UC Berkeley

Social groups are arenas for both cohesion and conflict. Whereas prevailing theories focus on how these processes unfold at the boundaries between groups, the authors focus on the tensions that emerge within groups and give rise to uncivil discourse. They develop a novel theoretical account of its network-structural antecedents. In polarized online groups, they hypothesize that the greater the structural similarity between two individuals, the less likely they will be to direct uncivil language toward one another. This relationship will be moderated by the degree of group polarization. Using a node-embedding algorithm to derive an omnibus measure of interpersonal structural similarity, they find support for the theory using a dataset encompassing more than 25 million comments made by over 1.7 million users in six polarized communities on Reddit. They discuss implications for research on intergroup animosity, group polarization, the measurement of structural similarity, and the interplay of structure and culture.