The Partner Abuse State of Knowledge Project Manuscripts and Online Data Base Overview of Findings by the Authors

# 11  The Crime Control Effects of Criminal Sanctions for Intimate Partner Violence
Christopher Maxwell and Joe Garner
72 pages.  Full manuscript available in Partner Abuse Vol. 3, Issue no. 4 (2012), pp. 469-500.

This research provides a systematic review of published studies that provide evidence regarding the crime control benefits from prosecution, conviction, and sentencing of IPV offenders, assesses the nature and quality of those studies, and summarizes the reported findings about the crime control effects of criminal sanctions.  Unfortunately, the research on the effectiveness of criminal sanctions for IPV does not address the underlying causal mechanisms and typically does not collect information which could permit distinguishing the effects of fear of sanctions from victim empowerment or other potential mechanisms by which sanctions might affect future behavior.  For this reason, we specify a generic crime control effect that is neutral about the mechanism by which sanctions affect future behavior.  In addition, our specification considers that all types of sanctions may not have an effect that is consistent in size or direction.  Therefore, we identify three hypotheses about the effects of criminal sanctions on IPV.  We label these the prosecution hypothesis, the conviction hypothesis and the sanction severity hypothesis.  This approach permits distinctions among three policy choices in criminal justice processing–the decision to charges, the decision to convict, and the type of sanction imposed upon convicted offenders. 

Among the 135 publications used to estimate the amount of prosecution and conviction by Garner and Maxwell (2009), we have identified 30 that assess the relationship between the application of sanctions by the criminal courts and repeat offending.  This review describes the characteristics of each of the studies, summarizes the substantive findings reported and evaluates the research designs, measures and methods used.  These 30 studies generated 143 statistical tests that inform one of these three crime prevention hypotheses.  For each study and by each hypothesis, we present the number of reported statistical tests that show that criminal sanctions were significantly associated with less repeat offending, more repeat offending, or do not significantly affect repeat offending.

Based upon the analyses and conclusions produced by these studies, we find that the most frequent outcome reported is that sanctions that follow an arrest for IPV have no effect on the prevalence of subsequent offending.  This finding holds for the prosecution, conviction and sanction severity hypotheses.  However, among the minority of reported analyses that do report a statistically significant effect, two-thirds of the published findings show sanctions are associated with reductions in repeat offending and one third show sanctions are associated with increased repeat offending. 

Our examination of the methods used by these studies identified seven common issues that suggest that, regardless of the substantive findings reported, the research designs used in these 30 reports are inadequate to assess the relevant public policies and criminological theories.  Based upon our systematic assessment of the published studies, we conclude that the preponderance of the reported findings show no effect for criminal sanctions; however, the quality of the methods used in this body of research provides us with an insufficient basis to support a conclusion about the use of criminal prosecution and sentencing for IPV.

To address this gap, we recommend that the field undertake a well-funded, long-term program of research that will rigorously test these and other crime control hypotheses.  This program must articulate the causal mechanisms under review, model when the effects begin and dissipate, use reliable and common outcome and sanction measures, distinguish selection effects from treatment effects, incorporate rigorous multivariate analyses, and meet contemporary standards for statistical power.  The National Institute of Justice-sponsored Spouse Assault Replication Program is an example of such an approach.  While imperfect, this program provided the rigorous, systematic evidence necessary to assess the crime control effects of arrest for IPV.   We also recommend using existing data from these 30 studies to improve our understanding of this body of research and of the crime control effects of criminal sanctions through secondary data analyses.