Prevention is political: political party affiliation predicts perceived risk and prevention behaviors for COVID-19

As of November 23, 2021, there were over 43.6 million SARS-CoV-2 infections in the US [1, 2] and more than 700,000 COVID-19 deaths [1, 3]. Both the infection and the death rates likely substantially underestimate the true population impact given that there are nearly 300,000 excess deaths in the US since the start of the pandemic [4, 5]. In addition to the mortality burden, there are long-term cardiac, respiratory, and other health consequences for COVID-19 patients [6, 7]. In many areas of the US, infection rates have shifted over time with changes in policy-mandated preventive strategies, decreasing when prevention-oriented policies were in place and then increasing again as prevention strategies were eased [8].

A centerpiece of public health strategies to control infectious disease spread is either recommending or mandating that members of the public engage in protective behavior to prevent disease transmission. In the case of COVID-19, staying 6 ft apart, wearing masks, working remotely when possible, avoiding public gatherings, and other strategies have been endorsed and communicated to the public by CDC, FEMA, WHO, and other national and international public health agencies [9‐11]. These preventive strategies are effective at slowing the rate of COVID-19 infection [12‐14]. However, for infectious disease prevention behaviors to be effective, they must be undertaken consistently by a sufficient proportion of the population to slow transmission [15‐17].

Most theoretical and empirical treatments of risk perception focus on individual cognitive and reasoning processes as the key determinant of a person’s risk perceptions [18, 19]. In contrast, the social amplification of risk framework [20, 21] describes the process by which scientific evidence, the ways in which people obtain information (e.g., news media), and political and cultural forces shape how individuals interpret and prioritize health risk information. According to this framework, perceptions of risk can be amplified or attenuated through social processes that influence: 1) the availability of risk information (e.g., via media and political sources) and 2) society’s response to the information (e.g., discourse about the veracity of the risk information in the media and interactions with cultural and peer groups [20]).

Influential communicators such as social/activist organizations and opinion leaders among social groups or organizations are key sources of information and can influence discourse surrounding risk information [21]. For example, political groups, parties, and leaders are prominent examples of influential communicators who may shape how their members and affiliates interpret risk information. These amplification and attenuation processes through influential communicators likely contribute to an alignment between, on one hand, people’s political affiliation and their underlying values and, on the other hand, their perceived risk.

During the pandemic, politicians in multiple countries have sought to control the amount and kind of information the public received about COVID-19 risk as well as actively disputing scientific discourse about the risk. Residents of several countries received messages from political leaders that minimized risk and raised doubt about preventive strategies. For example, the Prime Minister of Great Britain, Boris Johnson, publicly announced that he would not engage in social distancing and, specifically, would continue to shake hands [22] less than a week before the British government began to plan policy strategies to prevent transmission [23]. In Brazil, President Jair Bolsonaro made public statements that minimized the perception that the virus posed a risk, comparing it to the flu [24] and actively arguing against preventive policy strategies that were being put in place in Brazilian cities [25, 26]. Similar minimizing and contradictory statements can be found from the leaders of other countries [27].

In the United States, the country of focus in the current study, then President Donald Trump regularly made public statements that downplayed the threat posed by the virus [28, 29] in terms of both severity, comparing it to the flu [30], and the number of cases and deaths [31]. President Trump also downplayed the importance of preventive actions to protect against transmission, including arguing against mask mandates [32], pushing for early easing of stay at home and business closure orders [33], and holding public gatherings in spite of social distancing policies in place [34]. The President admitted to being motivated to downplay the risk despite receiving clear warnings about the severity and seriousness of the spread of COVID-19 ([35] p. xviii).

In addition to risk perception, individuals’ decision making about COVID-19 prevention behaviors has taken place in the context of a complex, saturated, fast-moving information environment, with multiple and sometimes contradictory messages from traditional media, social media, and government messaging [36]. There is evidence that political messaging has an influence on behavior for individuals who support the politician conveying the messaging. Specifically, a study of the impact of President Trump’s anti-vaccination messaging found that the exposure to the messages negatively impact vaccination engagement intentions, but only on the part of voters who voted for him [37].

Based on the social amplification of risk framework, one would expect that the selective communication of risk information by politicians and the conflating of politics and public health in media messaging would influence the public’s perceptions of risk and their decision making about behavioral strategies to mitigate risk. Moreover, given the current US phenomenon where some news media outlets provide partisan lenses on issues, one would expect this effect to be exacerbated as news media “amplify” the messaging about risk perception [38]. Thus, one would predict that political affiliation would, by affecting exposure, attention, processing, and response to SARS-CoV-2/COVID-19 messages, influence risk perception and behavior.

Given the empirical evidence and plausible, theory-derived mechanisms for the role of social amplification in risk perception and decision making about risk reduction messages, we hypothesize that political party affiliation will relate to engagement in preventive measures for SARS-CoV-2/COVID-19, such that – given the predominant minimization themes in President Trump’s statements, those who identify more strongly as Republicans will be less likely to engage in preventive measures and will perceive less risk relative to those who identify more strongly as Democrats.

We examined whether Americans’ political party identification is related to perceived risk of SARS-CoV-2 infection, the severity of COVID-19 illness, and engagement in a range of preventive behaviors. Although there have been examinations of how political affiliation relates to support for preventive measures against COVID-19 (e.g., policy measures to restrict gatherings, mask wearing), our approach adds to this literature in two ways. First, we utilize a nuanced, quasi-continuous assessment of strength of party affiliation, allowing for an understanding of how strength of political beliefs affect responses to COVID-19. Second, we examine the effects of political partisanship on risk perceptions for COVID-19 infection in addition to support for preventive measures, adding to the understanding of how politicians’ risk minimizing messages might affect individuals’ perceptions of the risks posed by COVID-19.

We conducted a cross-sectional, population-representative national survey of US adults. The study was reviewed and determined exempt by the University of Kentucky Institutional Review Board.

Six hundred eighty-three adults drawn from a representative sample of the Knowledge Panel were invited to participate. The survey was completed by 410 participants (completion rate of 60% [48]). The weighted demographics of the sample, per sampling design, mirror those of the US population. Of particular importance for this paper, 17% of the population identified as strongly Republican, 10% as weakly Republican, 20% as Independent/Lean Republican, 15% as Independent/Lean Democrat, 15% as weakly Democrat, and 23% as strongly Democrat.

Political party affiliation was consistently related to the degree of risk people perceived from SARS-CoV-2 infection and COVID-19 disease. This finding was consistent across multiple components of risk perception and for both perceptions of the person’s own risk and perceptions of risk to other people. Moreover, party affiliation was associated with the number of preventive behaviors a person engaged in and, for most the individual preventive behaviors, with the probability that a given person engaged in each individual behavior.

These findings are consistent with both laypeople’s observations of differences in behavior and with press reports over the course of the pandemic. In addition, they mirror non peer-reviewed reports [49]. There has been parallel work at a community/geographical level of analysis showing that the overall political makeup of counties in the US influences the degree to which residents restricted mobility during stay at home orders [50] and engaged in social distancing [51]. Importantly, to our knowledge this paper represents the first report from a nationally-representative, probability-based sample to examine both risk perceptions and preventive behaviors, thus providing strong descriptions of the effects of partisanship on risk perceptions and preventive behaviors among the American public. In addition, the finding that partisanship affects both beliefs about risk and actions that are intended to mitigate risk highlights that complex and multifaceted ways in which public responses to the pandemic have become politicized. Also, we find that our effects extend past perceptions of one’s own risk to the little studied but, in the context of infectious disease, equally important construct of perceived risk to others. Finally, our work extends beyond previous studies focusing on single behavioral responses (e.g., social distancing [51], vaccination [52] to demonstrate a remarkably consistent partisanship effect across a range of preventive behaviors with very different levels of effort, social involvement, etc. Each of these characteristics increases confidence in the conclusion that politicization of the coronavirus pandemic and of preventive measures influences the way in which the American public’s responses to the pandemic have played out over the past several months.

Finally, although our findings are framed in the context of SARS-CoV-2/COVID-19, they are likely to extend to other scientific and public health issues that become politicized. As described in the introduction, President Trump’s anti-vaccination Twitter messages have had a demonstrable effect on childhood vaccination attitudes of his supporters [37]. Politicization of public health has occurred in the past in response to such now commonplace measures as motorcycle helmet laws, milk pasteurization, water fluoridation, and mandatory seat belt use in automobiles [53‐55]. Thus, the implications of these findings extend past the current pandemic to the more general environment around science, public health, public policy, and community action to protect health and well-being.