Reconstructing Geographic Areas for Analysis of Homicide Rates in Small Populations

Van O'Brien, Northern Illinois University
Fahui Wang, Northern Illinois University

The rate estimates for rare events like homicide in small populations are suspectible to data reporting and other errors, thus compromise the validity of inferences. This research uses two Geographic Information Systems (GIS) methods to construct geographic areas with comparable population sizes from which more reliable homicide rates can be obtained. Using the data aggregated to the redefined analysis areas, the Nisbett-Reaves test of the herding-culture-of-honor hypothesis is replicated. The discrepancies between the results yielded from the original area units and the reconstructed units are evaluated. Analysis of other crime rates may also be subject to the problem of rate estimates in small populations, and thus may benefit from this research.

(Return to Program Resources)

Updated 05/20/2006