Old version: Be careful what you wish for, you might get it. New version: Be careful what questions you ask, you might get an answer. The New Republic is not afraid to ask questions. Tim Stelloh asks "Has one state discovered a simple way to combat domestic violence?"
In recent decades, one of the great grassroots movements of the twentieth century built a raft of protections designed to help abused women. These included a sprawling network of community shelters, gun restrictions for abusers, protection orders, and the nation's first federal anti-domestic violence legislation, the Violence Against Women Act (VAWA). Yet, despite this sustained effort—and even as overall homicides have plummeted nationwide—victims of domestic violence… are today killed in basically the same numbers as they were about 15 years ago. Between 40 and 50 percent of female homicide victims are killed by their husbands, boyfriends, and exes. And, for about half of these victims, police had been alerted to previous incidents of abuse.
There is, however, one exception to this grim trend: Maryland. Since 2007, domestic violence homicides in the state have fallen by a stunning 40 percent. What is Maryland doing that other states are not?
Stelloh tells a number of dreadful stories of domestic violence. His main story is how Jacquelyn Campbell, a professor of nursing at Johns Hopkins University, figured out what to do about it.
A few years after moving to Johns Hopkins in 1993, Campbell and a team of researchers began studying domestic violence murders in Maryland. Their work, which was published in 2002, sought to identify the key indicators that predicted whether a case of domestic violence was likely to become a domestic homicide. The study produced some surprisingly precise findings. If a man had a history of hitting his partner, that in itself was a predictor of murder. But certain kinds of behavior came with even higher chances of death. For instance, if a man choked his partner, she was five times more likely to be killed by him at some point. If he was unemployed, he was four times more likely to kill her. The researchers also found that only 4 percent of homicide victims had ever sought help from a shelter; in a follow-up study, they found that a stay in a safehouse decreased the risk of violent re-assault by 60 percent. Their findings offered new ways to measure risk. "It also informed the system about which cases needed heightened scrutiny," says Campbell…
Indeed, it was clear that the prevailing methods of dealing with domestic violence were inadequate. There had been a slight dip in domestic violence murders in the '90s following the passage of VAWA (which funded anti-domestic violence training and services) and the Brady Law (which required gun dealers to run background checks). But, since then, the number of women killed each year by their partners and exes has hovered around 1,600—that is, only about 500 fewer deaths per year than in 1976. For men, however, domestic homicides have declined from about 1,600 to 600. In other words, all the increased protections for women, the infrastructure of shelters and hotlines, had done a better job of protecting abusers rather than their victims…
By the end of 2005, the group had developed a series of questions that they called "the screen." The first three questions concerned the most important predictors of future homicide: Has the abuser used a weapon against you? Has he threatened to kill you? Do you think he might kill you? If the woman answered yes to any of those questions, she "screened in." If she answered no, but yes to four of the remaining eight questions, again, she was in. Among these were other, less obvious indicators of fatal violence: Has he ever tried to kill himself? Does she have a child that he knows isn't his?
This looks like one of those rare examples of a smashing policy success, from Stelloh's telling. "The screen" provides the essential information that almost always makes the difference in interventions: it tells the police which cases to focus their resources on.
I would note, however, that there is another term for "screening". It is "profiling". Profiling is exactly what Stelloh is talking about here. The term simply means that you come up with a list of criteria that help the police decide who to pay attention to.
Profiling has become a political stigma. This is so for very good reasons. Racial profiling, whether deployed by a cop looking for drug runners on a Florida highway or a TSA agent on the lookout for terrorists at an airport, carries two significant burdens. It presents inconveniences for persons of certain races and, worse, involves an assumption of guilt on the basis of race.
The profiling described by Stelloh seems innocent of these problems. It focuses on actions of the potential abuser as the potential victim reports them. But what if it turned out that Canadian men were significantly less likely to murder their spouses and girlfriends than, say, Texans? Would you omit that from the screen, even if it meant that some women would die as a consequence?
Profiling is an essential element in law enforcement. We are a long way from resolving the problems involved in it, but it is stupid to think we can dispense with it. Meanwhile, three cheers for Maryland and Jacquelyn Campbell. Protecting women from murderous abusers is what law is about if it is about anything.