The spatial dependency of crime increase dispersion

A number of analytical techniques (such as the Gini coeffi cient and the Lorenz curve) can identify unequal distributions in crime frequency among sub-areas within a study region; however, these tools are often aspatial and say nothing about the relationships between spatial units. Using dispersion analysis, a technique that measures the relative dispersion of a crime increase across a region allows for the identifi cation of particular spatial units that are suffi ciently infl uential to drive up the overall jurisdictional crime rate. In this article, a combination of the order of areal units from a dispersion analysis with a measure of the local level of spatial association is used to develop a tool that can identify clustered areas of emerging crime problems. The identifi cation of these second-order spatial processes may be benefi cial to police departments and crime prevention practitioners who are interested in the identifi cation of statistically signifi cant clusters of emerging crime hotspots. The process is demonstrated with an example of robbery rates in police sectors of Philadelphia, PA. Security Journal (2010) 23, 18 – 36. doi: 10.1057/sj.2009.16