Association for Public Policy Analysis & Management

The U.S. Department of Labor contracted with Social Policy Research Associates (SPR) and its subcontractors MDRC and National Opinion Research Center to conduct a multi-year impact evaluation of the Reintegration of Ex-Offenders (RExO). The purpose of the evaluation was to examine the impact of comprehensive employment-centered services provided by 24 community-based grantees across the country to formerly incarcerated individuals participating in RExO through a rigorous, random assignment evaluation. A total of 4,655 individuals was randomly assigned to either a program group (n=2,804) or control group (n=1,851), and the evaluation examined RExO’s impacts on participants’ post-program labor market outcomes and criminal recidivism by comparing the outcomes of these two groups. Each member of the research sample was tracked for at least three years after random assignment.

The programs funded under RExO primarily provided three main types of services: mentoring, which most often took the form of group mentoring, but also included one-on-one mentoring and other activities; employment services, including work readiness training, job training, job placement, job clubs, transitional employment, and post-placement follow-up; and case management and supportive services. These services tended to be relatively short in duration (at most a few weeks) and varied substantially across grantees.

Follow-up data collection included a telephone-based survey at both two and three years following random assignment, qualitative data collected through two rounds of site visits to each of the 24 grantees, and administrative data on recidivism (collected through state agencies in each of the 18 states in which RExO operated) and employment and earnings (collected through the National Directory of New Hires).

Results from the impact analysis found that the program group was significantly more likely to have received an array of services, but did not find any substantial impacts on recidivism, employment, or earnings. Possible reasons for the lack of impacts include the fact that the programs tended to be relatively short in duration, and perhaps did not address the full range of needs of study participants. Additionally, there was substantial variation in how RExO was implemented across the grantees, and the fact that the study was therefore not a test of any single model may have contributed to the lack of impacts. Further, the screening conducted by grantees prior to enrolling participants into the study ensured that the study sample was not reflective of the average population returning from state and federal prisons or local jails, and this fact also may have affected the study’s ability to find impacts.

The presenter, a Vice President at Social Policy Research Associates, was the Project Director for the Multi-Site Impact Evaluation of RExO.