Rehabilitation

As noted previously, Martinson's 1974 review of the research on correctional programs poignantly raised the question of whether correctional interventions "work." Harboring negative sentiments toward rehabilitation, many policymakers and criminologists embraced Martinson's critique of rehabilitation and embraced his conclusion that "nothing works" on corrections. Now that more than a quarter of a century has passed, we can provide a more balanced assessment of the question posed by Martinson. The evidence tells us that the heady optimism of early reformers was not warranted, but neither, it appears, is the pessimism of current-day critics of rehabilitation. Simply put, we know more now about what does, and does not, work to reduce offender recidivism (for a summary, see Cullen and Gendreau).

In recent years, correctional interventions have, at times, become more punitive and have sought to achieve recidivism by deterring offenders rather than by changing them. These intervention strategies, for example, have involved the intensive supervision of probationers and parolees, the electronic monitoring of offenders in the community, boot camps for those beginning a life in crime, and "scared straight" programs for juveniles. After considerable research, the evidence is clear: these deterrence-oriented programs do not work to reduce recidivism (Cullen and Gendreau; Cullen et al., 1996; Petersilia and Turner).

In contrast, it is now apparent that rehabilitation programs generally reduce recidivism and, when conducted according to the "principles of effective treatment" (Gendreau), cut reoffending substantially (Andrews and Bonta; Cullen and Gendreau; Lipsey and Wilson; Lurigio). Evidence favorable to rehabilitation has been generated by a statistical technique called meta-analysis. Traditionally, criminologists such as Martinson would read over a group of studies evaluating treatment programs. They would then either describe what the studies found—a narrative review—or try to count how many studies showed that offender treatment worked or did not work—the "ballot box" method. A meta-analysis, however, essentially computes a batting average across all studies, calculating the average impact of treatment on recidivism. Using this method, the existing research, which now involves hundreds of evaluation studies, shows that rehabilitation programs reduce recidivism about 10 percentage points. Thus, if a control group had a recidivism rate of 55 percent, the treatment group's rate of re-offending would be 45 percent.

A group of Canadian psychologists interested in crime—Don Andrews, James Bonta, and Paul Gendreau being its most prominent members—have taken the analysis of effective rehabilitation one step farther. They had two important insights. First, they believed that treatment should focus on changing those factors that are most strongly associated with or "predict" recidivism (e.g., antisocial values and peer associations, low self-control). Second, they hypothesized that rehabilitation programs that "worked" to reduce recidivism should share common features. Thus, it made sense to investigate what distinguished programs that decreased re-offending from those that did not.

Based on meta-analyses of treatment studies, they found that in rehabilitation programs that conformed to the principles of effective intervention, recidivism was about 25 percentage points lower in the treatment as opposed to the control group (Andrews and Bonta; Cullen and Gendreau). These principles include: (1) target the known predictors of recidivism for change; (2) use cognitive-behavioral treatments that reinforce prosocial attitudes and behavior, seek to challenge and extinguish criminal thinking patterns, and provide alternative, prosocial ways of acting; (3) focus treatment interventions on high-risk offenders; (4) try to take into account characteristics of offenders (e.g., I.Q.) that might affect their responsivity to treatment; (5) employ staff that are well trained and interpersonally sensitive; and (6) provide offenders with aftercare once they leave the program (Gendreau).