“…pre-analysis plan designated only demographics as variables used to pair respondents in the matched design. Rather than overlook other measured dimensions that could plausibly explain away perceived social norms/vaccination intentions relationships, we chose to add additional measures from the surveys-speci cally, trust in science (Koetke et al, 2021; Plohl & Musil, 2020; Stosic et al, 2021), trust in government (Han et al, 2021; Pagliaro et al, 2021), religiosity (Milligan et al, 2021; Olagoke et al, 2020), COVID-19 knowledge (Ning et al, 2020),and subjective assessments of that knowledge(Motoki et al, 2021). FigureS1shows that the patterns are the same without these additional variables, with one exception: Democrats displayed a strong relationship between perceived social norms of co-partisans and their own vaccination intentions that vanished once respondents were further matched on trust in science, trust in government, religiosity , COVID-19 knowledge, and subjective assessments of COVID-19 knowledge.…”