The micro dynamics of violent groups

J. Bruggeman, D. Weenink


To explain the dynamics of collective street violence, we examine video footage of 59 groups of (mostly) young men. During the rising turmoil of mutual provocations preceding a fight, participants are uncertain about the outcome, i.e. the public good of winning and its costs, and therefore become conformists to their group mates. We measure conformism in terms of tie-wise behavioral synchrony in the network of a focal group, and turmoil in terms of aggressing and moving around by opponents. We then use an asymmetric Ising model, reflecting the asymmetry of defection and cooperation, to explain the bursts of collective action that we often observe, as well as the exceptions without a burst. The model predicts that if turmoil rises among conformists, one or few individuals will accidentally contribute to the public good, i.e. act violently. At a critical level of turmoil, these initial cooperators trigger a cascade of cooperation. We thus recover the main results of critical mass theory without assumptions of rational choice or learning. The model also predicts that collective violence appears in small (sub)groups at relatively lower levels of turmoil, consistent with our observations. When simulating an increasing chance of tie interruptions, corresponding to group members losing sight of one another during a fight, the level of cooperation decreases. This model behavior predicts which de-escalation strategy will be effective, namely distracting participants by members of the focal group or the audience and thereby interrupting ties and decreasing exposure to opponents' provocations. The cases where this kind of de-escalation was practiced were indeed the exceptions where turmoil did not result in bursts of violence.

← Schedule