Are Interventions with Batterers Effective? A Meta-analytical Review

The inconsistency in the results both internally and between of previous meta-analyses on batterer intervention program efficacy, and the publication of new batterer interventions underscored the need for an up-to-date meta-analyticalreview. A total of 25 primary studies were found from literature search, obtaining 62 effect sizes, and a total sample of 20,860 intervened batterers. The results of a global meta-analysis showed a positive, significant, and of a mediummagnitude effect size for batterer interventions, but not generalizable. Nevertheless, the results exhibited a significantly higher rate of recidivism measured in couple reports (CRs) than in official records (ORs). As a consequence, intervention efficacy measuring in CRs was null, whilst in ORs was positive and significant. As for the intervention model, positive andsignificant effects were observed under the Duluth Model and cognitive-behavioural treatment programs (CBTPs), but a higher effect size was obtained with CBTPs in comparison to the Duluth Model (under this model, interventions may have negative effects, i.e., an increase in recidivism rate). In relation to intervention length, short interventions failed to reduce recidivism in ORs and may have negative effects, while long interventions were effective in reducing recidivism rate in ORs without negative effects. Efficacy evaluations in short follow-ups were invalid as artificially boosted recidivismreduction rate. Limitations of ORs and short follow-ups as measures of the intervention efficacy and implications of results for batterer intervention are discussed.