U.S. flag

An official website of the United States government, Department of Justice.

Do Batterer Programs Reduce Recidivism? Results From a Randomized Trial in the Bronx

NCJ Number
Justice Quarterly Volume: 25 Issue: 2 Dated: June 2008 Pages: 252-282
Date Published
June 2008
31 pages

This study discussed the availability of a definitive test for whether batterer programs reduced recidivism.


The study found that neither the batterer program nor either of two monitoring schedules produced a reduction in official re-arrest rates for any offense, for domestic violence, or for domestic violence with the same victim. The present study had sought to provide greater clarification over three of the four previous randomized trials which produced largely negative results. Over the past two decades, a growing number of courts have come to rely on batterer programs as their mandate of choice, especially when the legal issues in a case preclude the imposition of jail. The study randomly assigned misdemeanor domestic violence offenders in the Bronx, New York to either a batterer program or not; and to either monthly or graduated judicial monitoring, with the latter involving reduced court appearances in response to compliance and increased appearances in response to noncompliance. Similarly, 1-year victim interviews indicated that neither program assignment nor monitoring schedule significantly affected victim reports of re-abuse, either in general or with regard to specific forms of re-abuse: physical, threats, or other. While victims expressed greater satisfaction with the sentence when a batterer program was assigned, interpretation of this last finding is difficult in the absence of an effect on re-abuse. The study notes that the preponderance of evidence accumulated in the field calls into serious question the effectiveness of batterer programs based on the most prevalent national models. The randomized trial used in this study involved a two-by-two factorial design. Convicted male domestic violence offenders were randomly assigned to one of four experimental conditions. Study enrollment took place for a 19-month period from 2002 to 2004, with a total of 420 offenders subject to randomization. Tables, references

Date Published: June 1, 2008