In episodes of intergroup violence, which group members participate and which do not? Although such violence is frequently framed as occurring between distinct ethnic, racial or sectarian groups, it is easily overlooked that it is usually only a subset of the group's members who in fact participate in the violence. In predicting participation, extant research has privileged an atomistic approach and identified individual attributes indicative of a predisposition to violence. I suggest instead that a situational approach should complement the atomistic paradigm and present evidence that an individual's micro-spatial environment is an important predictor of differential participation in intergroup violence. Using GIS data on 3,426 residents from one community, I map the household locations of participants, non-participants, and victims of Rwanda's 1994 genocide. I find that participants are likely to live either in the same neighbourhood or in the same household as other participants. Specifically, as the number of violent to nonviolent individuals in an individual's neighbourhood or household increases, the likelihood of this individual's participation also increases. In explaining these neighbourhood and household effects, I suggest social influence is the mechanism at work. As micro-spatial distance decreases, micro-social interaction increases. Neighbours and household members exert influence for and against participation. Participation then may be as much the product of social interaction as of individual agency. What neighbours and family members think, say and do may influence participation in collective action such as intergroup violence. The conceptualization of neighbourhoods and households as micro-spheres of influences suggests the importance of social structure as a determinant of participation.
