Untangling altruism and parochialism in human intergroup conflict

Human intergroup conflict occurs on a scale unmatched in other mammals. Paradoxically, this capacity for war is closely linked to our exceptionally cooperative abilities. Models of "parochial altruism" describe how within-group cooperation and between-group competition may co-evolve, but it is unclear whether these models reflect human preference adaptation in real-world conflicts. Across five studies (total <i>N</i> = 1,121), we develop and validate a psychometric toolkit to test the core assumptions and predictions of parochial altruism models in groups involved in real conflicts of varying intensities. Our measures clearly distinguish interindividual altruism from intergroup parochialism, outperform prior metrics in capturing social preferences related to intergroup conflict, and improve predictions of individuals' conflict contributions. Notably, we find that parochialism varies for different outgroups-an unanticipated result that challenges existing theoretical models. Our work provides new tools for studying individual- and group-level social preferences in intergroup relations and presents novel evidence to inform substantive theoretical improvement.