Testing Labels: Comparing Network and Traditional Approaches to Identifying the Boundaries of Criminal Organizations

• In Event: Drugs, Gangs, and Co-Offending: Advances in Social Networks Methods

Abstract

Rigid labels assigned by police to organized crime groups make targets more tangible and, thus, more amenable to identification, investigation, and prosecution. But rather than accurately capture the qualities of co-offending networks, qualitative assignment of such labels may actually confine the way in which we consider measuring group structure within large, complex networks of co-offenders. We shift from traditional conceptualizations and measurements of crime groups, placing the “community” as the unit of analysis to test the boundaries of group labels assigned by police tasked with investigating serious and organized crime. Our data afford us this opportunity, capturing the nature of more than 3,000 unique ties comprising a network of 1,600 career criminals, serious and non-serious offenders, associates, and non-offenders, all of whom were knowingly or suspected to be involved with serious and organized crime in the Canadian province of Alberta. The research design consists of a multi-method approach to community detection analysis to uncover boundaries across the network. The group boundaries determined by community detection analysis are compared to those identified by law enforcement, thereby enabling us to assess whether the rigid boundaries imposed by group labels are warranted.

Authors

• Martin Bouchard, Simon Fraser University

• Mitchell Macdonald, Simon Fraser University