Predicting recidivism following participation in a treatment program for batterers

This study examined recidivism rates for 197 batterers who participated in a state-sponsored domestic violence treatment program. The primary objective of the study was to identify factors associated with liost-intervention recidivism. Significant factors identified through bivariate analyses were analyzed through logistic regression for the purpose of developing predictive models. Bivariate analyses identified 10 factors associated with post-intervention recidivism. Of these factors, logistic regression identified four factors that were predictive of recidivism. These four factors-psychopathology (personality disorders), psychiatric history, substance abuse, and child abuse in family of origin were able to 'predict 84% of all outcomes and 97% and 28% of abstainers and reoffenders, respectively. The data suggest that batterer characteristics, particularly batterer pathology and substance abuse, are more influential determinants of recidivism than systemic and programmatic factors. These findings are discussed with respect to their implications for policy, practice, and future research.