A fundamental cognitive bias increases support for violence against civilians in Gaza and Israel

War is often characterised by indiscriminate violence against civilians. It is critical to understand why ordinary people might support acts such as ethnic cleansing or genocide, as popular support facilitates campaigns of indiscriminate violence. Theory suggests that support may rest on ideologies and narratives that portray the target group as less than human and threatening, thereby creating a moral mandate for killing. However, there has been little empirical study of these mechanisms during outbreaks of extreme violence. Here we report studies carried out in an ongoing campaign of violence in Gaza and Israel (n = 2462), showing that alongside such narratives, popular support for violence against civilians derives from a common cognitive error we term the hate-motive bias: the tendency to overestimate hate motives, and underestimate defensive motives, of outgroup aggression. Hate-motive bias predicted support for various forms of violence against civilians even while statistically accounting for other predictors such as ideological orientation, dehumanisation of outgroups and perceived threat. Our results suggest that a common attribution bias may contribute to ordinary citizens supporting behaviour they might typically deplore. Efforts to correct this bias may offer a behavioural science lever for reducing popular support for violence against civilians.