The impact of violent behavior on co-offender selection: Evidence of behavioral homophily

Purpose: Research into adolescent co-offending has predominantly focused on individual differences in co-offender selection, with less attention to diversity in offending behavior. Research suggests that offenders with a violent history may be less desirable as co-offenders, potentially leading to homophily based on offending behavior. This study aims to examine homophily based on violent offending behavior among co-offenders.Methods: We applied social network analysis to a Belgian co-offending network (N = 33,815) using police data and introduced a measure of individual violence level based on offense types. We hypothesized that suspects tend to co-offend with suspects who share a similar violence level, and examined the relationship between network structure and the suspects’ violence level. Results: The results demonstrate a significant degree of homophily based on violence, with a more pronounced effect among non-violent suspects. This effect was consistent across different age groups.Conclusions: Our findings reveal behavioral homophily in violent offending behavior and highlight the importance of including older suspects to fully understand network dynamics. Intervention strategies should consider the broader criminal social environment to prevent the escalation of violence. Future research should employ dynamic network analyses to better understand the influence of violent behavior in co-offending networks over time.