Gauging the effectiveness of jail reentry programs

Millions of individuals enter and leave the nation's 3,600 jails each year; this study examines the effectiveness of a reentry program developed by a county correctional department to more effectively prepare inmates for release. The empirical study relies on administrative data and uses econometric methods to answer three research questions: first, whether program participants enrolled in more treatment programming; second, whether program participants incurred fewer disciplinary sanctions during the course of their incarceration; and third, whether program participants recidivated--as defined by a court arraignment for a new offense or a probation/parole violation within one year of release--at reduced rates than if they had not participated in the program. One goal of the paper was to explore the use of other outcome measures to determine program effectiveness besides reduced recidivism rates. The study exploits a natural experiment and employs a difference-in-differences estimator to determine program effectiveness. The author worked with the Department's Director of Research to compile a unique database on 2,105 individuals sentenced to the correctional facility during an eleventh month period prior and after the implementation of the program. The administrative records included extensive criminal history, socio-demographic, treatment programming, institutional movement, and post-release recidivism data. The study uses survival analysis and hazard modeling to explore whether the program had an effect on both the incidence and timing of recidivism.

The study found that the program did result in measurable increases in the rate of program participation by as much as 67%. Further, the study lent credence to the Department's staffs belief that that the Transitional Program motivated and gave incentive to individuals to seek more treatment programming, and also improved the institution's efficiency and timeliness of enrolling individuals into evidenced-based treatment programming better matched to the individual's assessed risks and needs. On the outcomes of disciplinary infractions and recidivism rates, the analyses did not find evidence of significant program effects. In exploring reasons for the null finding, the study did find that parole release served to increase recidivism significantly, but could not provide supporting evidence that increased rates of parole for program participants could have confounded a program effect to reduce recidivism rates.