People who want sweeping answers or solutions to the problem of domestic violence will not be comforted by the research of Lawrence W. Sherman, a University of Maryland criminologist and director of the Georgetown-based Crime Control Institute.
Ten years ago, Mr. Sherman conducted one of the first and best-known studies on domestic violence.
"The Minneapolis experiment" indicated that arrest seemed to have a deterrent effect. But Mr. Sherman shied away from recommending mandatory arrest policies until more studies could be conducted.
Six subsequent studies, by Mr. Sherman and others, came up with different results. In households where men were unemployed, for example, women were more likely to be beaten again if the man was arrested.
Yet communities across the nation continue to adopt mandatory arrest policies. And Sen. Joseph R. Biden Jr.'s Violence Against Women Act of 1993 supports mandatory arrest policies.
QUESTION: Can we prevent domestic violence?
ANSWER: Prevention is a matter of reducing probability. Can you lower the probability of a crime? Yes.
The Secret Service lowers the probability of a presidential assassination. Metal detectors in airports and courtrooms lower the probability of crimes there. When convenience stores started keeping less cash on hand, the number of armed robberies dropped.
When it comes to domestic violence, you have to start focusing on demonstrated cures and ask if they work.
Mandatory arrest doesn't work. And while you can predict in which cases the frequency of violence will escalate, you cannot predict homicides.
Q: Yet San Diego put together a special police unit and changed the way it deals with domestic-violence cases, and its homicide rate dropped for such crimes.
A: Fewer homicides in San Diego is such a statistically unreliable DTC idea; it's so small a number. A homicide rate in one city is not statistically significant.
In Milwaukee, we found that only one out of 33 homicides had ever contacted the police before in a domestic-violence case. The others had no prior contacts with police. We had similar results in other cities.
Q: So you can't predict who will be killed. What can we do?
A: The single recommendation is to have a community policing policy, a different policy for each neighborhood.
An example of that is "re-integrative shaming," which may be more effective, and indeed may be tougher than arrest. [What's meant by that] is a police-led confrontation over responsibility for the violence and reparations.
It includes the offender, and the victim and anyone else who might be involved -- family members, neighbors. I've seen this done with juveniles in Australia, and I think it ought to be tried.
Q: But some advocates for battered women aren't comfortable with the idea of the police dealing with this issue. What about special prosecutors, or courts?
A: People criticize me for saying that police are the only game in town, but it's the truth. The courts are bailing out of domestic violence.
Q: Why does arrest affect unemployed men differently than employed men?
A: Unemployment may simply be an indicator of some underlying characteristic that makes punishment more likely to backfire.
I have a theory that people who are not well-bonded to society are more likely to maintain their own sense of pride by fighting back against punishment and showing the state cannot squash them down.
They prove their independence by becoming more criminal.
Q: What about victims in famous domestic violence cases such as Kristin Lardner, who was killed by her ex-boyfriend?
According to the account by her father, Washington Post reporter George Lardner, the court system in Massachusetts had flaws that kept him on the streets. Could her death have been prevented?
A: I'm not sure her case was preventable. There are certain people, such as Kristin's ex-boyfriend, who are so bad, they cannot be gotten rid of. For us to lock these people up, we have to be willing to throw out the Constitution. But nothing allows this -- and for good reason. There are all kinds of angry people out there, most of whom are never going to hurt anyone.
We need to take threats a lot more seriously. One solution is to provide stalking legislation.
But then we need to study the laws. If we really care about the problem, then we have to make sure the solutions work.
Q: How are you going to convince legislators in these times to pass legislation that costs more money?
Let's be creative. Use the fines -- dedicate all the fines in the new law to support an evaluation of their effectiveness. Or tax alcohol. Alcohol is fair game, given the role of alcohol in domestic violence.
Q: How do feminists and others interested in domestic violence issues react to your rejection of mandatory arrest policies?
A: It's certainly not politically correct with feminists.
But it's fair to say that feminism is a white middle-class movement, and any research that splits a movement is controversial. No political movement wants to have anything that divides its ranks.
Q: What about people who believe batterers must be punished?
That's exactly what some people argue. They say that deterrence isn't the issue, that it's an issue of "just deserts."
That's a way of side-stepping the issue. And it's a prime example of how lawyers reason. They go backward from the result they want.
If you start backing away from criminal sanctions, advocates say you're soft on crime. [What matters is] what is effective. For my critics, it's a question of symbolic justice; for me it's an issue of preventing violence.
Perhaps I undervalue symbolic justice, but then I don't thirst for vengeance.