Abstract
Given partisanship’s powerful role in the U.S., a wealth of research has sought to identify its determinants. Toward this end, scholars have long highlighted (1) social identities and (2) group sentiments toward marginalized groups as two factors that shape citizens’ partisanship. Yet recent research has begun to “flip the causal arrow,” exploring how partisanship can influence social identities. In this Letter, we investigate one mechanism by which partisanship might also affect group sentiments. Using a well-powered, pre-registered experiment featuring historically marginalized groups, we test whether manipulating information about these groups’ voting behavior in presidential elections can affect partisans’ sentiments toward these groups. Across multiple outcome measures, we consistently find small, non-significant effects for each target group, irrespective of the respondents’ party. These null findings are important, suggesting that, despite partisanship’s enormous influence, Americans’ sentiments toward marginalized groups resist “evolving” even when these groups increasingly help their party’s electoral prospects.