Association for Public Policy Analysis & Management

For some time, researchers have been intrigued by the effect of pro-arrest policies for domestic violence crimes. The policies give guidance or instruction officers who are responding to domestic violence crime scenes on how to respond once a primary aggressor has been determined. Officers may use their discretion in making a warrantless arrest, are strongly encourage to make a warrantless arrest, or an officer is mandated to make a warrantless arrest. Berk and Sherman conducted six randomized controlled trials in the 1980s, but the results were inconclusive. Half of the randomized trials showed stricter arrest laws reduced subsequent domestic violence, while half of the trails showed stricter arrest laws had no effect and may have even increased domestic violence. Since these early experiments, a series of quasi-experimental designs have emerged in several literatures – though mainly crime and economics. As before, results are inconclusive.

This paper seeks to serve as a sort of meta-analysis of past research, but will also show that this question is difficult to answer with current methods. In reality, both of the previously employed methods leave policymakers with incomplete information, which is largely due to issues with measurement in both the policy and outcome variables. For example, a jurisdiction may have a variety of pro-arrest laws, which may interact in unknown ways, and they may have adopted more than one pro-arrest policy in their history (for example, a state may start with a discretionary law and then move to a mandatory law). This nuance has heretofore, not been explored. However, given the volatile and serious nature of domestic violence, this subtlety is very important for policymakers seeking to adopt policies regarding arrest.