one can overhear the answers. Also, we will replace your name with an ID number to identify your interview in the
computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, your parents, will be able to find out what you said in the interview. We will not use your name
in connection with any of your answers to the interview questions. Instead, we will enter an ID number into the computer. This ID
number replaces your name and only we will be able to link it to you. Once the study is over, we will delete the answers you gave us
and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are in immediate danger, that you intend to harm
someone else, or that you are being mistreated, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will contact you again for one more interview. We would like you to participate in this interview at approximately fifteen months
from when you were released. You will have another chance to read a consent form like this and decide if you want to do the
interview. If the interview is conducted outside of a correctional facility, you will receive a monetary payment of $50 for the fifteen
month interview. You will also receive an additional $5 if you call us to schedule an appointment for the final follow-up interview. If
you participate in all 4 interviews, you will receive an additional $50.

Nine-month Follow-up Interview – Community YOUTH

Page 2 of 2

□

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

Check here if the above statements are true.

Interviewer’s Signature

Date

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR ADULT PARTICIPANTS
Nine-Month Follow Up Interview - Facility
National Re-Entry Study
About the Study
The National Re-Entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, and
probably 3 months after you were released, you completed computer interviews with one of our interviewers. If you now agree to
continue to be in the study, we will ask you to complete a follow up interview. This interview will take about 1 and ½ hours. Like
last time, the interviewer will read questions from a computer screen and type your answers into the computer. The interviewer will
ask you questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other illegal
behavior, and services or treatment you have or may receive.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. In order to avoid that, we
will do the interview in a private setting where no one can overhear the answers. Also, we will replace your name with an ID number
to identify your interview in the computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, correctional facility staff, will be able to find out what you said in the interview. We will not use
your name in connection with any of your answers to the interview questions. Instead, we will enter an ID number into the computer.
This ID number replaces your name and only we will be able to link it to you. Once the study is over, we will delete the answers you
gave us and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are planning an escape, that you are in immediate
danger, or that you intend to harm someone else, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will contact you again for one more interview. We would like you to participate in this interview at approximately fifteen months
from when you were released. You will have another chance to read a consent form like this and decide if you want to do the
interview. If the interview is conducted outside of a correctional facility, you will receive a monetary payment of $50 for the fifteen
month interview. You will also receive an additional $5 if you call to schedule an appointment for the final follow-up interview.
Finally, if you participate in all 4 interviews, you will receive an additional $50.

Nine-month Follow-up Interview – Facility ADULT

Page 2 of 2

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

□Check here if the above statements are true.
Interviewer’s Signature

Date

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR YOUTH PARTICIPANTS
Nine-month Follow Up Interview-Facility
National Re-Entry Study
About the Study
The National Re-entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, and
probably three months after you were released, you completed computer interviews with one of our interviewers. If you now agree to
continue to be in the study, we will ask you to complete a follow up interview. This interview will take about 1 and ½ hours. Like
last time, the interviewer will read questions from a computer screen and type your answers into the computer. The interviewer will
ask you questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other illegal
behavior, and services or treatment you have or may receive.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. For example, if you tell us
that you used drugs or committed a crime, someone might overhear this. In order to avoid that, we will do the interview in a private
setting where no one can overhear the answers. Also, we will replace your name with an ID number to identify your interview in the
computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, your parents or correctional facility staff, will be able to find out what you said in the interview.
We will not use your name in connection with any of your answers to the interview questions. Instead, we will enter an ID number
into the computer. This ID number replaces your name and only we will be able to link it to you. Once the study is over, we will
delete the answers you gave us and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are planning an escape, that you are in immediate
danger, that you intend to harm someone else, or that you are being mistreated, we may need to inform the appropriate authorities
according to state and local law.
Future Contacts
We will contact you again for one more interview. We would like you to participate in this interview at approximately fifteen months
from when you were released. You will have another chance to read a consent form like this and decide if you want to do the
interview. If the interview is conducted outside of a correctional facility, you will receive a monetary payment of $50 for the fifteen
month interview. You will also receive an additional $5 if you call us to schedule an appointment for the final follow-up interview.
Finally, if you participate in all 4 interviews, you will receive an additional $50.

Nine-month Follow-up Interview – Facility YOUTH

Page 2 of 2

□

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

Check here if the above statements are true.

Interviewer’s Signature

Date

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR ADULT PARTICIPANTS
Fifteen-Month Follow Up Interview - Community
National Re-Entry Study
About the Study
The National Re-Entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, and
probably 3 and 9 months after you were released, you completed computer interviews with one of our interviewers. If you now agree
to continue to be in the study, we will ask you to complete a final follow up interview. This interview will take about 1 and ½ hours.
Like last time, the interviewer will read questions from a computer screen and type your answers into the computer. The interviewer
will ask you questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other
illegal behavior, and services or treatment you have or may receive. In an effort to thank you for taking the time for the interview, we
will give you $50.00 upon completion of the interview.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. In order to avoid that, we
will do the interview in a private setting where no one can overhear the answers. Also, we will replace your name with an ID number
to identify your interview in the computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else will be able to find out what you said in the interview. We will not use your name in connection with any of your answers
to the interview questions. Instead, we will enter an ID number into the computer. This ID number replaces your name and only we
will be able to link it to you. Once the study is over, we will delete the answers you gave us and the list that links your name and ID
number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are in immediate danger or that you intend to harm
someone else, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will not contact you again.
Questions
If you have any questions about the study, you may call 1-877-475-7056 (toll free) and leave a message so that someone from the
project can contact you; or you can write to us at National Re-Entry Study, 3040 Cornwallis Road, PO Box 12194, Research Triangle
Park, North Carolina, 27709-2194. If you have any questions about your rights as a research study participant, you may call the RTI
International Office of Human Subject Protections toll-free at 1-866-214-2043; or you can write to them at RTI International Office of
Human Subject Protections, 3040 Cornwallis Road, PO Box 12194, Research Triangle Park, North Carolina, 27709-2194.

Fifteen-Month Follow Up Interview – Community ADULT
Case ID Number: __ __ __ __ __ __ __ __

Do you have any questions that might help you decide if you do or do not want to continue to participate in the study?
By checking the box below, you are agreeing to continue to participate in the study. Please check the box only if:
9
9
9

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

□Check here if the above statements are true.
Interviewer’s Signature

Date

Page 2 of 2

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED ASSENT FOR YOUTH PARTICIPANTS
Fifteen-month Follow Up Interview - Community
National Re-Entry Study
About the Study
The National Re-entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, and
probably 3 and 9 months after you were released, you completed computer interviews with one of our interviewers. If you now agree
to continue to be in the study, we will ask you to complete a final follow up interview. This interview will take about 1 and ½ hours.
Like last time, the interviewer will read questions from a computer screen and type your answers into the computer. The interviewer
will ask you questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other
illegal behavior, and services or treatment you have or may receive. In an effort to thank you for taking the time for the interview, we
will give you $50.00 upon completion of the interview.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. For example, if you tell us
that you used drugs or committed a crime, someone might overhear this. In order to avoid that, we will do the interview in a private
setting where no one can overhear the answers. Also, we will replace your name with an ID number to identify your interview in the
computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, your parents, will be able to find out what you said in the interview. We will not use your name
in connection with any of your answers to the interview questions. Instead, we will enter an ID number into the computer. This ID
number replaces your name and only we will be able to link it to you. Once the study is over, we will delete the answers you gave us
and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are in immediate danger, that you intend to harm
someone else, or that you are being mistreated, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will not contact you again.
Questions
If you have any questions about the study, you may call 1-877-475-7056 (toll free) and leave a message so that someone from the
project can contact you; or you can write to us at National Re-Entry Study, 3040 Cornwallis Road, PO Box 12194, Research Triangle
Park, North Carolina, 27709-2194. If you have any questions about your rights as a research study participant, you may call the RTI
International Office of Human Subject Protections toll-free at 1-866-214-2043; or you can write to them at RTI International Office of
Human Subject Protections, 3040 Cornwallis Road, PO Box 12194, Research Triangle Park, North Carolina, 27709-2194.
Revised 7/06

Fifteen-month Follow Up Interview – Community YOUTH

Page 2 of 2

Case ID Number: __ __ __ __ __ __ __ __

□

You understand the information about the study in this form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

Check here if the above statements are true.

Interviewer’s Signature

Date

Revised 7/06

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED CONSENT FOR ADULT PARTICIPANTS
Fifteen-Month Follow Up Interview - Facility
National Re-Entry Study
About the Study
The National Re-Entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, and
probably 3 and 9 months after you were released, you completed computer interviews with one of our interviewers. If you now agree
to continue to be in the study, we will ask you to complete a final follow-up interview. This interview will take about 1 and ½ hours.
Like last time, the interviewer will read questions from a computer screen and type your answers into the computer. The interviewer
will ask you questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other
illegal behavior, and services or treatment you have or may receive.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. In order to avoid that, we
will do the interview in a private setting where no one can overhear the answers. Also, we will replace your name with an ID number
to identify your interview in the computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, correctional facility staff, will be able to find out what you said in the interview. We will not use
your name in connection with any of your answers to the interview questions. Instead, we will enter an ID number into the computer.
This ID number replaces your name and only we will be able to link it to you. Once the study is over, we will delete the answers you
gave us and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are planning an escape, that you are in immediate
danger, or that you intend to harm someone else, we may need to inform the appropriate authorities according to state and local law.
Future Contacts
We will not contact you again.
Questions
If you have any questions about the study, you may call 1-877-475-7056 (toll free) and leave a message so that someone from the
project can contact you; or you can write to us at National Re-Entry Study, 3040 Cornwallis Road, PO Box 12194, Research Triangle
Park, North Carolina, 27709-2194. If you have any questions about your rights as a research study participant, you may call the RTI
International Office of Human Subject Protections toll-free at 1-866-214-2043; or you can write to them at RTI International Office of
Human Subject Protections, 3040 Cornwallis Road, PO Box 12194, Research Triangle Park, North Carolina, 27709-2194.

Fifteen-Month Follow Up Interview – Facility ADULT
Case ID Number: __ __ __ __ __ __ __ __

You understand the information about the study in this consent form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

□Check here if the above statements are true.
Interviewer’s Signature

Date

Page 2 of 2

Page 1 of 2
Case ID Number: __ __ __ __ __ __ __ __

INFORMED ASSENT FOR YOUTH PARTICIPANTS
Fifteen-month Follow Up Interview-Facility
National Re-Entry Study
About the Study
The National Re-entry Study is a research study that is being led by RTI International and the Urban Institute and paid for by the
National Institute of Justice. We are hoping that this evaluation will help us understand the kinds of services and other activities that
might affect the lives of people leaving correctional facilities. We are interviewing approximately 4,000 individuals selected from 15
states nationwide.
Participation in the Study
You may remember that you first enrolled in the study before you were released from a correctional facility. At that time, and
probably 3 and 9 months after you were released, you completed computer interviews with one of our interviewers. If you now agree
to continue to be in the study, we will ask you to complete a final follow up interview. This interview will take about 1 and ½ hours.
Like last time, the interviewer will read questions from a computer screen and type your answers into the computer. The interviewer
will ask you questions about your housing situation, mental and physical health, family and friends, job, alcohol and drug use, other
illegal behavior, and services or treatment you have or may receive.
Voluntary Participation
The choice of whether to continue to participate in this study is completely up to you. No one will be upset or angry if you decide not
to participate. It will not affect your treatment or supervision in any manner. If you decide to participate in the study, you can refuse
to answer any of the questions asked in the interview. Just tell the interviewer you want to skip a question and the interviewer will go
on to the next question.
Benefits
There are no direct benefits to you from participating in this study. However, by participating, you are helping us learn more about the
kinds of services and activities that might help people as they return to the community after being in a correctional facility.
Risks of Study Participation
There are two risks involved in study participation. One risk is that the questions we ask might make you feel uncomfortable or be
upsetting. If you should feel uncomfortable or upset during the interview, you may ask the interviewer to take a break and/or to skip
any of the questions. The other risk is that someone might find out what you tell us during the interview. For example, if you tell us
that you used drugs or committed a crime, someone might overhear this. In order to avoid that, we will do the interview in a private
setting where no one can overhear the answers. Also, we will replace your name with an ID number to identify your interview in the
computer.
Confidentiality
We will keep what you tell us in the interview confidential. Only the people working on the study will be able to see your answers.
No one else, including, for example, your parents or correctional facility staff, will be able to find out what you said in the interview.
We will not use your name in connection with any of your answers to the interview questions. Instead, we will enter an ID number
into the computer. This ID number replaces your name and only we will be able to link it to you. Once the study is over, we will
delete the answers you gave us and the list that links your name and ID number.
Exceptions to Confidentiality
There are some exceptions to our promise of confidentiality. If you tell us that you are planning an escape, that you are in immediate
danger, that you intend to harm someone else, or that you are being mistreated, we may need to inform the appropriate authorities
according to state and local law.
Future Contacts
We will not contact you again.
Questions
If you have any questions about the study, you may call 1-877-475-7056 (toll free) and leave a message so that someone from the
project can contact you; or you can write to us at National Re-Entry Study, 3040 Cornwallis Road, PO Box 12194, Research Triangle
Park, North Carolina, 27709-2194. If you have any questions about your rights as a research study participant, you may call the RTI
International Office of Human Subject Protections toll-free at 1-866-214-2043; or you can write to them at RTI International Office of
Human Subject Protections, 3040 Cornwallis Road, PO Box 12194, Research Triangle Park, North Carolina, 27709-2194.
Revised 7/06

Fifteen-month Follow Up Interview-Facility YOUTH

Page 2 of 2

Case ID Number: __ __ __ __ __ __ __ __

□

You understand the information about the study in this form,
You have had all of your questions answered fully, and
You want to continue to participate in the study.

Check here if the above statements are true.

Interviewer’s Signature

Date

Revised 7/06

Questions and Answers About
the National Re-Entry Study
What is the purpose of the
study?

The National Re-Entry study is a
research study being conducted by RTI
International, in partnership with the
Urban Institute, and is sponsored by the
National Institute of Justice.
We hope that this study will help us
understand the kinds of services and
other activities that might affect the
lives of people leaving correctional
facilities.

Why was I selected?
You were chosen through scientific
sampling procedures from a list of
people who are scheduled to be
released from a correctional facility.
The information you provide will be kept
confidential and will not be shared with
anyone in a manner in which it can be
linked to you.

What is involved?
If you agree to participate, you will take
part in a computerized, face-to-face
interview conducted in a private setting
by one of our field interviewers. We will
interview you again three months after
your release, then again at both nine
and fifteen months after your release.
You can refuse to answer any question,
and you may also stop the interview at
any time.

Do I have to participate?
No. You do not have to participate in
this survey or respond to any questions
you do not want to answer. Your
decision to participate or not will not
affect your incarceration term or any
social service(s) you may be receiving.
In fact, correctional facility staff will not
know you agreed to participate in this
study unless you choose to tell them.

Will I be paid?

What types of questions will
be asked?
The questions will touch on your
attitudes and experiences, your health,
your use of drugs and alcohol, and any
programs and services you may have
received.

How long will it take?
The time varies, but each of the
interviews generally takes about an
hour to an hour and a half.

You will receive a study information
packet after the first interview. You will
then receive $35 for the first follow-up
interview you complete and $50 for
each of the remaining two follow-up
interviews. If you keep your contact
information up-to-date with us and
speak with one of our supervisors prior
to each follow-up interview, you will
receive an additional $5 upon
completion of each interview. If you
complete all four interviews, you will
receive an additional $50 upon
completion of the last follow-up
interview.

Are my answers safe in the
computer?

Parts of the interview will be made
inaccessible as we move through the
questions, and the entire interview will
be inaccessible once it is complete.
On the day the interview is completed,
it is electronically transmitted to RTI
International and identified only by a
code number. Your answers are then
combined with all other participants’
answers and turned into statistics for
analysis.

Who will see my answers?
Only the researchers and project staff
at RTI International will see the data,
and your name and any other
identifying information will not be linked
with your individual answers. All
information will be kept strictly
confidential and is protected under
Federal law by 28 CFR Part 22
—Confidentiality of Identifiable
Research and Statistical Information,
which regulates the disclosure of
identifiable information and requires
that RTI International protect the
confidentiality of the information you
provide.

What is RTI International?

RTI International is an independent,
not-for-profit contract research
organization located in North Carolina.
Dedicated to conducting research that
improves the human condition, RTI
International performs various types of
laboratory and social research for
government and industrial clients.
For more information, write
National Re-entry Study, 3040
Cornwallis Rd., P.O. Box 12194,
RTP, NC 27709-2194, or call
1-877-475-7056 (toll-free) and leave
a message. Someone from the
project staff will contact you.
If you have questions about your
rights as a study participant, you may
write RTI International’s Office of
Research Protection (same address
as above) or call 1-866-214-2043

Additional information about the National
Institute of Justice is available at:
http://www.ojp.usdoj.gov/nij
Additional information about RTI
International is available at:
http://www.rti.org
RTI International is a trade name of Research
Triangle Institute.

QUESTIONS AND
ANSWERS ABOUT THE

National
Re-Entry Study
Sponsored by
National Institute of
Justice (NIJ)

Conducted by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC
27709

Questions and Answers About
the National Re-Entry Study

What is the purpose of the
study?

What types of questions will
be asked?
The questions will touch on your
attitudes and experiences, your health,
your use of drugs and alcohol, and any
programs and services you may have
received.

How long will it take?
The time varies, but each of the
interviews generally takes about an
hour to an hour and a half.

Are my answers safe in the
computer?

QUESTIONS AND
ANSWERS ABOUT THE

For more information, write
National Re-entry Study, 3040
Cornwallis Rd., P.O. Box 12194,
RTP, NC 27709-2194, or call
1-877-475-7056 (toll-free) and leave
a message. Someone from the
project staff will contact you.
If you have questions about your
rights as a study participant, you may
write RTI International’s Office of
Research Protection (same address
as above) or call 1-866-214-2043

National
Re-Entry Study
Sponsored by
National Institute of
Justice (NIJ)

What is RTI International?

Conducted by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC
27709

Questions and Answers About
the National Re-Entry Study
What is the purpose of the
study?

What is involved?
We would like your child to participate
in a computerized face-to-face interview
conducted by one of our field
interviewers in a private setting at the
correctional facility. We will interview
your child again three months after his
or her release, then again at both nine
and fifteen months after release. In
each interview, your child can refuse to
answer any question, and he or she
may also stop the interview at any time.
At three and fifteen months, we will also
request an oral fluids sample, which
your child, not the interviewer, will
collect. Your child is free to allow the
interview but decline participation in the
oral fluids sample collection.

How was my child selected?
Your child was chosen through
scientific sampling procedures from a
list of people who are scheduled to be
released from a correctional facility.

How long will it take?
The time varies, but each of the
interviews generally takes about an
hour to an hour and a half.

What types of questions will
be asked?
The questions will touch on your child’s
attitudes and experiences, health, use
of drugs and alcohol, and any programs
and services he or she may have
received. The information your child
provides will be kept confidential and
will not be shared with anyone in a
manner in which it can be linked to him
or her.

Does my child have to
participate?

No. Your child does not have to
participate in this survey or respond to
any questions he or she does not want
to answer. The decision to participate
or not will not affect his or her
incarceration term or any social
service(s) he or she may be receiving.
In fact, correctional facility staff will not
know your child agreed to participate in
this study unless your child chooses to
tell them.

Are there any risks to
participating?

The only risks are that the questions
might make your child feel
uncomfortable or be upsetting, or that
someone might hear what your child
tells us during the interview. If your
child feels uncomfortable or upset
during the interview, he or she can ask
the interviewer to take a break and/or to
skip any of the questions. To avoid
someone overhearing what your child
tells us in the interview, we will do all
interviews in a private setting. For any
follow-up interviews conducted in your
home, we would appreciate your help
identifying a private setting.

What happens to the
information?

Parts of the interview will be made
inaccessible as we move through the
questions, and the entire interview will
be inaccessible once it is complete. On
the day the interview is completed, it is
electronically transmitted to RTI
International and identified only by a
code number. Your child’s answers are
then combined with all other
participants’ answers and turned into
statistics for analysis.

What about confidentiality?
Only the researchers and project staff
at RTI International will see the data,
and your child’s name and any other
identifying information will not be linked
with his or her individual answers. All
information will be kept strictly
confidential and is protected under
Federal law by 28 CFR Part 22
—Confidentiality of Identifiable
Research and Statistical Information,
which regulates the disclosure of
identifiable information and requires
that RTI International protect the
confidentiality of the information your
child provides.

What is RTI International?

For more information, write
National Re-entry Study, 3040
Cornwallis Rd., P.O. Box 12194,
RTP, NC 27709-2194, or call 1-877475-7056 (toll-free) and leave a
message. Someone from the project
staff will contact you.
If you have questions about your
rights as a study participant, you may
write RTI International’s Office of
Research Protection (same address
as above) or call 1-866-214-2043

QUESTIONS AND
ANSWERS FOR PARENTS
ABOUT THE

National
Re-Entry Study
Sponsored by
National Institute of
Justice (NIJ)

Conducted by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC
27709

Preguntas y respuestas sobre el
Estudio nacional de reingreso

¿Qué hay que hacer para
participar?

¿Es obligatoria la participación
de mi hijo(a)?

¿Cuál es el propósito del estudio?

El Estudio nacional de reingreso (National
Re-Entry Study) es un estudio de
investigación que realiza RTI International,
en colaboración con el Instituto de
Urbanización (Urban Institute) y es
patrocinado por el Instituto nacional de
justicia (National Institute of Justice).
Esperam os que este estudio nos ayude a
comprender los tipos de servicios y otras
actividades que pudieran afectar las vidas
de las personas que salen de los
establecimientos correccionales.

¿Cómo seleccionaron a mi
hijo(a)?

Su hijo(a) fue seleccionado(a) por m edio de
procedimientos de m uestreo científicos, a
partir de una lista de personas que van a ser
liberadas de un establecim iento
correccional.

Nos gustaría que su hijo(a) participe en una
entrevista computarizada en la que un
entrevistador personalm ente le haría la
preguntas en un lugar privado en el
establecimiento de la correccional.
Volverem os a entrevistar a su hijo(a) tres
meses después de que haya sido puesto(a)
en libertad. Luego nuevamente
entrevistaremos a su hijo(a) nueve meses y
quince meses después de haber salido de la
correccional. Durante cada entrevista su
hijo(a) se puede rehusar a contestar
cualquier pregunta que él o ella no desee
contestar; y además puede detener la
entrevista en cualquier momento. A los tres y
quince meses, también le vamos a pedir a
su hijo(a) que nos proporcione una muestra
de saliva, que su m ismo(a) hijo(a) obtendrá,
sin la intervención del entrevistador. Su
hijo(a) puede decidir libremente participar en
la entrevista y puede negarse a participar en
la recopilación de m uestras de saliva.

¿Qué tipos de preguntas le van a
hacer?

¿Cuánto tiempo va a tomar?
El tiempo para com pletar la entrevista varía,
pero por lo general, cada una de las
entrevistas se puede completar en una hora
o una hora y media.

Las preguntas se refieren a las actitudes y
experiencias de su hijo(a), la salud, el uso de
drogas y alcohol, y sobre los programas y
servicios que él o ella pudieran haber
recibido. La información que su hijo(a)
proporcione se mantendrá confidencial y no
se compartirá con ninguna persona de tal
manera que se pudiera asociar con su hijo(a).

No, su hijo(a) no tiene que participar en esta
encuesta ni responder a ninguna pregunta
que él o ella no desee contestar. La decisión
de participar o de no hacerlo, no afectará el
periodo de encarcelamiento ni los servicios
sociales a los que él o ella pudiera estar
recibiendo. De hecho, el personal del
establecimiento correccional no sabrá que
su hijo(a) estuvo de acuerdo en participar en
este estudio, a menos que su hijo(a) decida
informarles al respecto.

¿Hay algún riesgo al participar?

Los únicos riesgos pudieran ser: que la
preguntas hicieran que su hijo(a) se sintiera
incómodo(a) o enojado(a), o que alguien
pudiera escuchar lo que dice su hijo(a)
durante la entrevista. Si su hijo(a) se siente
incómodo o enojado(a) durante la entrevista,
él o ella puede pedir un descanso al
entrev istador, o le puede pedir que se salte
todas las preguntas que no desee contestar.
Para evitar que alguien llegará a escuchar lo
que su hijo(a) nos dice durante la entrevista,
haremos todas las entrevistas en un lugar
privado. Para las otras entrevistas de
seguim iento que se realicen en su hogar, le
pediremos a usted que por favor nos ayude
a encontrar un lugar privado para realizar la
entrev ista.

¿Qué sucede con la información?
Conform e pasamos de una pregunta a otra,
se bloqueará el acceso a ciertas partes de la
entrevista y toda la entrevista se bloqueará
una vez que se complete la entrevista. El
mism o día en que se completa, se transm ite
de manera electrónica a RTI International y
sólo se identifica con un número de código.
Entonces las respuestas de su hijo(a) se
combinan con las de todos los otros
participantes y se analizan en forma de
estadísticas.

¿Cómo se mantiene la
confidencialidad?

Sólo los investigadores y el personal del
proyecto en RTI International verán los datos
y el nombre de su hijo(a). Ninguna otra
información que lo/la pudiera identificar será
asociado con sus respuestas individuales.
Toda la información se mantendrá
estrictamente confidencial y está protegida
de acuerdo a la parte 22 del Código del
Reglam ento Federal 28 llamada: ‘La
confidencialidad de la información de la
investigación y estadística identificables’, la
cual regula la divulgación de inform ación
identificable, que requiere que RTI
International proteja la confidencialidad de la
información que proporcione su hijo(a).

¿Qué es RTI International?

RTI International es una organización de
investigación independiente, no lucrativa
ubicada en Carolina del Norte. Esta
organización está dedicada a realizar
estudios que mejoren la condición humana,
y realiza varios tipos de investigación de
laboratorio y social para dependencias del
gobierno y clientes industriales.
Para obtener más información, puede
escribir al Estudio nacional de reingreso,
3040 Cornwallis Rd., P.O. Box 12194,
RTP, NC 27709-2194, o puede llamar al 1877-475-7056 (gratis) y dejar un mensaje.
Alguien del personal del proyecto se
comunicará con usted.
Si tiene preguntas sobre sus derechos
como participante en un estudio, usted
puede escribir a la Oficina de RTI
International para la protección de
participantes (a la misma dirección arriba
mencionada) o puede llamar al 1-866-2142043 (gratis).

La inform ación adicional acerca del Instituto
nacional de justicia está disponible (en
inglés):
http://www.ojp.usdoj.gov/nij
La inform ación adicional acerca de RTI
International está disponible en:
http://ww w.rti.org
RTI International es una marca registrada del
Research Triangle Institute.

PREGUNTAS Y
RESPUESTAS PARA LOS
PADRES, ACERCA DEL

Estudio nacional
de reingreso
Patrocinado por el
Instituto nacional de justicia
(National Institute of Justice)

Realizada por
RTI International
3040 Cornwallis Road
Research Triangle Park, NC
27709

Questions and Answers About
the National Re-Entry Study
What is the purpose of the
study?

Why was I selected?
When you were selected for the first
interview, you were chosen through
scientific sampling procedures from a
list of people who were scheduled to be
released from a correctional facility.
Now we need to follow-up with you and
gather additional information. Like
before, the information you provide will
be kept confidential and will not be
shared with anyone in a manner in
which it can be linked to you.

What is involved?

If you agree to participate, you will take
part in a computerized, face-to-face
interview conducted in a private setting
by one of our field interviewers at a time
most convenient to you. As before, you
can refuse to answer any question, and
you may also stop the interview at any
time.

What types of questions will
be asked?
Similar to the first interview, the
questions will touch on your attitudes
and experiences, your health, your use
of drugs and alcohol, and any programs
and services you may have received.

How long will it take?
The time varies, but each of the
interviews generally takes about an
hour to an hour and a half.

Do I have to participate?
No. You do not have to participate in
this survey or respond to any questions
you do not want to answer. Your
decision to participate or not will not
affect your supervision term or any
social service(s) you may be receiving.
In fact, your supervision officer will not
know you are participating in this study
unless you choose to tell him or her.

Will I be paid?
You will receive $35 for the first followup interview you complete and $50 for
each of the remaining two follow-up
interviews. If you keep your contact
information up-to-date with us and
speak with one of our supervisors prior
to each follow-up interview, you'll
receive an additional $5 upon
completion of each interview. If you
complete all four interviews, you will
receive an additional $50 upon
completion of the last follow-up
interview.

Are my answers safe in the
computer?

Who will see my answers?
Only the researchers and project staff
at RTI Internationalwill see the data,
and your name and any other
identifying information will not be linked
with your individual answers. All
information will be kept strictly
confidential and is protected under
Federal law by 28 CFR Part 22
—Confidentiality of Identifiable
Research and Statistical Information,
which regulates the disclosure of
identifiable information and requires
that RTI International protect the
confidentiality of the information you
provide.

What is RTI International?

QUESTIONS AND
ANSWERS ABOUT THE

National
Re-Entry Study
Sponsored by
National Institute of
Justice (NIJ)

If you have questions about your
rights as a study participant, you may
write RTI International’s Office of
Research Protection (same address
as above) or call 1-866-214-2043

Additional information about the National
Institute of Justice is available at:
http://www.ojp.usdoj.gov/nij

Additional information about RTI

International is available at:
http://www.rti.org
RTI International is a trade name of Research
Triangle Institute.

Conducted by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC
27709

Questions and Answers About
the National Re-Entry Study

What is the purpose of the
study?

What is involved?

How long will it take?
The time varies, but each of the
interviews generally takes about an
hour to an hour and a half.

Do I have to participate?
No. You do not have to participate in
this survey or respond to any questions
you do not want to answer. Your
decision to participate or not will not
affect your supervision term or any
social service(s) you may be receiving.
In fact, your supervision officer will not
know you are participating in this study
unless you choose to tell him or her.

Are my answers safe in the
computer?

Who will see my answers?

Only the researchers and project staff
at RTI International will see the data,
and your name and any other
identifying information will not be linked
with your individual answers. All
information will be kept strictly
confidential and is protected under
Federal law by 28 CFR Part 22
—Confidentiality of Identifiable
Research and Statistical Information,
which regulates the disclosure of
identifiable information and requires
that RTI International protect the
confidentiality of the information you
provide.

QUESTIONS AND
ANSWERS ABOUT THE

What is RTI International?

Additional information about the National
Institute of Justice is available at:
http://www.ojp.usdoj.gov/nij

Additional information about RTI

International is available at:
http://www.rti.org
RTI International is a trade name of Research
Triangle Institute.

National
Re-Entry Study
Sponsored by
National Institute of
Justice (NIJ)

Conducted by
RTI International
3040 Cornwallis Road
Research Triangle Park, NC
27709

Questions asked of Juvenile Respondents to Ensure they Understood the
Consent Form
1) So that I can be certain you understand how much of your time I need today,
please tell me how long this interview will take.
ANSWER: 1 ½ hours. IF R DOES NOT SAY “1 ½ HOURS,” RE-READ THE
FOLLOWING STATEMENT: This interview will take about 1 and ½ hours.
2) Since some of the questions are sensitive, I also need to be sure you understand
the topics I will be asking you about. What topics will we ask about in the
interview?
ANSWER: HOUSING, MENTAL AND PHYSICAL HEALTH, FAMILY AND
FRIENDS, JOB, ALCOHOL AND DRUG USE, OTHER ILLEGAL
BEHAVIOR, SERVICES OR TREATMENT. IF R DOES NOT LIST ALL THE
TOPICS, RE-READ THE FOLLOWING STATEMENT: We will ask you
questions about your housing situation, mental and physical health, family and
friends, job, alcohol and drug use, other illegal behavior, and services or treatment
you have received.
3) Again so that I can be certain you understand what I’ve just read; please tell me
whether you can choose not to do the interview.
ANSWER: YES. IF R DOES NOT ANSWER THAT HE/SHE CAN CHOOSE
NOT TO DO THE INTERVIEW, RE-READ THE FOLLOWING
STATEMENT: The choice of whether to participate in this pilot study is
completely up to you. You can choose not to do the interview.

4) If you choose to do the interview, tell me what you should do if you do not want
to answer a specific question?
ANSWER: TELL ME YOU DON’T WANT TO ANSWER THAT QUESTION.
IF R ANSWERS INCORRECTLY, RE-READ THE FOLLOWING
STATEMENT: You can refuse to answer any of the questions asked in the
interview. Just tell me you want to skip a question and I will go on to the next
question.
5) Again, I need to be sure you understand what I’ve just read to you. Please tell me
what the risks are of doing the interview.
ANSWER: THE QUESTIONS MIGHT MAKE ME FEEL UNCOMFORTABLE
OR BE UPSETTING; SOMEONE MIGHT FIND OUT WHAT I TELL YOU
DURING THE INTERVIEW. IF THE R DOES NOT LIST BOTH RISKS, REREAD THE ENTIRE PARAGRAPH ABOVE.

6) I want to make sure you understand what I just read about confidentiality. Can
you please tell me what confidential means?
ANSWER: IT MEANS THAT NO ONE ELSE BESIDES THE PEOPLE
WORKING ON THE STUDY WILL BE ABLE TO SEE MY ANSWERS OR
FIND OUT WHAT I SAID IN THE INTERVIEW. IF THE R DOES NOT
STATE THIS, RE-READ THE FOLLOWING STATEMENT: It means that no
one else besides the people working on the study will be able to see your answers
or find out what you said in the interview.
7) Can you tell me what it means to replace your name with an identification
number?
ANSWER: IT MEANS THAT RATHER THAN ENTERING MY NAME INTO
THE COMPUTER, YOU WILL ENTER AN ID NUMBER INSTEAD. THIS ID
NUMBER REPRESENTS ME BUT ONLY YOU WILL BE ABLE TO LINK IT
TO ME. IF THE R DOES NOT STATE THIS, RE-READ THE FOLLOWING
STATEMENT: It means that rather than entering your name into the computer,
we will enter an ID number instead. This ID number represents you but only we
will be able to link it to you. Once the study is over, we will delete the answers
you gave us and the list that links your name and ID number.
8) It is really important that I’m sure you understand why we might need to break
our promise of confidentiality, so I need you to answer one more question before
we finish reading this form. What are some things you could tell us that might
cause us to break our promise of confidentiality?
ANSWER: I AM PLANNING AN ESCAPE; I AM IN IMMEDIATE DANGER;
I INTEND TO HARM SOMEONE ELSE; I AM BEING MISTREATED. IF
THE R DOES NOT LIST ALL 4 CONDITIONS, RE-READ THE ENTIRE
PARAGRAPH ABOVE.

Appendix F. Administrative
Data Protocols and Data
Transfer Protocol

F-1

SVORI Multi-site Evaluation Administrative Data Collection Protocol
Information for Adult Correctional and Juvenile Justice Agencies with authority
over Probation/Parole data
This document provides information on the SVORI multi-site evaluation, describes the types of
administrative data needed from the impact sites for the impact evaluation, and identifies the
target population for which data are needed as well as the anticipated schedule for obtaining
data.

Background
The SVORI impact evaluation is focused on determining the extent to which SVORI program
participants have better post-release outcomes than comparable individuals who did not
participate in SVORI programming. The evaluation design is shown in Figure 1. Annual surveys
(in 2003, 2005, and 2006) of the local SVORI program directors are informing the
implementation assessment. Data for the impact evaluation are being collected from two
sources—offender interviews and agency administrative data.
Figure 1. SVORI Evaluation framework.

Community Context
y
y
y
y
y

Inputs:
The SVORI
y Federal Funding &
Other Resources
y Technical Assistance
y Federal Grant
Requirements

Offender Context
y
y
y
y
y
y

Population Characteristics
Unemployment Rates
Service Availability
Residential Stability
Post-release Supervision Structure

Outputs:
Implementation

Throughputs
y Local Partnership
Formation &
Functioning
y State & Local
Resources

Population Characteristics
Criminal History
Mental & Physical Health
Substance Abuse
Education/Training/Work Experience
Family Ties

Outcomes

In-Prison
y
y
y
y
y

Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services

y
y
y
y
y

Coordination/Supervision
Education/Training
Family Services
Health Services
Transition Services

Community

Post-Supervision

Offender
y
y
y
y
y
y
y
y

Community involvement
Employment
Family Contact/Stability
Health/Mental Health
Housing
Recidivism
Substance Use
Supervision Compliance

Systems
y Rearrest Rates
y Reincarceration Rates
y Systems Change

y Community Reintegration
Activities

Evaluation
Components

Impact
Evaluation

Implementation Assessment

Cost-Benefit Analysis

The SVORI evaluation model begins with the use of Federal funding, technical assistance and
Federal grant requirements (Inputs) to facilitate local partnership formation and function in
combination with state and local resources (Throughputs) to yield a comprehensive SVORI
program that begins during imprisonment and continues into the community (Outputs).
SVORI Administrative Data Protocol

Page 1 of 10

January 2007

The outputs include programming and services provided to and received by subjects prior to
and after release. Examples of programming and services include the following:
•
•
•
•
•
•

counseling programs (e.g., anger management, batterer intervention treatment
programs, mental health counseling, sex offender treatment, and substance abuse
treatment)
educational programs (e.g., GED classes or vocational training)
employment programs (e.g., resume preparation, interview preparation)
health services (e.g., dental or medical)
life skills classes
reentry (or release) planning

SVORI participants are expected to receive more programming and services than comparison
subjects. However, comparison subjects are also expected to receive some services and, in
fact, some early analyses suggest that comparison subjects in some impact sites are receiving
more services than SVORI subjects in other sites. This was expected since each site began its
SVORI program from a baseline level of services that varies across states. Thus, it is important
for the impact evaluation to have individual-level data on services and programs. The subject
interviews ask about service receipt, but it is important to obtain any data from agency records
to supplement the interview data. 1
The primary outcome of interest is recidivism, which will be measured from official records as
the following:
•
•
•
•

Technical violation and/or revocation
Arrest
Conviction
Reincarceration

In addition, we will have self-report criminal behavior information from follow-up interview data.
The key intermediate outcomes are as follows:
•
•
•
•
•

Employment
Mental and physical health
Housing
Substance use
Supervision compliance

Other intermediate outcomes that will be more difficult to evaluate because they pose particular
measurement challenges are community reintegration and family contact/stability. These
intermediate outcomes will be assessed from interview data.
Figure 1 also explicitly acknowledges the potential impact of Community and Offender Contexts
(characteristics) on the provision of services (i.e., the outputs) and outcomes. These
characteristics include measures at the community level (e.g., population demographics and
employment rates that will be obtained from official statistics) and of subject characteristics,
including, importantly, measures of the following:
•
1

Demographic information

Cost-benefit evaluations are being conducted in a subset of impact sites. The availability of detailed, individual-level
data on services and programming is particularly important to these evaluations as they are the source of
information for the estimation of costs.

SVORI Administrative Data Protocol

Page 2 of 10

January 2007

•
•
•
•

Criminal history
Education, training, and work experiences
Physical and mental health
Substance use

In addition, to the extent that data are available, measures of family stability will be important for
the analyses (these data will come primarily from interview data).

Data sources
Data for the impact evaluation come from two sources:
•

Interviews with SVORI participants and comparison subjects—four rounds of
interviews are being conducted. Baseline data collection ended in November 2005 with
interviews being completed with approximately 2,700 SVORI/non-SVORI respondents
in fourteen states (16 SVORI programs). Follow-up interviews at 3 and 9 months post
release have also been completed. The final wave of follow-up interviews is being
conducted 15 months post release and will be completed in April/May 2007.

•

Administrative data from state agencies will be used to supplement the offender
interview data.
o

Data are needed from all states originally selected as adult impact sites (IN, IA,
KS, ME, MD, MO, NV, OH, OK, PA, SC, VA, and WA) from the following
agencies:
adult correctional agencies (DOC)
probation and parole agencies (P/P) [if separate from DOC]
state-level agencies who manage arrest data (Arrest Data Agency)

Data are needed from all states originally selected as juvenile impact sites (CO,
FL, KS, ME, and SC) from the following agencies:
juvenile justice agencies (DJJ)
adult correctional agencies (DOC)
probation and parole agencies (P/P) [if separate from DJJ/DOC]
state-level agencies who manage arrest data (Arrest Data Agency)

In four sites (IA, MD, PA, and SC), we also plan to be seek employment-related
data from the state Employment Security Commission (ESC). (This is a special
data collection undertaking and is not addressed further in this document.)

Figure 2 shows the SVORI administrative data sources (other than ESC), the relationship
between these sources and the schedule for data collection. Importantly, the figure illustrates
the need to have appropriate identification numbers in order to link the various data files at
the individual subject level. The individual-level data from the agencies will be merged with
the individual interview data. Because we will merge data from various sources, all data
must have identifiers that will allow us to link subjects across different data sources. 2

All aspects of this evaluation, including obtaining and managing administrative data, have been and continue to be
thoroughly reviewed by an RTI Institutional Review Board (IRB). We will be happy to provide detailed
information on our protocols and any other information pertinent to the security of these data.

SVORI Administrative Data Protocol

Page 3 of 10

January 2007

Figure 2 SVORI data collection.
2004

2005
7/04

Baseline

2006

4/06
9-month

11/06
15-month

4/07

+ ID

+I
Ds

2008

11/05

3-month

Interview
Data

2007

Corrections

Data
Data

Juvenile
Justice

Impact
Subjects
Data Base

Data

Arrest Data
Agency

2004

Data

2005

2006

2007

2008

The DOC and DJJ agencies will be asked for data in February 2007. These data should
provide information through December 31, 2006 (or the date of file creation) These data will be
used for our initial outcome analyses, and importantly will be the main source for offender
identifiers that we will subsequently supply to other agencies. 3 We will make a second request
for data in early 2008; this second data set should include information for calendar year 2007.
We are asking for recidivism and updated demographic information data twice in order to
accomplish the following:
1. Provide short-term impact findings: Data through the end of December 2006 will provide
us with at least 12 months of follow-up data on all subjects and will facilitate interim
reporting.

The issue of cross-linking identifiers may be more difficult for juvenile agencies and subjects and will be addressed
individually with each site.

SVORI Administrative Data Protocol

Page 4 of 10

January 2007

2. Provide long-term impact findings: Data through the end of December 2007 will provide
us with at least 24 months of follow-up data on all subjects and will be used for the final
report.
The administrative data collection involves two components, each of which are addressed
below:
1. For whom will we request administrative data?
2. Which data elements (variables) and When (schedule) will we request data?

For whom will we request administrative data?
Figure 3 provides a diagram illustrating the three potential groups of subjects for whom a
corrections or juvenile justice agency could provide administrative data. The core subjects of
interest to the impact evaluation are those who participated in offender interviews, i.e., the
interview respondents. However, we are requesting data on one of the following two groups:
(1) all individuals enrolled in the SVORI program between the start of the program and
December 31, 2005 and all individuals comparable to those individuals who were released
in 2004 and 2005 (but did not participate in SVORI) 4 OR
(2) all individuals on the lists of eligible respondents 5 provided by the agency during the
baseline enrollment period for the offender interviews (7/04-11/05).
Figure 3. Potential groups for administrative data.

All SVORI & Comparison Subjects

Potential respondents

By having administrative data on either (1) all
SVORI participants (enrolled through
12/31/2005) and all comparison subjects
released between 1/1/2004 and 12/31/2005
or (2) all eligible respondents submitted to
the evaluation, we will have larger site-level
samples than are available through our
interview respondent samples.

These expanded samples will provide greater
statistical power to detect treatment effects.
Specifically, these “admin-only” analyses will
Interview
respondents
be based on substantially larger samples than
“interview-only” analyses, providing greater
statistical power to assess treatment effects
(albeit with fewer measures). As the numbers
of subjects interviewed very greatly over the
16 programs included in the impact
evaluation, the need to obtain data on additional subjects is especially critical in some sites and
for some demographic groups, such as women. To the extent that sites are interested in sitelevel analyses, it is important that they provide us with administrative data on larger samples.

Although some impact sites released SVORI participants in 2003, most released few participants prior to 2004. For
simplicity, we are seeking comparison subjects who were released only in 2004 or 2005.
5
For confidentiality reasons, it is important that we obtain data on the full list of potential respondents so that we do
not reveal to agencies which individuals did and did not choose to participate in the interviews. This “masking”
affords an additional protection of the confidentiality that was promised to those who consented to interviews.

SVORI Administrative Data Protocol

Page 5 of 10

January 2007

In addition, these larger samples will be useful as we examine selection issues: How
comparable are the subjects we interviewed to those eligible for interviews? How comparable
are the subjects we interviewed and the subjects eligible for interviews to all SVORI participants
and 2004-2005 comparisons? Answering these questions will improve our ability to generalize
our findings beyond our interview samples.
In summary, therefore, we would like the sites to provide administrative data on one of the
following groups listed in Table 1. The groups are listed in order of our preference (reflecting
our ability to conduct the most thorough examination of impact).
Table 1. Subjects for whom administrative data are requested (listed in order of preference).

SVORI
1. All program participants from local SVORI
program initiation through 12/31/2005

OR
2. All SVORI participants submitted to the
evaluation as eligible respondents.

COMPARISON
1. All ‘comparison’ individuals released
between 1/1/2004 and 12/31/2005. This
group would include both:
a. Those submitted to the evaluation as
eligible respondents and
b. Any other individuals who met SVORI
eligibility criteria but who did not
participate in SVORI (excluding those
who refused to participate in the
SVORI program if known)
OR
2. All comparison subjects submitted to the
evaluation as eligible respondents.

Please Note: If your agency routinely produces annual release cohort research data files, it
may be easier for you to provide us with these files (since in so doing, you will not have to
produce a “special” dataset for the SVORI evaluation). If you wish to provide annual release
files for the years of interest—e.g., all releases during calendar year 2004 and 2005—we will
accept these files instead of a “special” SVORI file as long as the files contain necessary
identifiers (including name, gender, DOB, and any agency and state identifiers [e.g., DOC# or
DJJ#]) so that we can link the administrative data to our interview data as well as to the
administrative data we receive from other agencies (arrest and employment security
commissions).

Which data elements will we request and when will we request them?
The analysis plan for the evaluation is based on estimating models of outcomes that include
independent variables (either control or explanatory variables) as measures of factors
theoretically linked to the outcome of interest. Thus, for example, our models of criminal
recidivism will include measures of criminal history as well as demographic variables plus
measures of treatment needs and service receipt. In particular, it will be important—to the extent
that data are available—to have the following:
•
•
•
•
•

Pre-release measures of the outcomes of interest (e.g., previous criminal history for
recidivism models; employment quality (stability, etc.) for employment models)
Demographic characteristics
Measures that describe risk factors or treatment needs related to each outcome
Measures of program and service delivery
Measures of outcomes.

SVORI Administrative Data Protocol

Page 6 of 10

January 2007

We will work with each DOC or DJJ agency to determine the most efficient way to extract
records for the appropriate individuals, depending on which of the two options in Table 1 is
approved. This may involve working with existing annual release files (as mentioned above),
providing the agency with a file (in whatever format would be best) that contains identifiers 6
(e.g., DOC numbers, DOB), or developing a set of filters to limit the records to those offenders
who meet SVORI criteria.
Table 2 provides information on the types of data and relevant time periods we would like to
obtain by year of request. We are aware that the availability and quality of data will vary by
state. In some cases, the information may only be available and/or retrievable manually, which
we will need to note. 7
Information is needed for both the “instant incarceration” (the admission related to the sentence
being served at inclusion in study) and any returns to incarceration (reincarcerations) 8 . The
data on the instant incarceration will be used to construct independent variables for the outcome
models. The reincarceration data will be used to construct dependent variables for the outcome
models.
Community supervision (probation/parole) data will also be needed to develop measures of
post-release service/program receipt and performance on the supervision following the instant
incarceration. In particular, measures of technical violations, revocations, positive drug tests,
and other indicators of supervision compliance (payment of fees for example) will be used to
develop outcome indicators (dependent variables). Information on program participation (e.g.,
participation in substance abuse treatment) will be used to construct independent variables.
Our initial data request to correctional and juvenile justice agencies will be made in early 2007
for the time period through December 31, 2006. A second request will be made in early 2008
for data through December 31, 2007.
Finally, as noted earlier, we are willing to accept 2004 and 2005 release cohort data files if
these are easier for the agency to produce.

These identifiers would be provided for respondents (which includes SVORI and comparison offenders) provided to
us during the baseline enrollment period (July 2004 through December 2005), as well as an expanded list of
SVORI participants enrolled outside of our baseline sampling period that was provided to us by SVORI program
directors in early 2006. We do not have identifiers for potential comparison offenders other than those provided
during the baseline sampling period and would therefore need the agency’s assistance to identify potential
comparison offenders from the remaining months in the two-year range.
7
If automated data are not available, the researchers may decide to manually code some data in certain sites,
depending on the type of data needed by the research team.
8
In addition to returns to DOC, we will also be making a separate request to the appropriate state agency for
information regarding subsequent arrests and convictions.

SVORI Administrative Data Protocol

Page 7 of 10

January 2007

Table 2. Data constructs for Correctional Agencies (O=Obtain; U=Update)
Note on 2008 update requests: If it is easier for an agency to provide complete information in 2008
(rather than updating information), we will accept a full “data dump.”

Constructs

2007

2008

Instant Incarceration (incarceration that led to inclusion in study) data
SVORI flag

Identifiers (specific to the person, admission, sentence, or charge); any
available statewide ID numbers used across agencies within the state
would be particularly helpful
Demographic information: age (DOB), gender, race, ethnicity,
education
Employment at time of admission (employed, wages, other indicators
as available)
Family status at time of admission including marital status, number of
children
County (and/or city) of residence at time of admission

County (and/or city, circuit, or judicial district) of conviction

Criminal history: number of prior convictions, commitments, or
community supervisions; type of prior convictions, commitments, or
community supervisions
Instant offense(s) [i.e., offense(s) for incarceration that led to inclusion
in the dataset)]: number and types of charges, sentence length(s), date
of admission
Assessments during incarceration (risk, needs, substance use, other)
information (scores and dates)
Infractions (if available)

Visitation records (if available)

O
O
O

Programs, services, and treatment participation information (if
available), including types of programs, length of participation (dates)
Release information: date of facility release, sentence end date (if
different from date of release), parole/community supervision flag (type
of release)
Enrollment criteria: fields needed to identify comparison offenders that
are not otherwise included in the request
Instant Community Supervision (community supervision immediately
following prison sentence that led to inclusion in study) data
Contact information (e.g., current address if available—to help us
locate interview participants for follow-up interviews)
Type of supervision (e.g., probation, parole, or any state-specific
labels)
Offense(s) that led to the supervision term: number and type of
offenses, length of supervision term, date supervision began

SVORI Administrative Data Protocol

Page 8 of 10

O
O

January 2007

Constructs
Instant Community Supervision (community supervision immediately
following prison sentence that led to inclusion in study) data (cont.)
Initial county of supervision (if available)

2007

2008

Supervision level (initial and final; all changes if available)

Initial conditions of supervision (any offender-specific conditions in
addition to standard conditions, if available)
Compliance with conditions (payment of fines, fees and restitution)

Violations and infractions (number, date, type, disposition)

Supervision contacts (type and number of each; dates and length of
contacts if available)
Programs, services, treatment (begin and end dates, completion
status, if available)
UA results (type tests, drugs positive, dates if available)

Employment (months employed, type of employment, etc. if available)

Date supervision ended (projected or actual termination dates)

Termination code [if applicable—some cases will not yet have
terminated] (e.g., successful, unsuccessful, any available detail on
reason for termination)

Recidivism data
Reincarcerations
Reincarceration offense(s): number and type of charges, sentence
length(s), date of admission
Release information: date of facility release(projected if still
incarcerated), sentence end date (if different from date of release),
parole/community supervision flag (type of release)
If revocation, type and date
New community supervisions
Type of supervision (e.g., probation, parole, or any state-specific
labels)
Initial county of supervision
Offense(s) that led to the supervision term: number and type of
offenses, length of supervision term, date supervision began
Termination code (e.g., successful, unsuccessful, any available
detail on reason for termination)

*Release information for the instant incarceration will be needed in 2007 and 2008 because not all
individuals in either the interview sample or the potential-interview sample had been released when
we concluded baseline interviews in November 2005.

SVORI Administrative Data Protocol

Page 9 of 10

January 2007

Data Transfer
RTI and UI evaluation staff will work with the relevant agencies on the format and transmission
of the data. The data must be transferred to RTI in a secure manner and options have been
developed to submit the files securely using either the evaluation website or via
FedEx. Evaluation staff working with the agencies will ask that two sets of electronic data files
be generated as an additional measure to protect this potentially sensitive information:
1. one data file should contain only the identifiers and a special unique ID# (e.g., an auto or
sequential number) generated for the purpose of transfer of these data;
2. the other data file should contain only the requested data elements and the special
unique ID#.
For either transfer method (website or FedEx), the two files should be sent in two separate
transmissions. If the agency chooses to upload data using the evaluation website, a short
period of time should be allowed to elapse between transferring the two files. If FedEx is used,
the files containing the data should be sent one day and the file containing the ID#s should be
sent the next day. These measures are important to assuring the security of the data and the
confidentiality of the individuals whose data we are obtaining.

Summary
In summary, the purpose of the administrative data collection is to obtain data that will allow us
to construct measures for the models that will be used to assess the impact of SVORI in the
sixteen impact sites.
The schedule for data collection is as follows:
•

•

Early 2007: Initial data run to obtain all requested information on identifiers, criminal
history, instant incarcerations and community supervisions through December 31,
2006; recidivism data will also be obtained for those returned to prison or starting new
community supervision sentences by 12/31/2006.
Early 2008: Updates to obtain additional release, programming (if available), and
recidivism information through December 2007.

SVORI Administrative Data Protocol

Page 10 of 10

January 2007

SVORI Multi-site Evaluation Administrative Data Collection Protocol
Part 2: Internal Guide for RTI and UI Staff
This internal document was developed for RTI and Urban staff members who are requesting
administrative data from state and local agencies for the SVORI Multi-site evaluation. This
document describes the process and schedule for requesting data and outlines the information
we need to collect from each agency as negotiations proceed.
Purpose: We will use administrative data from Departments of Corrections or Juvenile Justice
[DOC/DJJ], Probation and Parole/Community Corrections Agencies [P/P], agencies that
manage state arrest data [Arrest Data Agency or ADA] and Employment Security Commissions
[ESC] to supplement the offender interview data being collected in the multi-site evaluation
impact sites (as well as data for Maine juveniles and Virginia adult SVORI programs). This data
collection will:
1. Provide additional information, including recidivism information, on survey respondents
and
2. Expand, where feasible, the sample of SVORI participants and non-participants by
including all SVORI participants and additional comparison subjects.
In addition, we will use the administrative data to examine issues related to selection bias (i.e.,
to determine whether our interview respondents are representative of the sampling frame from
which they were drawn and whether the sampling frame is representative of the entire
population of SVORI participants and comparable, non-SVORI participants).
This document accompanies “Part 1” of the protocol. The companion document, which is
intended to be shared with local site staff, provides information on the SVORI multi-site
evaluation, describes the types of administrative data needed, identifies the target population for
which data are needed, and includes a less-detailed schedule for obtaining data. We have
created different versions of this companion document for each type of agency (DOC/DJJ, P/P,
and Arrest) in order to reduce confusion of those receiving the documents (e.g., the P/P version
doesn’t go into nearly as much detail about the various pools of offenders we are requesting
data for or make such a strong pitch for getting identified data).

For whom will we request administrative data?
The core subjects of interest to the impact evaluation are those who participated in offender
interviews, i.e. the interview respondents. However, we are requesting data on one of the
following two groups:
(1) all individuals enrolled in the SVORI program between the start of the program and
December 31, 2005 and all individuals comparable to those individuals who were released
in 2004 and 2005 (but did not participate in SVORI) 1 OR
(2) all individuals on the lists of eligible respondents provided by the agency during the baseline
enrollment period for the offender interviews (7/04-11/05).
PLEASE NOTE: We are also willing to accept a cohort release file (data on all offenders
released during a calendar year) if that is easier for the site and specific agencies involved. If
the agency prefers to provide cohort release files, we will need data that include releases during
Calendar Years 2004 and 2005. There is also a small set of sites where the DOC may limit the
1

SVORI Administrative Data Protocol—Part 2

Page 1 of 12

January, 2007

data they provide only to those offenders who agreed to release administrative data on their
consent forms.
Table 1 shows the options for the samples for which we wish to obtain administrative data. The
specific sample on which we request data will depend upon which is easiest for a specific site.
The groups are listed in order of our preference (reflecting our ability to conduct the most
thorough examination of impact).
PLEASE NOTE: “Part 1” of the protocol that we are sharing with sites emphasizes Options 1
and 2 (and mentions the possibility of 3—providing cohort release files); option 4 is listed only in
this internal document as it is not relevant in most sites and would be potentially confusing.
Table 1. Subjects for whom administrative data are requested (listed in order of preference).

OPTION
Option 1

OR
Option 2
OR
Option 3
OR
Option 4

SVORI
All program participants from local
SVORI program initiation through
12/31/2005

COMPARISON
All ‘comparison’ individuals released
between 1/1/2004 and 12/31/2005. This
group would include both:
1. Those submitted to the evaluation
as eligible respondents and
2. Any other individuals who met
SVORI eligibility criteria but who
did not participate in SVORI
(excluding those who refused to
participate in the SVORI program
if known)
OR
OR
All SVORI participants submitted to
All comparison subjects submitted to the
the evaluation as eligible respondents. evaluation as eligible respondents.
Option 3. All offenders released in Calendar Years 2004 or 2005.
OR
All SVORI participants who completed
baseline interviews and signed
administrative data release consent
forms.

OR
All comparison subjects who completed
baseline interviews and signed
administrative data release consent
forms.

For DOC/DJJ agencies
For Option 1 (all SVORI enrollees, plus comparison subjects released in 2004 and 2005), we
will provide to the agency:
1. a list of DOC/DJJ numbers (and any other identifiers the agency might need to uniquely
and correctly identify records) for all offenders enrolled as SVORI participants based on
enrollment lists we obtained from each SVORI program;
2. a list of DOC/DJJ numbers (and any other necessary identifiers) for all comparison
subjects submitted to us as eligible respondents on the monthly release lists.
PLEASE NOTE: The second list will not include the full set of comparison offenders released
since the beginning of 2004 because our list of eligible respondents only includes offenders
released between July 2004 and December 2005. Therefore, we need to work with the DOC (or
DJJ for juvenile sites) staff person fulfilling our administrative data request to develop and apply
a set of filters that approximates the original process used to identify the eligible comparison

SVORI Administrative Data Protocol—Part 2

Page 2 of 12

January, 2007

offenders on the monthly lists. The process the site (MIS person and/or PD) used to identify
appropriate comparison subject was documented earlier by site liaisons at RTI or UI. We will
need to work with the administrative data MIS person to ensure that similar filters are developed
for the expanded comparison set.
For Option 2, we will provide the agency with a list of DOC/DJJ numbers (and any other
identifiers the agency might need to uniquely and correctly identify records) for all offenders
contained on our “master list” of SVORI and comparison offenders submitted to us as eligible
respondents on monthly release lists. 2
For Option 3, which represents full release cohorts, we will not need to provide identifiers to the
agency.
For Option 4, we will provide the agency with a list of DOC/DJJ numbers (and any other
identifiers the agency might need to uniquely and correctly identify records) for SVORI and
comparison subjects for whom we have consent forms.

For other agencies
Once we have worked with the DOC/DJJ to identify the full set of offenders for whom we will
receive administrative data, we need to get as many identifiers as possible from DOC/DJJ so
that we can provide a list of appropriate identifiers to the probation/parole and arrest data
agencies, as well as the adult DOC for juvenile sites (see Table 2 below).
Note on our enrollment lists: We have requested (or compiled based on rolling enrollment
information we have been receiving from sites) enrollment lists that are supposed to contain a
complete list of every offender who has ever been enrolled in SVORI for each site. We
requested identifiers (in addition to other information) for these lists to enable us to merge the
information with our offender data, and we are also hoping to use these numbers to identify
offenders for administrative data requests, as described above. Table 2 shows the types of
identifiers contained in each enrollment list.
Table 2. Enrollment list identifiers
Site
Adults:
IA
IN
KS
MD
ME
MO
NV
OH
OK
SC
WA

Last
Name

First
Name

Middle
Name

x
x
x
x
x
x
x
x
x
x
x*

x
x
x
x
x
x
x
x
x
x
x

x
some

some initials

some

DOC/DJJ#

DOB

SSN

x
x
x

x
x
x
x

x
x
x
x
x

x
x
x
x

x
x
x
x
x
x
x

Other

specify

SID

SID, Loc Code (?)

For confidentiality reasons, it is important that we obtain data on the full list of potential respondents so that we do
not reveal to the agencies who did and did not choose to participate in the interviews (unless, as noted in option 4, an
agency agrees to provide data only on consented individuals). This “masking” affords an additional protection of the
confidentiality that was promised to those who consented to interviews.

SVORI Administrative Data Protocol—Part 2

Page 3 of 12

January, 2007

Table 2. Enrollment list identifiers
Site
Juveniles:
CO
FL
KS
SC

Last
Name

First
Name

x
x
x
x

Middle
Name

some
some

DOC/DJJ#

DOB

SSN

x
x
x

Other

specify

*some names missing

Administrative Data Sources
We are seeking administrative data in all impact sites from adult correctional agencies (DOC),
probation and parole agencies (P/P), and state-level agencies who manage arrest data (Arrest
Data Agency). In the juvenile impact sites, we are also seeking administrative data from the
juvenile justice agency (DJJ). In four sites (IA, MD, PA, and SC), we also hope to be able to
obtain employment-related data from the state Employment Security Commission (ESC). Figure
2 in the Part 1 document for DOC/DJJ agencies shows the SVORI data sources (other than
those from ESC), the relationship between these sources and the schedule for data collection.
Importantly, the figure illustrates for the agencies the need to have appropriate IDs in order to
link the various data together at the individual subject level. The goal is to obtain individual-level
data from appropriate agencies that will be merged with the individual interview data. Because
we will merge data from various sources, all data will have to have identifiers that will allow us to
link subjects across different data sources. Also, with all files, we need to be certain they
communicate to us the date of extraction.
Table 3 summarizes the approach to data collection, identifying the five agency types from
which data will be sought and the programs for which data will be required. The DOC/DJJ data
provide the base for the administrative data collection by (1) identifying all of the individuals who
will be included in the analyses, and (2) providing administrative data on the instant
incarceration that led to inclusion in the sample. In addition, these data will be used to provide
reincarceration information. Thus, data from these agencies will need to be collected first; but
will also be needed later to track recidivism. (Per NIJ’s request, we plan to follow all subjects for
at least 24 months following release, which means that we will need recidivism data from these
agencies through December 2007.)
Negotiations with DOC and DJJ agencies began in spring 2006 and are ongoing. We need to
obtain DOC/DJJ data first. The first data file to be received should contain data through
12/31/2006. We anticipate that most agencies will be able to generate this data no later than
February of 2007, although this will need to be confirmed with each site. We would like to
finalize the list of data elements and file format with DOC/DJJ agencies by the end of 2006 so
that we are prepared to make the formal request in early 2007. We anticipate providing
identifiers to P/P and ADA agencies with a formal request for data through December 2006 (or
to date) in May or June of 2007 and would like to obtain data from them no later than August
2007.
Note that the list of data elements in the “Part 1” document for agencies has been
updated, so you should compare any list you reviewed with agencies earlier in 2006 with
the current list to see if there any new elements we need to pursue.

SVORI Administrative Data Protocol—Part 2

Page 4 of 12

January, 2007

Table 3. Data sources
Agency
Department of
Correction (DOC)

Department of Juvenile
Justice (DJJ) [juvenile
impact programs only]

Probation/parole [i.e.,
community supervision]
departments (P/P)

Agency that maintains
statewide arrest records
(Arrest Data Agency—
ADA)
Employment Security
Commission (ESC)
[Adult programs in IA,
MD, PA and SC]

Domains and Time Frames (see Part 1 companion
document for more detail)
Data through December 2006
• Demographic data
• Criminal history (including information on current and prior
convictions and incarcerations)
• Institutional data on instant incarceration (e.g., programs,
assessments, treatment, infractions)
• Family visitation and contact information (if available)
• Identifiers for use in obtaining data from Probation/Parole
and Arrest Data Agency
• Reincarceration (revocation or new offense—including
offense, sentence, length of stay)
Data through December 2007
Reincarceration—including offense, sentence, length of stay,
infractions), updated release and programming information
Data through December 2006
• Demographic data
• Criminal history (including information on instant
incarceration)
• Institutional data (e.g., programs, assessments, treatment,
infractions)
• Family visitation and contact information (if available)
• Identifiers for use in obtaining data from DOC,
Probation/Parole and Arrest Data Agency
• Reincarceration (revocation or new offense—including
offense, sentence, length of stay)
Data through December 2007
• Reincarceration—including offense, sentence, length of
stay), updated release and programming information
Data through December 2006
• Criminal history
• Performance on instant supervision
• Recidivism (including offense, sentence, and termination
information)
Update through December 2007
• Performance on instant supervision
• Recidivism (including offense, sentence, and termination
information)
Data through date of data extract
• Criminal history (e.g., prior arrests)
• Recidivism
Update through date of data extract
• Recidivism
Data through 2007
• Post-release employment and wages

SVORI Administrative Data Protocol—Part 2

Page 5 of 12

When to
Request
February 2007

February 2008
February 2007

February 2008
June 2007

June 2008

June 2007

June 2008
June 2008

January, 2007

Departments of Correction/Juvenile Justice
•

•

These are the agencies from whom we obtained our lists of potential respondents. We
have research agreements and/or MOAs with these organizations. However, not all of
these agreements negotiated in 2004 include agreements for administrative data so an
additional agreement may be needed for some sites.
Data from these agencies will provide data for our impact analyses, but also are
needed to obtain identifiers that will be used to request (and match) data from other
agencies. Specifically, we will need the appropriate identifier(s) to request
probation/parole data and arrest data. Also, for our juvenile subjects, we will need
identifiers/identifying information in order to request DOC data. When discussing
available identifiers with the agency, it may be helpful to ask what identifiers are used
by the state agencies themselves when merging data across agencies.
o Necessary identifiers include full name, DOB, DOC/DJJ#, state identification
number (if available in the state), and, if the state does not use a state
identification number, SSN. If it is possible to avoid obtaining SSN’s yet still
enable the probation/parole agency and arrest agencies to obtain a match on our
respondent, we should make every attempt to do so. The cases in which this
should be possible are states that use a statewide identification number to link
data for individuals across all state agencies.
o Please see notes in the Data Transmission section (page 9) about special
procedures that are necessary if a transmission contains SSN.
Care needs to be taken in obtaining records to be sure that the correct identifier is
used. For example, some states assign different ‘admit-IDs’ for each separate
admission that an individual offender may have. That means that a single offender may
have multiple different ‘DOC IDs.’ However, in order to obtain criminal history and
subsequent offending beyond readmission to the DOC, we need a single unique
‘person-ID.’ Thus, we will need both the identifier associated with the incarceration of
interest and the unique person identifier (if the state assigns one) or some other unique
person identifying information (e.g. DOB) to locate all records for the individual
subjects. Please see notes in the Data Transmission section (page 9) about special
procedures that may be necessary if a transmission contains SSN.
The issue of identifiers is particularly difficult for juvenile cases. Many states do NOT
assign a unique subject identifier or even create a fingerprint record for a juvenile upon
arrest. Thus, we need to work very closely with the juvenile justice agencies to obtain
sufficient identifying information to be able to obtain the criminal record and subsequent
adult DOC data we need for these juvenile subjects. The challenges in obtaining
administrative data for our juvenile subjects include (1) assuring that we obtain records
for each juvenile admission for each of our juvenile subjects; (2) obtaining identifiers to
be able to track the juvenile into the adult system (DOC, P/P); and (3) obtaining
sufficient information to be able to obtain adult arrest records.
When reviewing the list of data elements with DOC/DJJ agencies, please be sure to
note any elements we need that are not available electronically but could be obtained
by extracting information from hard-copy files. Although we are likely to pursue this only
in economic impact sites (IA, MD, PA, and SC) and possibly juvenile sites (CO, FL, KS,
and SC) and only for certain types of elements (likely programs and services data),
please document what is available on paper in all sites. In order to be useful to us,
these data need to be (1) available for both SVORI and comparison offenders, and (2)
available systematically in a consistent location.

SVORI Administrative Data Protocol—Part 2

Page 6 of 12

January, 2007

Probation/Parole Agencies
•

•

•
•

•

In many cases, Probation and Parole agencies (P/P) are NOT part of the Department
of Correction. (Note that Probation and Parole may also be two separate agencies or
is some cases may be county based). Study participants may be released to parole
and/or to probation supervision—depending upon the state’s sentencing structure and
the offender’s sentence.
We need state-level information from P/P for the following purposes:
o Information on performance on supervision following release for the instant
sentence, including revocation/new offending information (recidivism indicators).
o Information on criminal history (i.e. probation sentences and parole supervision
prior to the instant sentence).
o Information on future offending that is separate from the instant sentence (e.g.,
individual successfully completed parole but is subsequently arrested and
sentenced to probation).
We will use identifiers obtained from the DOC and DJJ to generate lists of individuals
for whom we need data. Thus, our request for P/P data follows that for DOC/DJJ data.
We will begin working with P/P agencies in 2007 as soon as we have received
DOC/DJJ data and know the identifiers we will have for each site. Agencies’ abilities to
compile calendar year data for our requests will vary, but we hope to have 2006
calendar year data for all sites by August of 2007. We would like to obtain P/P data for
all individuals included in the dataset provided to us by the DOC or DJJ agency.
If probation and parole is part of the DOC/DJJ, it may be fairly easy for the agency to
include probation/parole (community supervision) data at the same time they send us
data on incarcerations. If this is the case, please explore the degree to which the
DOC/DJJ is willing and easily able to provide P/P data along with the data request
made in early 2007.
In juvenile sites, we will need information on our juvenile subjects from both the juvenile
parole agency and from the adult P/P agency.
Initial data request for 2006 data: This request should be for data that includes all
historical information for each individual (i.e., every probation or parole term up through
12/31/2006.
Second data request for 2007 data: The agencies can provide either (1) an update that
includes information for each individual for the year 1/1/2007 through 12/31/2007; or (2)
a complete “dump” of records for the individual that would include both the data we
obtained in 2007 and the new data obtained in 2008.

Arrest Data Agencies
•

•

We would like to obtain from each State Arrest Data Agency arrest information (arrest
date, offense date, charge(s), and, if available, conviction and sentence information) for
all individuals included in the dataset provided to us by the DOC or DJJ agency.
We will need identifiers from DOC (& DJJ) to provide to the agency for matching
purposes. It would be helpful to find out from the arrest agency what information we will
need to provide to them in order to request data so that we can be sure to obtain it from
the DOC (or DJJ). (Note that DOC/DJJ personnel may know which identifiers are
needed to obtain arrest data.)
We will need to negotiate data agreements with the agency in each state that maintains
the state’s arrest records.

SVORI Administrative Data Protocol—Part 2

Page 7 of 12

January, 2007

•

We plan to request these data in early 2007 and early 2008. Each request will be for
complete data (i.e., the full set of available arrest data through to the date of the data
extract) because in some sites arrest data entry may be backlogged.

Employment Security Commissions
•
•
•
•

We plan to obtain Employment Security Commission data in selected sites (IA, MD,
PA, and SC—a subset of the Cost-Benefit sites).
We will be able to obtain these data only for adults who signed administrative data
consent forms.
We will request these data only once—in early 2008—for all activity through December
2007.
Because we are collecting ESC data in so few states, members of the management
team will carry out the requests to these agencies.

Data Constructs and Elements
In the companion (“Part 1”) document to be distributed to DOC/DJJ agencies, the section titled
“Which data elements will we request and when will we request them?” outlines the
rationale for obtaining various types of administrative data elements and describes the overall
constructs for which we are seeking data. If the agency you are working with has multiple data
items related to a construct (for example, multiple indicators of ‘sentence length’), it is better to
get all of them. If you have not already done so, we strongly encourage each site lead and
liaison to schedule a conference call with the site’s MIS/research staff to review our request and
the data they collect in order develop a plan for specifying the sampling frame and the specific
data items to be received.
RTI project staff are working on a “meta-data” database where we will be able to store
information about the data (we will develop a separate database to store the actual data).
Therefore, it is very important that you request as much documentation as possible about the
individual data items we will be receiving. We need to know how each data item relates to our
various data constructs, what each item measures, format information, labels for values, data
quality limitations, relationships to other items, etc. We will also use the database to begin
developing common cross-site measures. When asking for documentation, although we would
prefer not to receive the entire data manual for their MIS system if it is possible for them to
provide documentation specific to the individual items we receive, please assure the site that
they should provide whatever is easiest for them.

Data Transmission
Detailed instructions for transmitting data to agencies and receiving data from agencies are
contained in a separate document: “Admin data protocol Part 3_Data Transmission.doc”. Mark
Pope will coordinate the production and transmittal of these files.

Detailed Timeline
Table 4 on the following page outlines the timeline and process for requesting administrative
data from the data sources described above.

SVORI Administrative Data Protocol—Part 2

Page 8 of 12

January, 2007

Table 4. Timeline and tasks
When, Which , Who
– December 2005
– DOC or DJJ
– Site liaisons

– January 2006
– DOC or DJJ
– Mgmt staff
– Feb 2006–Jan 2007
– DOC or DJJ
– Site leads & liaisons

– February 2007
– DOC or DJJ
– Site liaisons
– Feb-March 2007
– DOC or DJJ
– DOC/DJJ staff
– Jan- April 2007
– P/P (if separate from
DOC/DJJ), DOC (in juvenile
sites), & ADA
– Site liaisons

What
Work with site staff to identify who to contact and whether a research agreement or MOU will be needed.
In many cases, the request for administrative data was included in the research agreement with DOC and
DJJ.
Site liaisons confirm those agencies for which MOUs are complete and identify other agencies for which
we will need to execute MOUs (or research agreements) to acquire data.
Letter sent to the identified contacts, explaining the project and requesting a telephone meeting between
the contact and Senior Staff to discuss the project’s data needs.
Administrative site leads conduct telephone call with agency contacts. This call should conclude with an
understanding of:
(1) whether a formal request should be submitted (e.g., letter, MOU, research application);
(2) any data elements that are not available or are available only through manual extraction of paper
records;
(3) the cohorts for which the data will be obtained (see Table 1);
(4) the time ranges of releases for whom data can be obtained;
(5) if a full release cohort is NOT being provided: the identifiers required to access their data (e.g., specific
admission identifiers versus unique person identifiers) including the format in which the identifying data
should be sent to the agencies (e.g. ASCII, excel, SAS; CD or disk);
(6) whether the agency can provide a sample “mini-dump”; and
(7) the site-specific timeline for requesting and receiving the full administrative dataset.
Site liaisons prepare any MOUs or research agreements that will be needed for these data requests.
Site liaisons submit initial request for data (through December 31, 2006) to sites. The exact date to be
determined in negotiations with sites
Agencies send initial administrative data.
If probation/parole data are not obtained as part of DOC or DJJ request:
⋅▪ Work with community corrections agency staff to identify who to contact and whether a research
agreement or MOU will be needed. As we have not had contact with separate P/P, we will need
agreements with these agencies unless DOCs are willing to obtain data on our behalf.
Site liaisons work with site staff to identify who to contact for obtaining (1) community supervision data (if
not obtained through DOC/DJJ request), (2) adult incarceration data (for juvenile sites), and/or (3) state
arrest records.

SVORI Administrative Data Protocol—Part 2

Page 9 of 12

January, 2007

Table 4. Timeline and tasks
When, Which , Who
– January-May 2007
– P/P (if separate from
DOC/DJJ), DOC (in juvenile
sites), & ADA
– Site leads & liaisons

– June 2007
– P/P (if separate from
DOC/DJJ), DOC (in juvenile
sites), and ADA
– Site liaisons
– August 2007
– P/P (if separate from
DOC/DJJ), DOC (in juvenile
sites), and ADA
– Agency staff
– January 2008
– DOC, DJJ, P/P, and ADA
– Site liaisons
– Feb-April 2008
– DOC, DJJ, P/P, and ADA
– Agencies

What
Administrative site leads conduct telephone call with agency contacts. This call should conclude with an
understanding of:
(1) what type of formal request should be submitted (e.g., letter, MOU, research application;
(2) any data elements that are not available or are available only through manual extraction of paper
records;
(3) the cohorts for which the data will be obtained (see Table 1);
(4) the time ranges of releases for whom data can be obtained;
(5) the identifiers required to access their data (e.g., state identification number, FBI number) including the
format in which the identifying data should be sent to the agencies (e.g. ASCII, excel, SAS; CD or disk);
(6) whether the agency can provide a sample “mini-dump”;
(7) the site-specific timeline is for requesting and receiving the full administrative dataset.
Site liaisons work with site staff to identify request protocols and prepare MOUs or research agreements
to obtain access to the data.
Site liaisons submit requests for data (through December 31, 2006). This will be the initial request for P/P
agencies not connected with DOC/DJJ, for DOC agencies in juvenile sites, and for arrest data agencies.

P/P and Law Enforcement Agencies send administrative data. The timing of when agencies will be able to
provide data through the end of the previous calendar year will vary. We would like to have agencies send
data to us as soon as they can—by August at the latest.
Site liaisons submit requests for data (through December 31, 2007). This will be an update request for all
agencies.
Agencies send administrative data.

SVORI Administrative Data Protocol—Part 2

Page 10 of 12

January, 2007

Talking Points
The following “talking points” may be helpful in creating an agenda for phone calls as we
negotiate access to administrative data. We need to determine (and enter in the project
database), for every agency providing administrative data to us, the following pieces of
information:
•

•

•
•

•

Type of offender for whom the agency is willing to provide data, i.e., which of the four
options in Table 1:
1. All program participants from local SVORI program initiation through 12/31/2005
AND all comparison individuals released between 1/1/2004 and 12/31/2005
• this second group includes two subsets: (a) those submitted to us as
eligible respondents PLUS (b) any other individuals who met SVORI
eligibility criteria but who did not participate in SVORI (excluding those who
refused to participate in the SVORI program if known). OR
2. All SVORI participants and comparison subjects submitted to the evaluation as
eligible respondents OR
3. All offenders released in 2004 or 2005 OR
4. All SVORI participants and comparison subjects who completed baseline
interviews and signed administrative data release consent forms (do not offer this
option unsolicited—it is not relevant in most sites and would be potentially
confusing).
If the site is providing data to us based on a list from us (i.e., not giving us the full
release cohort) we need to find out the particular identifiers we need to send them so
that we can uniquely identify the correct individuals, and the format for these identifiers
(e.g., Excel, SPSS, SAS). If providing data for consented offenders only, what do we
need to provide to them to "prove" consent? (Ask this last question ONLY if the agency
is only willing to provide data for consented individuals.)
Are there constraints on the time period for which they are willing to provide data? We
would like to get a full criminal history (i.e., of all prison terms) in response to our initial
request—can they give us all historical and current information (as of the date of the
data extract)? If providing release cohorts, can they give us releases from 2004 and
2005 calendar years?
As you walk through the list of data constructs and elements, are there constructs for
which they have no available data? Note that if they have multiple fields for a specific
construct, discuss with them how the various fields are defined and then get guidance
about which ones are most appropriate. If there is no obvious choice, get all the
possible fields.
Are there fields that are likely to contain a lot of missing or poor-quality data?
Are there any fields we need that are not available electronically but could be obtained
by extracting information from hard-copy files? Note that we are likely to only pursue this
in economic impact sites (IA, MD, PA, and SC) and possibly juvenile sites (CO, FL, KS,
and SC) and only for certain types of elements (likely programs and services data), but
please document what is available on paper in all sites. We can only use data that are
(1) available for both SVORI and comparison offenders, and (2) available systematically
in a consistent location.
What identifiers are stored in their dataset (e.g., DOCNUM, state ID#, individual, case, or
sentence identifiers) and how reliable/complete are they? We may need to combine the

SVORI Administrative Data Protocol—Part 2

Page 11 of 12

January, 2007

•

•
•
•

•

•
•

data we receive from this agency with data from other agencies, and they may use
different identifiers.
What kind of documentation can they provide to help us understand the data fields they
are providing to us? At a minimum, we need names for each field; labels for codes,
descriptions of what’s in the field, etc.
What do we need to do as part of the formal request? Memo, MOU, research
application? Who does the request go to? Any details on what we need to do or include
should be entered in the database.
When can we submit the formal request? Some sites might want us to wait until after a
certain time.
What format will the data be in (e.g., SAS, SPSS, Excel or Access database, text file)?
What will the file structure of the data be? Will it be a flat file? What will be the unit of
analysis (a person? an admission?) Or will it be a relational database (where there are
separate ‘files’ for content areas such as demographics or charge information.) If it is
relational, what will the keys to link across subfiles be?
Will the site be able to create separate files—one containing only identifiers and a row
ID# and the other(s) containing the data elements and the row ID# (but not the
identifiers)?
Will the site be able to Fed-ex the files to us in two separate shipments on two separate
days? (We will pay for the shipping charges.)
Once they receive our formal request, how long will it take at their end until we can we
expect to get the full set of cases?

SVORI Administrative Data Protocol—Part 2

Page 12 of 12

January, 2007

SVORI Multi-site Evaluation Administrative Data Collection Protocol
Part 3: Data Transfer Procedures

Transmissions received from agencies
RTI and UI evaluation staff will work with the relevant agencies on the format and transmission
of the data. The data must be transferred to RTI in a secure manner, and Mark Pope will be
coordinating all data transfer arrangements. Essentially, evaluation staff working with the
agencies will ask the agencies to generate two sets of electronic data files:
1. One data file should contain only the identifiers (without SSN) and a special unique ID#
(e.g., an auto or sequential number) generated by the agency for the purpose of transfer
of these data;
2. The second data file should contain only the requested data elements and the special
unique ID#.
If SSN must be obtained from the agency (see page 6 of the internal “Part 2” protocol), please
consult with Mark Pope to work through additional steps required (likely the generation of a third
file containing only the special ID# generated by the agency and the SSN).
The password-protected files can be transferred in one of two ways: via a secure website
(preferred) or in separate Fed-Ex submissions.

Transmission via website
Data files will be transferred from the local SVORI sites to RTI using the following steps:
1. A page located on the SVORI public website (http://www.svori-evaluation.org/admin) will
be used by the various state agencies to upload administrative data files. This page will
also contain text fields where the site can enter its state, the organization that is
providing the data, and a description of the file being uploaded. This site uses Secure
Sockets Layer which means that the uploaded file and text information are encrypted
during the transfer. This page will not be "advertised" on the SVORI public website and
only those individuals who are providing data to us will be given the URL. The site will
also be hidden from any webcrawlers or bots that might try to index it.
2. Upon uploading, the file will be saved to a secure directory on the RTI webserver that is
not accessible from the web except by using the upload function on the above
mentioned page (i.e., anyone trying to type the directory structure directly into their
browser cannot get to this directory).
3. After the file has been uploaded, the information entered into the text fields above and
the file name are logged to a table in the SVORI database and a notification e-mail is
sent to Mark Pope alerting him that a new file has been uploaded. He will then remove
the uploaded file from the webserver and transfer it to the SVORI share drive on RTI's
private network.
The files should be uploaded in two (or three, depending on whether SSN will be transmitted)
separate transmissions. When the agency uploads files to the website, a short period of time
should be allowed to elapse between transferring files containing data and files containing
identifiers. After uploading one file, the webpage automatically reports that the file uploaded
successfully and then gives the user the option of uploading another file.

Transmissions via FedEx
The password-protected files should be copied onto media such as a CD or flash drive (which
will be provided by RTI on request) and sent to RTI via Fed-Ex (using pre-paid labels provided
by RTI).
The files should be sent in two (or three, depending on whether SSN will be transmitted)
separate transmissions. The files containing the data should be sent one day and the file
containing the ID#s should be sent the next day (the file containing SSN should be sent on a
third day).

Post-transmission Procedures
Once received by RTI, the data files will be merged. The identifiers will be retained only in the
master files for the project. The analysis files used by project staff will not contain any
identifying information; individual cases will be identified only by a study-assigned ID. The
linked files will be destroyed at the end of the study, unless an IRB amendment is approved to
extend the period of storage of these data.

Transmissions sent to agencies
If our request to an agency involves our staff submitting a list of identifiers to their staff, the
transmission procedures differ depending on the content of the list and the preference of the
agency. Mark Pope will coordinate the production and transmittal of these files that contain
identifiers and will ensure RTI/UI staff have the software they need to properly encrypt files.
Files provided to the sites from RTI/UI will be encrypted prior to transfer using PGP to create a
self-decrypting archive. A self-decrypting archive encrypts the file but allows it to be decrypted
by someone who does not have PGP software on their machine, provided that the recipient has
the passphrase that was used to create the self-decrypting archive. The encrypted file will be
placed on the regular project FTP site developed for the project (used to transmit release lists
and confirmation files), unless the agency specifically requests that the file (also encrypted) be
submitted on CD/diskette (via FedEx). The passphrase will be given to the agency contact
person via telephone to ensure security.