Prioritizing Burglars: Comparing the Effectiveness of Geographical Profiling Methods

Abstract

The effectiveness, in prioritizing suspects, of six geographical profiling methods are compared by determining the rank to which each of 92 prolific burglars was assigned, from the total of 400 known burglars, who were selected from a large metropolitan database because they resided in the borough in which the crimes occurred. Using mean and median ranked prioritization of actual offenders, as well as the percentages that appeared in the top 5% of rankings and the area under the curve of a specially developed ‘Ranked Prioritization Function,’ RP (f), it was found that Dragnet using a logarithmic decay function and the distance from the centre of gravity produced the lowest average ranks, with 72% of the actual offenders in the top 5% of prioritized rankings. The implications of the findings are discussed.