The Minneapolis Domestic Violence Experiment
"If the facts don’t fit the theory, change the facts."
— Albert Einstein
I received more emails about my last column, Mandatory arrest: A flawed policy based on a false premise, than any column to date. Most of the comments were similar to those I received following my presentation at the February domestic violence conference (PDF) in Sacramento, Calif. that was sponsored in part by the National Family Violence Legislative Resource Center.
At the conference I stated that, to date there is not a single empirical study that documents mandatory arrest works best for victims, offenders, or law enforcement. Apparently there are some people that attended the conference or read my column who believe the Minneapolis Domestic Violence Experiment (MDVE) documents that arrest works best. However, if you read the MDVE report you will discover that is not what the MDVE demonstrates.
Similar to the following paper What Does Research and Evaluation Say About Domestic Violence Laws? (PDF), it is almost universally written and accepted by most domestic violence advocates, academics, public policy makers, and the electronic and print media that that the MDVE proved that arrest works best concerning law enforcement domestic violence intervention. The above paper claims:
In the most-far-reaching of these studies, an experimental-design study conducted in Minneapolis reached the conclusion that arrest of batterers results in an overall reduction of recidivism: reduced repeat assaults (Miller, 2005, p.14).
Some misinterpretations may be the fault of the MDVE co-authors, Lawrence W. Sherman and Richard A. Berk. On the first page of the MDVE, Sherman and Berk write, “It [the MDVE] found that arrest was the most effective of three standard methods police use to reduce domestic violence.” This claim, in and of itself, certainly seems to demonstrate that arrest does work best.
However, Sherman and Berk are only partially responsible for the ubiquitous misinterpretations of their study because in the same paragraph they write, “These were not life-threatening cases, but rather the minor assaults which make up the bulk of police calls to domestic violence.” The MDVE clearly documents that it screened out the serious cases and researched only minor incidents. What the MDVE does seem to demonstrate is that arrest works best for minor domestic violence (family conflict), not serious battering behavior.
Even that conclusion contains some important caveats. In the next paragraph they write, “The findings, standing alone as the result of one experiment, do not necessarily imply that all suspected assailants in domestic violence experiments should be arrested.” And, “Other experiments in other settings are needed to learn more.”
And did the advocates, public policy makers and the electronic and print media listen to these concerns and wait to learn more? Just a few years after the MDVE, 90 percent of law enforcement agencies had either “encouraged” or “required” arrest policies. In 2008 almost half the states have mandatory arrest and prosecution policies and all have some form or type of preferred or encouraged arrest policies.
Spouse Assault Replication Program (SARP)
The National Institute of Justice (NIJ) and the Centers for Disease Control and Prevention (CDC) co-sponsored five programs, as suggested by the MDVE. On the first page of the SARP the authors write that the use of arrest was only occasionally associated with statistically significant reductions in reduced repeat assaults. The SARP also reported that the majority of men discontinued their offending without an arrest.
The authors of the SARP note that policies mandating arrest for all suspects may unnecessarily reduce the ability of the criminal justice system to serve diverse individual victims and complex cultural communities. The SARP also suggests that the best response, given the limited resources and personnel of the criminal justice system, may better serve the community by identifying and focusing on the most violent offenders and those victims most at risk and in need of assistance.
Some of my past columns demonstrate that there are a growing number of studies that document “one-solution-fits-all” criminal justice intervention processes have produced some unforeseen and unintended negative consequences. These “one-solution-fits-all” polices and practices may save some lives and may make some families safer. However, at the same time a growing number of studies documents that “one-solution-fits-all” styled intervention can have negative consequences and do endanger some of the victims they are intended to protect.
In my column National Institute of Justice Studies Ignored, I detailed some of the unintended consequences in the following studies: Controlling Violence Against Women, Forgoing Criminal Justice Assistance, Safety and Justice for All (PDF), Effects of No-Drop Prosecution of Domestic Violence Upon Conviction Rates, Exposure Reduction or Backlash? (PDF) and Advancing the Federal Research Agenda on Violence Against Women.
The study Effects of Victims' Experiences With Prosecutors on Victim Empowerment and Re-occurrence of Intimate Partner Violence, Final Report (PDF) is also relevant. This study found that no-drop policies had the effect of lowering victims’ empowerment and were unrelated to the reoccurrence of violence in victims’ lives.
The article Mandatory Arrest and Prosecution Policies for Domestic Violence is on the National Violence Against Women Prevention Research Center Web site. This article suggests, as do many others, that “one-solution-fits-all interventions” may have very negative effects on some victims. It suggests that all domestic violence interventions should be tailored to fit the diverse and often complex needs of the victims.
And what should be most troubling for advocates, academics, public policy makers and the electronic and print media is a recent study, Does the Certainty of Arrest Reduce Domestic Violence? Evidence from Mandatory and Recommended Arrest Laws. This study provides evidence that mandatory arrest laws may have played a role in harming the people they are intended to help by increasing the number of intimate partner homicides.
This increase in intimate partner homicides is documented in the Bureau of Justice Statistics (BJS) online report Homicide Trends in the U.S. under, “The proportion of all homicides involving intimates by gender of victim, 1976-2005.” From 1976 to 1993 the number of female intimate partner homicides decreased from 34.5% to 28.2%. From 1994 to 2005, while the percentage of non-intimate or unknown homicides of women decreased from 72.0% to 66.7% the number of intimate partner homicides of women increased from 28.0% to 33.3%.
Advocates, researchers, or public policy makers need only to read the MDVE to discover that it did not include the full spectrum of complex and multifaceted domestic violence incidents. In fact the MDVE was a very limited experiment: “The design applied only to simple (misdemeanor) domestic assaults, where both the suspect and the victim were present” (MDVE, p.2). Hence, the MDVE demonstrates that arrest may work best only for what is labeled “family conflict” or minor domestic violence incidents. The MDVE did not provide any relevant data concerning deterrence for serious long term violence or “battering behavior.”
The MDVE also warns that the socioeconomic and cultural demographics of Minneapolis may not be comparable in many other urban settings and the effects of “one-solution-fits-all” policies may prove to be different in different settings. In fact, the SARP clearly demonstrates that there are different effects in different settings.
The MDVE made no recommendations for the implementation of mandatory arrest or mandatory prosecution policies. In fact the MDVE project director, Lawrence W. Sherman in his book, Policing Domestic Violence: Experiments and Dilemmas calls for the repeal of mandatory arrest laws. Sherman does not believe that mandatory arrest will provide a general deterrent effect concerning the general public and he fears that mandatory arrest may actually be detrimental to many people who live at lower end of the socioeconomic educational strata of society. The rise in intimate partner homicides may provide verification to Sherman’s fears.
And before you challenge or disagree with the conclusions of this column you should first read the hyperlinked relevant studies this column provides and decide for yourself what the facts are, rather than simply accepting as fact what others claim the MDVE documents.
And because I believe that complaining about what is wrong without suggesting what might be right, is in and of itself the wrong thing to do, my next column will make some suggestions concerning what policies and procedures might be more effective than “one-solution-fits-all” intervention policies.