The Effect of Homophily on Co-Offending Outcomes

This study explores how similarity preference among offenders impacts the structure of co-offending networks and the diffusion of information, mod-eled through skill exchange. Co-offending provides certain advantages, but factors such as trust and usefulness of co-offenders significantly influence these criminal collaborations. Using an Agent-Based Model, we simulate in-teractions between offenders based on varying levels of similarity prefer-ence, allowing us to observe network characteristics such as density, transi-tivity, average degree, tie weight, and skill acquisition. The results show that stronger similarity preferences lead to sparser but more stable criminal net-works with a higher number of repeated interactions between the same of-fenders. However, increased exclusivity in partner selection reduces the in-formation diffusion within the network, limiting the number of skills ac-quired by offenders. Conversely, inclusive partner selection facilitates great-er skill exchange but results in fewer strong ties between offenders. This study highlights the trade-offs between stability and skill diffusion in crimi-nal networks. Networks with high homophily tend to be more stable but less skilled, while more open networks allow for greater exchange of knowledge at the cost of connection strength. These findings contribute to the under-standing of how offender collaboration shapes criminal network structures and the spread of criminal opportunities.

Sep 19, 2024

Social Simulation Conference 2024