Summary: New research shows that a strong friendship network can be as important as a charismatic leader in driving group violence.
Study Finds Friendship Networks Drive Raiding Behavior Among the Nyangatom; Leadership Matters Most for Initiation
Researchers studying the social dynamics of the Nyangatom, a nomadic pastoralist community in East Africa, report that dense friendship networks among men play a major role in shaping participation in intergroup raids. While experienced or well-connected individuals help trigger raids, the decision of most participants to join is driven more by their ties within the wider social network than by loyalty to a single leader.
The findings, published in the Proceedings of the National Academy of Sciences (PNAS), come from a three-year longitudinal study that mapped interpersonal connections among Nyangatom men and tracked who initiated and joined raiding parties. The study’s results have implications for understanding collective violence in settings ranging from small-scale societies to urban gangs, insurgent groups, and other contexts where informal networks shape risky collective action.
“Social interactions embedded in networks are central to many forms of collective behavior,” said Nicholas Christakis, co-director of the Yale Institute for Network Science and senior author of the study. “Networks can foster cooperation and innovation, but they also enable the rapid emergence and spread of violence.” Christakis, who holds appointments in sociology, ecology & evolutionary biology, biomedical engineering, and medicine at Yale, emphasizes that studying real-world social ties reveals how group actions develop outside formal institutions.
The research identifies two complementary roles played by social structure in the formation of raiding parties. First, individuals who have participated in many raids and who occupy central positions in the network tend to act as initiators. These leaders are more likely than others to spark a raid because of their reputation, experience, and the breadth of their connections.
Second, once a raid is initiated, recruitment extends far beyond the immediate friends of those leaders. The study shows that non-leaders collectively have a larger influence on who ultimately joins a raid because they bring their own distinct sets of friendships into play. In other words, the overall pattern of participation is distributed across the broader social network rather than being confined to a leader’s clique.
“Collective action rarely begins with a single charismatic figure pulling random followers into action,” said Alexander Isakov, co-first author and a postdoctoral researcher at the Human Nature Lab. “People are more likely to take part because of their friendships and the network positions they occupy relative to other participants.”
Co-first author Luke Glowacki, a research fellow at the Institute for Advanced Study in Toulouse, pointed out that the Nyangatom lack formal political leaders or centralized institutions. “That absence of formal hierarchy makes it easier to observe how informal social ties organize real-world collective behavior,” he said. The researchers note that even in the absence of formal rank, distinct roles emerge that resemble a leadership structure: some individuals repeatedly initiate raids while others act as core or peripheral participants.
From a policy perspective, the study suggests two different levers for influencing violent collective behavior. Targeting or diminishing the influence of frequent raid initiators may reduce the likelihood that a raid is launched in the first place. However, because participation can quickly draw individuals from across the entire social network, interventions aimed only at leaders may not prevent mobilization once a raid is underway. Network-level strategies — for example, reducing the density of risky ties or strengthening alternative, nonviolent connections — could therefore be important complements to leader-focused measures.
The study used longitudinal social network mapping and field observations to document how raiding parties formed and who participated. Additional co-authors include Richard Wrangham (Harvard University), Rose McDermott (Brown University), and James Fowler (University of California, San Diego). The research provides a mathematical framework to identify leaders and to track how multiple groups mobilize over time.
Journal: Proceedings of the National Academy of Sciences (PNAS).
Lead authors: Luke Glowacki and Alexander Isakov. Senior author: Nicholas A. Christakis.
The study titled “Formation of raiding parties for intergroup violence is mediated by social network structure” reports that intergroup violence among the Nyangatom is shaped by within-group social dynamics. Initiation of raids depends on specific leaders who participate frequently, have more friends, and occupy central network positions. Yet raid participants come from across the entire population rather than only from the direct friends of leaders. An individual’s choice to join a raid is strongly associated with their network ties to other participants, and non-leaders exert a larger total effect on participation than leaders. The findings indicate that social networks can support risky collective action, amplify the emergence of raiding parties, and facilitate intergroup violence in small-scale societies.
Research source: Yale University. Image credit: Luke Glowacki. Original research published in PNAS (October 2016).