Social Network Structures and Crime Mobility: A Discrete Choice Analysis of Co-Arrest, Co-Victimization, and Overlap Networks

• In Event: Quantitative Applications in Criminal Justice Research

Abstract

This study examines the role of three types of social network structures—co-arrest, co-victimization, and arrest-victimization overlap networks—in shaping crime mobility patterns. Using Dallas police arrest and incident data from 2014 to 2024, along with 2010 census data, we construct both homogeneous (co-arrest/co-victimization) and heterogeneous (arrest-victimization) networks. The spatial unit of analysis is block group. Through link prediction, we first simulate social network and generate network structures. We then assess how these network structures influence offender and victim locational choice using discrete choice models (DCM). In the DCMs, we estimate the probability of an offender/victim traveling to specific block groups based on different network features. Our findings seek to broaden offender-centric theoretical frameworks, extending their applicability to co-arrest, co-victimization, and arrest-victimization overlap in offender and victim mobility.

Author

• Yuki Wang, University of California, Irvine