Rethinking Rehabilitation in Prisoner Reentry Programs: Capturing the Chaos of Desistance

Purpose: This paper presentation reports an evaluation of a statewide prisoner reentry program that found mixed results relative to an equivalent comparison group and measured by positive urinalysis, new convictions, and returns to prison. Similar ambiguous findings for other reentry programs have shown these programs to have largely failed at rehabilitation, at least measured against traditional recidivism outcomes. However, based on the current results, it is argued that a recalibration in how rehabilitation is measured is needed, placing less emphasis on recidivism and more on capturing process and intermediate outcomes that likely build success during the non-linear and chaotic path toward desistance.   

Methods: Propensity scores, based on an estimated logistic regression model, and caliper matching identified a comparison group for 537 reentry participants using age, race, gender, and Level of Service Inventory-Revised (LSI-R) composite scores as matching variables. Utilizing these matched samples, descriptive comparisons (i.e., chi-square test, t-test, ANOVA) were completed. Then to examine differences in positive urinalysis, new convictions, and returns to prison, reentry participants who released prior to the end of the study timeframe (N=358) and their matched comparisons were analyzed using survival analysis, logistic regression analysis, and best subsets logistic regression analysis.

Results: In descriptive comparisons, reentry participants served longer on the index incarceration and acquired a significantly greater number of positive urinalyses than comparison subjects. Survival analysis revealed that reentry participants were at a significantly increased hazard for return to prison, but an apparent group difference in survival curves was not evident over the entire study period. In contrast, logistic regression findings indicated that reentry participants had significantly lower odds of incurring a new conviction when compared to subjects in the comparison group.  Best subsets logistic regression found no significant differences in recidivism between the groups.

Conclusions and Implications: Study results suggest a mixed picture in terms of recidivism outcomes on whether the reentry program succeeded at rehabilitation. In order to achieve a more holistic picture on how reentry participants fared, the rehabilitation definition needs to expand beyond recidivism. As described in the literature, desistance is often a nonlinear and chaotic process, which cycles between criminal and non-criminal behavior (Laub & Sampson, 2001). Public health echoes this conceptualization of desistance, emphasizing that changes in human behavior do not occur in a linear, rational pattern (Resnicow & Page, 2008). However, reentry programs have largely been based on the model that improvements in intermediate outcomes will lead to linearly matched improvements in recidivism (Lattimore et al., 2010). Reentry programs seen through the lens of this linear model centered on recidivism may be failing to capture the process that actually leads to desistance. Social workers are well positioned to capture the chaotic progress towards desistance by reframing and expanding the concept of rehabilitation to match what behavioral change actually looks like. Then, all benefits of reentry can be articulated beyond simply recidivism, which is often as much a product of structural influences as personal decision-making.