Predicting Domestic Homicide: Prior Police Contact and Gun Threats (From Woman Battering: Policy Responses, P 73-93, 1991, Michael Steinman, ed. -- See NCJ-129473)

• NCJ Number: 129477
• Author(s): L W Sherman; J D Schmidt; D Rogan; C DeRiso
• Date Published: 1991
• Length: 21 pages
• Grant Number(s): 86IJCXK043
• Sponsoring Agency: National Institute of Justice (NIJ); US Dept of Justice NIJ Pub
• Sale Source: Anderson Publishing Co
• Publication Type: Research (Applied/Empirical)
• Language: English
• Country: United States of America

Annotation

Information gathered regarding domestic violence in Milwaukee and a reanalysis of earlier studies conducted in Boston, Kansas City, and Minneapolis indicated that predicting domestic homicide is more complicated than previously thought.

Abstract

The Kansas City study found that in about 90 percent of the cases of domestic homicide, police had responded to at least one call for service at the address of the victim or suspect in the previous 2 years and to 5 or more calls in about half the cases. The Boston and Minneapolis data on repeat calls were consistent with these results. However, the records focused on addresses rather than specific people, reflected small samples, and did not indicate the seriousness of the prior violence. The Milwaukee research overcame some of these problems, using 1989 data on gun threats as revealed in records of police calls to the city's domestic violence hotline. Results indicated that police can clearly identify buildings with high predictability of recurrent domestic disturbances and couples with high predictability of recurrent minor domestic battery. However, predictions of domestic homicide based on prior police contacts with couples have a false-positive error rate of more than 99 percent, and little would be gained from police intervention. Figure, tables, and 15 references