An Examination of Offense Specialization With Self-Report Data: A Statistical Model, Test, and Replication

Christopher J. Schreck, Illinois State University
D. Wayne Osgood, Pennsylvania State University

This study develops a new approach for studying specialization in property versus violent crime that (1) distinguishes this specialization from the general propensity to offend and (2) permits researchers to determine relationships of explanatory variables to both general propensity and specialization. We use these methods to examines whether a range of crime correlates influence the likelihood that an adolescent will specialize in either property or violent crime. We define specialization in terms of the variety of crimes that an offender commits. While relatively few studies use this definition, existing research that does faces many limitations that make conclusive results unlikely. One of these limitations is the inability of available statistical techniques to easily control for multiple independent variables. We employ hierarchical linear modeling to implement our approach with two sets of longitudinal data: the Monitoring the Future study and the Montreal Youth Project. This analysis strategy enables researchers to control for many independent variables and to determine whether these factors influence participation in violent or property crime. We report preliminary results from these analyses and speculate on future directions for specialization research as well as other applications of this analysis technique.

(Return to Program Resources)

Updated 05/20/2006