Developments in geographical offender profiling: commentary on Bayesian journey-to-crime modelling

The Bayesian approach covered in this special issue complements the earlier ways of exploring offenders' crime location choices. It is different from earlier approaches in dealing directly with large local data sets of solved crimes to develop statistical models as compared with attempts to derive more generic ways of conceptualising an offender's geographical crime pattern. The two approaches can thus be seen to enshrine different epistemologies; the Bayesian being about a process for deriving patterns in geographically specific data sets, the earlier centre-of-gravity-oriented approaches seeking to establish general theories and principles that are broadly independent of any given locality. They are consequently sensitive to different types of bias in the data available. However, the interesting results achieved through the analysis of Bayesian patterns within crime data does further the consideration of the similarities across offenders in their use of locations for crime and thus has theoretical as well as practical implications. Therefore, these two approaches need to build on each other. This complementarity provides a useful example for other areas of Investigative Psychology. Copyright © 2009 John Wiley & Sons, Ltd.