Evaluating the Outcomes of a Prison-Based Domestic Violence Treatment Program

Background and Purpose: Between 30-60% of incarcerated men report using domestic violence in their intimate relationships. Many community-based domestic violence treatment programs have targeted attitudes associated with domestic violence as outcomes with varied success, while prisons across the country have begun to implement correctional-based treatment programs as a means to reduce recidivism rates. The purpose of this study is to evaluate a correctional-based domestic violence treatment program by exploring benefits to attitudes and recidivism for program participants. Our research questions were: 1) Does participation in a prison-based domestic violence treatment program improve attitudes toward women and lower levels of criminal thinking? and 2) Does participation in a prison-based domestic violence treatment program reduce recidivism rates?

Methods: Approximately 600 offenders over 10 years participated in a manualized program called STOP. STOP is a CBT program that focuses on culturally-sanctioned power of men over women and anger management. Data were provided by prison administrators from the North Carolina Department of Corrections, including program participants and 200,000 inmates released over the ten-year period. A sample of 176 participants was used to assess attitude changes, and a sample of 253 STOP participants was used to assess STOP’s influence on recidivism. In regards to recidivism, we used propensity score matching (PSM) to create a matched comparison group based on nine covariates associated with recidivism and program selection. A statistically similar match was identified for all program participants, yielding a final recidivism sample of 506. PSM enhances confidence in the validity of findings when a true-experimental design is not feasible, as it helps control for selection bias. Attitude changes were assessed using the Attitudes Towards Women Scale and the Texas Christian University Criminal Thinking Scale. Participants were considered recidivists if they were re-incarcerated in a North Carolina prison during the five and seven years post-release.

 Results: Program participants had improved scores on the ATW (t=-7.70, p<.001, 95% Confidence Interval (CI) [-6.24, -3.72]) and all six subscales of the TCU Criminal Thinking Scale. The mean score on ATW increased, indicating more positive attitudes towards women. The mean scores on all six subscales of the TCU scale decreased, indicating lower criminal thinking.  Using binary logistic regressions, we found that program participation was likely to yield lower recidivism rates, as indicated by the 95% CIs. For the five-year recidivism rate, the odds ratio point estimate was .624 (p = .092, 95% CI [.360, 1.080]), while the seven-year odds ratio point estimate was .611 (p = .160, 95% CI [.307, 1.215]). While not statistically significant, confidence intervals indicate that STOP would lead to lower recidivism than a match control group in the majority of samples, particularly for 5-year recidivism rates. 

Conclusion and Implications: These results suggest that the STOP program yields positive outcomes. Improved attitudes related to domestic violence can lead to lower perpetration rates, leading to fewer victims, while lower recidivism rates are beneficial for prisons, states, and the federal government in terms of monetary costs and gains in societal safety.