Constructing Geographic Areas for Analysis of Homicide in Small Populations: Testing Herding-Culture-of-Honor Proposition

Abstract

The rate estimates for rare events like homicide in small populations are very susceptible to data errors, and thus compromise the validity of inferences. This chapter discusses two geographic information systems (GIS)-based methods for constructing geographic areas with sufficiently large base populations to permit reliable estimates of homicide rates to be obtained. One is the spatial order method, and the other is the ISD method (after the Information & Statistics Division of the Health Service in Scotland, where it was devised). Both methods construct new analysis areas based on spatial proximity of basic units. For demonstration, the methods are applied to testing the herding-culture-of-honor hypothesis proposed by Nisbett and Reaves, and the result shows that the herding-culture-of-honor proposition is merely an artifact of unreliable estimate of homicide rates. The methodology, in general, is applicable to analysis of any rates with small base populations.