1 year, a finding that the evaluators attribute to a combination of implementation difficulties, program design, and a mismatch between participant needs and program content. In response to the evaluation report, Marlowe (2006) argues that the evidence base for the program was flawed from the beginning, with weak designs and unproven, unstandardized interventions. Edward Rhine of the Ohio Department of Rehabilitation and Correction-and colleagues (2006) are more optimistic about reentry programming in general, and point out that Project Greenlight was not notably different from other failed reentry programs and that the treatment was not delivered appropriately (see also Wilson and Davis, 2006; Visher, 2006; Marlowe, 2006).

This review of programs and services for former prisoners suggests three main conclusions regarding their effectiveness in reducing recidivism and problem behaviors. First, there is scientific evidence that several programs and approaches reduce violations of community supervision requirements, arrests for new crimes, and drug use. They include cognitive-behavioral therapeutic approaches and frequent testing for drug use, coupled with treatment. Mentoring programs and comprehensive multiservice employment initiatives show promise but require further, more rigorous research.

Second, inadequate implementation of program principles and procedures appears to be a significant obstacle in the way of being able to determine program effectiveness or finding out whether a program might have benefits for participants.

Third, a major limitation of current program evaluation results is the failure to account fully for self-selection bias. Random assignment of persons to treatment and control conditions remains rare in research on the reentry process. Greater use of experimental designs, when such designs are feasible, is essential for drawing valid conclusions about reentry program effectiveness. When such designs are not feasible, greater attention should be paid to the selection of comparison groups and statistical adjustments for existing differences between program participants and nonparticipants. Although the field has moved beyond “nothing works” in assessments of program effects on reentry outcomes, it can identify with high confidence only a very few best practices for reducing recidivism and enhancing desistance among people leaving prison to return to local communities. More research, especially more experimental research, is needed to identify interventions that could significantly improve outcomes of community supervision for parolees (see Chapter 6).