The 0. J. Simpson affair is only the latest in a series of events that have focused attention on domestic violence. Using data collected by Murray Strauss and his colleagues, Lawrence Sherman (1992) estimates that each year there are 18 million incidents of domestic violence fitting the criminal-justice classification of an assault and that police officers handle at least two million cases of violence involving a spouse or lover. Domestic-violence assault is more common than all other forms of violence combined. Recognizing that domestic violence is a social as well as a private issue, a number of police departments have participated in experiments designed to learn how they can best handle domestic-violence calls. In this paper, we use data from the experiment in Minneapolis to determine how police treatments in cases of domestic violence affect the couple's subsequent violence. We also examine how socioeconomic factors such as age and employment are related to domestic violence. These factors are considered in part because they may affect the police actions. Our model is a stochastic dynamic model that allows the effects of the police actions to vary over time. In brief, we find that arrest is more effective than advising or short-term separation but that the differential effect is transitory.

We estimate the general deterrent effect of criminal justice resources on criminal behavior using panel data for a sample representative of young men in U.S. urban areas. Our data, which combine individual-level information on arrests and personal characteristics with aggregate information on criminal justice resources, allow us to obtain deterrence measures that reflect theoretical concepts and are of potential policy relevance. We find robust evidence of a general deterrent effect flowing from criminal justice, particularly police, resources. The nature of our data also allows us to conduct extensive exogeneity tests and to explore a number of possible 'third causes' for the deterrent effect. We find strong evidence that our deterrence variables are exogenous and no evidence that our deterrence result stems from commonly hypothesized third causes. Copyright 1994 by MIT Press.