When can individual partisanship be tempered? Mass behavior and attitudes across the COVID-19 pandemic

Abstract How do partisan differences in mass behavior and attitudes vary across contexts? Using new individual-level panel data on the COVID-19 pandemic from 54,216 US adults between March 2020 and September 2021, we consider how partisan differences vary according to the personal costs and benefits of behaviors, their public symbolism, and elite-level policy choices. Employing various panel data estimators, including difference-in-differences, we evaluate how partisan gaps evolve across changes to the political and health contexts, including the national vaccine rollout, individual vaccination status, and within-state policy variation. We find partisan divides are substantial even in (ostensibly) apolitical domains, although they are tempered by higher net personal costs to actions, lower public symbolism, and elite policy choices that counter national party cues.