Modelling Property Crime Using the British Crime Survey. What Have We Learnt?

Abstract

The British Crime Survey (BCS) has been used to develop a number of statistical models that describe property crime victimization at the level of the individual household. This paper gives an overview of what has been learnt from these studies. In terms of the predictors of crime, it is now well established that both household and area characteristics play important roles: in particular, household affluence has a positive effect on crime victimization, in contrast to the negative effect exerted by area affluence. However, findings of the latter part of the 1990s have emphasized that crime victimization cannot be regarded as random even when the statistical model is conditioned on these known characteristics. Based on a more general model which allows for this nonrandomness, the present study uses simulations to illustrate the roles of household and area characteristics in respect of the following: how household types differ in predicted property crime victimizations, both in aggregate and differentially by crime type; how area characteristics are associated with rates of victimization; how household victimization histories affect predictions of subsequent crime events.