The Serious and Violent Offender Reentry Initiative (SVORI) funded agencies to develop programs to improve criminal justice, employment, education, health, and housing outcomes for released prisoners. SVORI was a goal-oriented initiative that specified outcomes that should be achieved by programs that were developed locally. The purpose of the SVORI multi-site evaluation was to determine whether the selected programs accomplished the overall goal of the Reentry Initiative -- increasing public safety by reducing recidivism among the populations served by the program -- and determine the relative costs and benefits of the program. The SVORI multi-site evaluation included an impact evaluation to determine the effectiveness of programming. Specifically, the purpose of the impact evaluation was to determine whether individuals who participated in enhanced reentry programming, as measured by their enrollment in SVORI programs, had improved post-release outcomes than comparable individuals who did not participate in SVORI programming.
Under NIJ Grant 2004-RE-CX-0002, the principal investigators conducted an impact evaluation of SVORI. Impact evaluation data collection for both SVORI and non-SVORI participants consisted of four waves of in-person, computer-assisted interviews and oral swab drug tests conducted in conjunction with follow-up interviews. The research team collected data on a total of 2,391 individuals including 1,697 adult males (Part 1), 357 adult females (Part 2), and 337 juvenile males (Part 3).
As part of the impact evaluation, experienced RTI field interviewers conducted pre-release interviews with offenders approximately 30 days before release from prison and a series of follow-up interviews at 3, 9, and 15 months post-release. All interviews were conducted in private settings 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. The pre-release interviews obtained 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 to the community. Post-release interviews were conducted from December 2004 through May 2007. Interviews were conducted in the community, and in jails or prisons for those who were re-incarcerated. The post-release interviews were similar in content across waves and obtained data on reentry experiences, housing, employment, family and community integration, substance abuse, physical and mental health, supervision and criminal history, service needs, and service receipt. The interview instruments were developed through an extensive 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.
Under NIJ Grant 2009-IJ-CX-0010, follow up interview data was added for males, females, and juveniles from 11 of the 12 original SVORI sites.
In developing criteria for site selection for the Serious and Violent Offender Reentry Initiative (SVORI) impact evaluation, the principal investigators focused on identifying factors that would provide the best assurance that a program would be evaluable. Six criteria were identified to guide site selection:
- Program had clearly defined elements and goals.
- Program was implemented (or was likely to be implemented).
- Program target population was accessible and of sufficient size.
- Appropriate comparison population was available and accessible for inclusion in the study.
- Administrative data were of good quality and available for the evaluation.
- Program was 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
- Review of SVORI grantee proposals and work plans and follow-up. Telephone interviews with program directors to obtain information not gleaned from the review, clarifications and updates on the programs' status.
- Visits to the sites of a selected subset of programs.
- Review of all information to develop a list of recommended programs for inclusion in the impact evaluation that was submitted to NIJ for approval.
Based upon these criteria and procedures, a total of 16 out of all 89 SVORI programs were included in the impact evaluation, comprising 12 adult programs and 4 juvenile programs located in 14 states (adult only unless otherwise 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. 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, quasi-experimental 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.
From these 16 programs, a total of 4,354 cases were fielded for inclusion in SVORI impact evaluation study. A total of 1,963 cases were dropped from the sample including 718 cases that were released before interviews could be scheduled, 635 cases that were ineligible for the evaluation, 370 refusals, 192 cases were dropped because the respondents were not released while the first post-release interview was being conducted, and 48 other non-interviews. Thus, the final sample of evaluation-eligible respondents for the impact evaluation was comprised of 2,391 individuals -- 1,697 adult males (Part 1), 357 adult females (Part 2), and 337 juvenile males (Part 3). Specifically, the final sample included 863 SVORI and 834 non-SVORI adult males, 153 SVORI and 204 non-SVORI adult females, and 152 SVORI and 185 non-SVORI juvenile males.
Longitudinal: Cohort / Event-based
Adult male offenders who received Serious and Violent Offender Reentry Initiative (SVORI) programming or were selected as comparison subjects in 12 states between July 2004 and November 2005 (Part 1). Adult female offenders who received Serious and Violent Offender Reentry Initiative (SVORI) programming or were selected as comparison subjects in 11 states between July 2004 and November 2005 (Part 2). Juvenile male offenders who received Serious and Violent Offender Reentry Initiative (SVORI) programming or were selected as comparison subjects in 4 states between July 2004 and November 2005 (Part 3).
Four waves of in-person, computer-assisted interviews with SVORI program participants and comparison subjects
Laboratory results data from the oral swab drug tests that were conducted during the 3- and 15-month interviews for respondents who were interviewed in a community setting
computer-assisted personal interview (CAPI)
The Adult Males Data (Part 1), Adult Females Data (Part 2), and the Juvenile Males Data (Part 3) each contain the same 5,566 variables from the original Multi-site Evaluation of SVORI including variables from the 4 waves of offender interviews, 10 drug test lab results variables and 3 weight variables. (Note: Some interview questions were only asked of adults, and other questions were only asked of juveniles.) Offender interview variables include demographics, housing, employment, education, military experience, family background, peer relationships, program operations and services, physical and mental health, substance abuse, crime and delinquency, and attitudes.
From the offender interviews, data include:
- Demographic variables: age, facility type, gender, race, acculturation, and duration of incarceration.
- Housing variables: location, type of housing, duration of housing, housing expectation, composition of household, contribution to housing costs, owner/tenant status, housing stability, barriers to housing, living with criminally-/drug-involved people, and neighborhood quality.
- Employment variables: ever had job, sources of support/employment status, job stability/reasons for not working, unemployment insurance, most recent job information, lifetime employment duration/termination, expectation to return to previous job, barriers to employment, job satisfaction, and job stress.
- Education variables: educational attainment, school attendance/stability, and school suspension/expulsion.
- Military experience variables: ever served, type of discharge, and currently serving.
- Family background variables: marital/partner status, parental status, children/primary care responsibilities, child support, family affiliation, family criminal history, parent/guardian information, family emotional support, family instrumental support, in-prison contact, parental relationship, victimization, perpetration of violence, quality of intimate partnership, child custody and visitation, and relationship with children.
- Peer relationships variables: peer criminal behavior and peer instrumental support.
- Program operations and services variables: assessment and case management, service need, release planning, services received (child support/child care, juvenile services, identification/life skills/attitudes, parenting/domestic violence/mentoring/anger management, education/transportation/housing/accessing resources, employment services), medical/dental care, mental health care, substance abuse treatment, most/least helpful services, work release, and experiences in the first 24 hours after release.
- Physical and mental health variables: physical health status, physical health-related limitations, physical health diagnoses and medications, physical health hospitalizations, vision/hearing/dental health, mental health status, mental health-related limitations, mental health symptoms, PTSD symptoms, mental health treatment, mental health medications, and mental health hospitalizations.
- Substance abuse variables: use of alcohol, use of sedatives, use of tranquilizers, use of stimulants, use of pain relievers or opiates, use of methadone, use of anabolic steroids, use of marijuana, use of hallucinogens, use of cocaine, use of amphetamines, and use of inhalants.
- Crime and delinquency variables: criminal history, gang membership, court appearances, supervision status and officer contacts, supervision conditions and violations, sanctions and rewards, attitudes toward parole officer, perceptions of factors related to recidivism, and perceptions of factors related to desistance.
- Attitude variables: self-efficacy, locus of control, readiness for change, spirituality, legal cynicism, substance abuse treatment motivation, and civic action.
The 10 drug test lab results variables include results for amphetamines, cannabinoids, cocaine, opiates, and phencyclidine for the oral swab drug tests that were conducted during the 3-month (Wave 2) and 15-month (Wave 4) interviews.
Weight variables include individual probability, population average treatment effects (PATE) weight, and average treatment effect on the treated (ATET) weight.
Additional variables, on the same topics as the original data collection, were added to each part of the study via follow-up interviews resulting in a final variable count of 5,751 for each part.
Of the 4,354 cases fielded for inclusion in the original Serious and Violent Offender Reentry Initiative (SVORI) impact evaluation study, a total of 2,391 respondents completed the Wave 1 (30 days pre-release) interviews, yielding a response rate of 54.9 percent, which is based on the total number of cases fielded, including both eligible and ineligible cases.
A total of 1,464 respondents completed the Wave 2 (3-month post-release) interviews, yielding a response rate of 61.2 percent.
A total of 1,527 respondents completed the Wave 3 (9-month post-release) interviews, yielding a response rate of 63.9 percent.
A total of 1,637 respondents completed the Wave 4 (15-month post-release) interviews, yielding a response rate of 68.5 percent.
The SF-12 Health Survey was used to measure respondents' physical and mental functioning, and the SA-45 (Global Severity Index and Brief Symptom Inventory) and the Positive Symptom Total index were used to measure respondents' mental health. The SA-45 includes subscales indicating symptoms of specific psychopathologies including anxiety, depression, hostility, interpersonal sensitivity, obsessive-compulsive disorder, paranoid ideation, phobic anxiety, psychoticism, and somatization. Several Likert-type scales were also used.