Beggarman or thief: Methodological issues in offender specialisation research

The problem of offender specialisation is one of the long-standing themes in theoretical and empirical criminology. Specialised offending suggests that there are specific causal factors that give rise to particular kinds of crime, whereas generalised offending is evidence for broader social and economic causes. Criminological research has shown consistent but weak evidence for specialisation. However, studies of offending specialisation entail methodological assumptions about how episodes of offending should be conceptualised and classified. This study examines two aspects of the methodology of offender specialisation research and their impact on specialisation indices like the Forward Specialisation Coefficient. The issues examined are first, how offending is represented in terms of offence classifications, and the impact of using a narrow versus broad classification system, and second, the impact of different rules for classifying episodes that comprise multiple offence types. Both of these methodological issues were found to have an impact on measures of offence specialisation. In general, higher levels of specialisation were found when broad offence categories were used compared with narrow ones, and when mixed offending episodes were defined by a Most Serious Offence rule compared with a generic 'combination' category. These findings suggest that specialisation research needs to give greater attention to the assumptions made in classifying and aggregating offence data.