Trying to Relate: Victims, Offenders, and Their Classification in Homicide Research

Marc Riedel, Southern Illinois University - Carbondale
Wendy C. Regoeczi, Cleveland State University

Victim-offender relationships may be central to explaining homicide. The disaggregation of homicides by type has salience for both testing theory and public policy. However, the classification of victim-offender relationships in widely available homicide data sets such as the Supplementary Homicide Reports or the Homicides in Chicago data file include numerous detailed categories designed to cover the broad spectrum of relationships that may exist between the victim and the perpetrator of a homicide. Researchers tend to find the analysis of such classifications far too unwieldy and consequently they are inclined to collapse them into a more manageable number of categories. The problem with this aggregation is that it often occurs haphazardly, with little use of theory for direction. This has led to tremendous variation across studies in what kinds of relationships are considered "acquaintances," "family," or "friends," making comparisons of findings difficult and applications to policy problematic. In this paper we propose using clustering methods to create a classification of victim-offender relationships using variables identified as important on the basis of theory or previous research, with the ultimate goal of developing a widely applicable classification of victim-offender relationships. Analyses are based on homicide data for California covering the years 1994-1999.

(Return to Program Resources)

Updated 05/20/2006