Microcosms of violence among street gang members: Social contagion, propensity to violence, and gang embeddedness

Gang members face a paradox: while they may join a gang for protection from violence, they are actually more likely to be victimized than non-gang members. Although it is known that gang affiliation increases the risk of violence perpetration and victimization, little is understood about the factors within gangs that influence these risks. This study examines the relationship between violent perpetration and victimization within the context of gang networks. Using 20 years of police data, we mapped the incidents of violent victimization and perpetration among 1587 Haitian street gang members and their affiliates in Montreal, Canada. Our results show that violence occurs in clusters within these groups and that victimization and perpetration are more likely to happen in the same locations within the network. Regression models revealed that victimization was strongly related to: (1) having committed violence, (2) having more violent perpetrators in one's entourage, and (3) having more victims in one's entourage. These three effects were interdependent, creating a mutual aggravation effect: members who had perpetrated high levels of violence, high levels of victimization in their network, and who had violent peers were 15 times more victimized than members who were not directly or indirectly involved in violence. The structure of peer relationships was also important. Denser networks provided some protection against victimization, but this was dependent on members' own level of violence. Violent perpetrators did not benefit from the protection offered by a close-knit network. Our findings show that violence within gangs is not equally distributed and is concentrated in certain areas of the network. Perpetration and victimization are linked, and the local density of the network can reduce the impact of violence in the network. Thus, the idea that gangs can provide protection may not be as paradoxical as it seems. In tightly knit groups, and for members not directly involved in violence, gang affiliation did not increase violence risk. This understanding may improve targeted interventions to prevent both the experience and commission of violence.