Published by Lawrence Sherman and Richard Berk in 1984, the Minneapolis Domestic Violence Experiment was the first to attempt to assign police responses randomly after domestic violence incidents. The findings suggested that arrest did reduce recidivism (the relapse into abusive and/or criminal behavior), and the findings were widely distributed. The researchers found, according to victims’ and official reports, that arrest for misdemeanor domestic violence was significantly more effective than other police actions in reducing repeated violence during a 6-month follow-up period. The study received a great deal of attention and seemed to influence public policies.
Background
From the beginning of the modern battered women’s movement in the 1970s through the mid-1980s, most supporters of battered women emphasized the lack of police responsiveness to their needs and advocated for more active police intervention. Newspapers often publicized incidents in which battered women were unprotected after calling the police. In most jurisdictions prior to the mid- to late 1980s, mediation and advice were standard police responses to domestic violence, and arrest was rare. When police officers intervened, they usually talked to the bat-terer, urged him to walk around the block to calm down, and then allowed him to return home. Such an approach often led to resumed violence after the perpetrator returned.
Methods
The study included 314 misdemeanor domestic violence incidents that were handled by police in Minneapolis in 1981 and 1982. Misdemeanor domestic violence crimes differ from felonies, which usually involve serious injuries or use of a weapon.
Volunteer officers agreed to arrest, mediate, or separate couples after an incident according to instructions on the top page of a randomly organized color-coded pad of paper. Then the researchers followed the cases by looking for official reports of subsequent incidents and by interviewing the victims twice weekly for 6 months. The employment of victims’ as well as official reports to measure subsequent violence was extremely important because victims could report on incidents that were not officially documented.
Results
The arrest treatment showed a significantly smaller recidivism level over a 6-month period than the recidivism level for perpetrators who were ordered to leave. The victim interviews indicated a significantly lower recidivism rate for those who were arrested versus those who received advice.
After publication of these findings, there were criticisms related to selection of cases, low level of participation by police officers, and lack of complete adherence to the randomized police responses. In addition, since violence can be cyclical, a 6-month follow-up period is not long enough to demonstrate a deterrent effect on batterers with long cycles.
Impact
In the late 1980s and in the 1990s, the percentage of police departments using arrest as their preferred or mandated policy increased greatly. It is unclear whether the change was due to the impact of the study and its replications (Spouse Assault Replication Project), to the influence of some successful lawsuits against police departments, or to the influence of advocates. A debate about effectiveness has accompanied the spread of preferred arrest policies, and there are philosophical disagreements about whether mandatory arrest promotes victim empowerment.
