An Exploratory Analysis of Homicides in Washington, DC

Elizabeth R. Groff, National Institute of Justice
Julie Wartell, Institute for Law and Justice
J. Thomas McEwen, Institute for Law and Justice

This research is an exploratory data analysis of homicides that occurred in the time period 1990-2000 in Washington, D.C. The main objective of the research is to use spatial statistics to describe the patterns of all homicides and of several subsets of homicide. The subsets examined are homicides by race, sex and age of victim. Both point pattern and area level spatial techniques are used. Point pattern techniques are used to describe both the degree of clustering in the distributions and the spatial scale at which the clustering occurs. However, in order to provide a better understanding of homicide in its community context, it is essential to include the social structural charateristics of the areas. Spatial autocorrelation of homicide rates is computer and a regression tree model is developed to predict homicide rates. This study clearly indicates the significant clustering of homicides in general and by age, race and sex of victims. A regression tree analysis reveals that education is the most significant variable in predicting homicide rate at the blockgroup level.

(Return to Program Resources)

Updated 05/20/2006