Title: The Influence of Affiliation Within Criminal Groups

• In Event: Crime and social Networks

Abstract

Abstract: Introduction. Co-delinquency and organized crime inspired research for many years. Multiple avenues have been explored, mostly with longitudinal research. Those studies were aiming to understand the impact of juvenile co-offending patterns on adulthood criminal behavior. Afterwards, co-delinquency was mostly explained by criminal offenders’ history to predict future delinquency patterns in criminal groups. Then, closer to this research, came social links mixed with other affiliation theories to describe affiliation patterns in criminal groups offending. But most recently, social network analysis, from the anthropological branch, was introduced as a newcomer in the criminological field to describe the structure of criminal groups, allowing this research to bring light to unexplored criminal organized group affiliation behaviors and practices. This presentation, associating social network methods and regression analysis, aims to provide a whole new perspective to understand the influence of affiliation on offenders joining organized criminal groups. Methodology. The dataset is divided into four equal temporal parts in which offenders were possibly arrested multiple times in different time periods. Every time period in the dataset is then analyzed with a two-mode network (offenders and events) to create a complete one-mode network (offenders only) followed by cluster analysis then to detect communities. This method allows, at first, to statistically identify affiliated offenders. Then, it allows observing the criminal career trajectory of offenders, but also to identify potential predictors with linear and logistic regression analysis. Results. Results suggest that age, gender and ethnicity as well as criminal career parameters like offense types could be potential predictors of affiliation. Finally, quite interestingly, criminality could be as much important in affiliation predictions as affiliation by itself.

Authors

• Henrik Vaugeois, Université Laval

• Yanick Charette, Université Laval