Trumpism, climate and COVID: Social bases of the new science rejection

views have been well studied, we adapt those results into specific hypotheses, tested here using new data on both ACC and COVID-19 vaccine rejection. H1: Trump support will be greater among older respondents, and those who identify as white, evangelical Christian, conservative and Republican. H2: Rejecting the reality of ACC, and personal rejection of COVID-19 vaccination, will both be more common among ideologically conservative, Republican and Trump-supporting individuals. H2a: Support for ex-president Trump thus can be modeled as an intervening (mediating) vari- able, related to ideology, party and other background characteristics, with both direct and indirect effects on ACC and vaccine rejection. H3: Individuals who self-identify as white or as evangelical Christians (traits often linked to conservative identity) will be more likely to reject ACC and vaccination. H4: Education×ideology interaction effects, widely observed in ACC studies, will also affect vaccination responses in similar ways: ideological divisions on both topics widening with education. H5: Partisan divisions on Trump support, ACC and vaccine rejection will be wider among respondents who report having mostly same-party friends (friends×party interaction). H6: After controlling for the background, sociopolitical, interaction and identity effects noted above, Trump support, rejection of ACC and rejection of COVID-19 vaccination will also be more common among individuals who are receptive to unrelated conspiracy beliefs.

The next section describes a new survey dataset, and the analytical methods used to test these hypotheses in a unified multivariate model. 4. Data and methods 4.1 POLES 2021 survey Data analyzed here are from the Polar, Environment and Science survey, conducted in sum- mer and early fall 2021 (POLES 2021). This representative US survey is well suited to the research questions posed above because it carried both climate and COVID-19 items, along with a range of different indicators for sociopolitical identity and respondent background. The survey also carried a number of conspiracy-belief items [41], several of them on topics that have not been politicized and are unrelated to climate or COVID. Other POLES 2021 ques- tions on science-related topics replicated those asked five years earlier on the POLES 2016 sur- vey [14,81,82]. In addition, the 2021 survey carried new questions reflecting changes in the US political context. Finally, and necessitated by the pandemic-driven shift from telephone inter- viewing to online data collection, the newer survey carried attention-check items to screen out thoughtless respondents—such as those who answered all the questions too quickly, or “straightlined” their agreement or disagreement with blocks of incompatible statements. Research protocols were approved by the Institutional Review Board for the Protection of Human Subjects in Research at the University of New Hampshire (IRB-FY2021-38), and the online questionnaire distributed in two stages: June/July and September/October 2021. After

quality screening, we obtained 1,134 valid completions. Sampling for this survey, organized by Qualtrics, aimed for a nationally representative profile of US adults with respect to age, gender, race, education and political party. PLOS ONE Trumpism, climate and COVID PLOS ONE | https://doi.org/10.1371/journal.pone.0293059 January 10, 2024 4 / 20