Abstract

The United States underperformed its potential in responding to the COVID-19 pandemic. The authors use original survey data from April 2020 to March 2022 to show that political partisanship may have contributed to this inconsistent response by distinguishing elites and citizens who took the crisis seriously from those who did not. This division was not inevitable; when the crisis began, Democrats and Republicans differed little in their viewpoints and actions relative to COVID-19. However, partisans increasingly diverged when their preferred political leaders provided them with opposing cues. The authors outline developments in party politics over the last half century that contributed to partisan division on COVID-19, most centrally an anti-expertise bias among Republicans. Accordingly, Republicans’ support for mitigation measures, perception of severity of COVID-19, and support for vaccines gradually decreased after the initial outbreak. Partisan differences also showed up at the state level; Trump's vote share in 2016 was negatively associated with mask use and positively associated with COVID-19 infections. Diverging elite cues provided fertile ground for the partisan pandemic, underscoring the importance of political accountability even in an era of polarization.

The COVID-19 crisis served as a stress test for the American body politic. By most measures, it failed. In October 2019, the Global Health Security Index at Johns Hopkins University ranked the United States as the country best equipped to handle a public health emergency (NTI/JHCHS 2019). But living up to this potential would, among other things, require citizens to embrace the best practices suggested by the nation's public health apparatus. Scientific consensus emerged in the early months of the pandemic, recommending the importance of staying at home, social distancing when in public, avoiding large indoor gatherings, and wearing a face covering when around others. Unfortunately, many Republican leaders, most notably President Donald Trump, downplayed the deadliness of COVID-19 and failed to embrace these recommendations. Moreover, the extraordinarily effective COVID-19 vaccines, which were developed in record time, also became the subject of partisan politics. As a result, the United States drastically underperformed the potential identified in the Johns Hopkins assessment.

A pandemic is bound to kill a substantial number of people, but according to Deborah Birx, a key member of President Trump's COVID-19 task force, many could have been saved. While the initial surge of 100,000 deaths that occurred by May 2020 was probably unavoidable, Birx estimated that at least 130,000 deaths that took place after that but before vaccines became widely available could have been mitigated by following guidelines and recommendations put forward by the public health and scientific communities (ORSSCC 2022). Failure to observe best practices followed a partisan pattern. We show that Republican and Democratic citizens developed very different perceptions about the pandemic and that red states and blue states differed markedly in their infection rates in the last months of the Trump presidency. After vaccines became available in spring 2021, political partisanship again emerged as the deepest fault line distinguishing those who got vaccinated from those who did not. The unnecessary carnage almost surely fell disproportionately on Republican partisans, although antimitigation decisions negatively affected Americans of all political stripes.

For those who do not study politics for a living, what unfolded must be vexing. How could so many people have chosen to follow political cues rather than medical and scientific ones in the face of a public health crisis? We start from a time-honored premise in public opinion and political psychology: when it comes to navigating a complicated set of circumstances, citizens often lack enough information or expertise to make good judgments on their own, so they look to their information environment for recommendations from sources they trust (Lenz 2012; Lupia and McCubbins 1998). Facing a novel coronavirus, people were bound to search for advice, but from whom? In the polarized political environment in which the pandemic occurred in the United States, Americans’ personal identities are increasingly tied up in their partisan identities (Mason 2018). That has encouraged them to rely increasingly on political leaders to inform their opinions and behaviors, even in areas where partisan cue-taking might seem suboptimal. After the pandemic became politicized by party, the result was a public health disaster.

Although the bulk of this article traces the implications of diverging elite partisan cues about the pandemic, we begin by providing some conjecture as to why Republican leaders broke from scientific orthodoxy on the pandemic. Being antiscience is not, historically speaking, a Republican position. In the 1950s, for example, a different Republican president, Dwight Eisenhower, threw the weight of the federal government behind a mass vaccination program that eventually eradicated polio in the United States. However, the base of that era's party is much different from the one today. The Republican Party has evolved from one centrally motivated by a philosophy favoring small government to a party centrally motivated by a philosophy favoring the maintenance of time-honored traditions. As Galileo Galilei might attest, the latter is often accompanied by a marked skepticism about science.1

Although our insights about the parties suggest many Republicans in the electorate were disposed to doubt the scientific consensus on COVID-19, our results indicate that a partisan pandemic was still not inevitable. When leaders of both parties reflected the beliefs of public health experts, as they did in the early months, the attitudes of Republicans and Democrats in the electorate did not differ much. Only after Republican leaders started to downplay the health risks associated with COVID-19 did partisan differences on mitigation widen into chasms. Even then, not all Republicans were equally susceptible to leader cues, with those who feared becoming seriously ill from the virus having mitigation preferences much like those of Democrats.

Last, we turn to data gathered during the period after vaccines became available, which reveal a similar and deadly pattern. Among Republicans, a spillover effect has developed from attitudes about COVID-19 vaccines in particular to attitudes and behaviors about vaccination in general. As we begin, let us be clear that our intent is not to definitively test the application of these theories to the COVID-19 crisis, but rather to use public opinion evidence to help explain the public's response.

Partisanship in Washington and in the Electorate

The prevailing wisdom in political science is that citizens’ sense of partisan identification exerts an exceptional influence on their political opinions (Gerber and Huber 2010; Lenz 2012). No force is more important, no variable is more central, to citizen opinion and behavior (Campbell et al. 1960). Because most Americans know less about political matters than one might ideally hope and expect, they tend to use shortcuts to get by. The most common shortcut is to look to their favored party's leaders to deduce where they should stand and what they should believe on issues of the day (Converse 1964). These long-standing insights are arguably more relevant today than at any other time. Polarization has made Americans’ party identities central to their personal identities (Mason 2018), supercharging the leader-follower dynamic and ensuring that COVID-19 took place in a political context in which citizens are even more likely to rely on party cues (Druckman, Peterson, and Slothuus 2013; Leeper and Slothuus 2014).

Ironically, COVID-19 is the type of issue that actually had the potential to break the grip of partisanship. Many political issues, such as affirmative action, gender equality, and LGBT rights, are “easy” issues for citizens to engage with because they have a long history of salience or they contain easily accessible symbolic referents that citizens can use to develop a position (Carmines and Stimson 1980). “Harder” issues, such as tax policy, typically require a higher degree of content knowledge (Johnston and Wronski 2015), something not often found in abundance among average citizens (Converse 1964). COVID-19 might be considered an especially hard issue; it was a new type of risk, it developed quickly, and even public health experts needed time to assess and understand the situation. As such, it was the kind of issue on which Republicans in the electorate might have followed scientific expertise had their leaders encouraged them to do so (see, e.g., Leeper and Slothuus 2014 on contextual elements constraining the use of party cues). And for a time, they did. As with other Western democracies where citizens supported their governments’ efforts to combat the pandemic regardless of party (Mordecai and Connaughton 2020), our evidence suggests that Republicans and Democrats alike took cues from public health scholars and the Centers for Disease Control and Prevention (CDC) at the beginning.

As the pandemic dragged on into the summer, the information environment evolved. On the one hand, a scientific consensus hardened around the importance of a range of mitigation strategies. On the other hand, Republican political leaders began to diverge from that consensus. To understand why, it is useful to remember that doubting science and being Republican do not necessarily go together, historically speaking. Today's GOP is much more populist than has been typical. Not long ago, the country club—a symbol of elitism—was an exemplar of Republican identification. While it is true that the United States has always had an anti-intellectual streak (Hofstadter 1963), that tendency has not always found its home in the Republican Party. Rather, the most noteworthy populist-minded politicians of America's first 200 years include Andrew Jackson, William Jennings Bryan, Huey Long, and George Wallace—all Democrats.

The Republican turn to populism over the past 20 years is important because party followers will not blindly follow any cue put forward by a party leader. If, for example, Republican politicians one day endorsed affirmative action and abortion rights, most Republicans in the electorate would not follow. Opposition to both are central to Republicans’ understanding of politics and society. Partisans will more closely follow leaders when their stances are broadly consistent with why they favored their party in the first place. Through the first two-thirds of the 20th century, the line that divided the parties was best understood in terms of differences in opinion on government spending and regulation (Ellis and Stimson 2012). Most Democrats wanted government to do more of both, whereas most Republicans wanted it to do less, producing parties divided by social class and income. However, a class divide does not suggest a clear connection to belief in scientific expertise, which is likely part of the reason why the Eisenhower administration embraced the polio vaccine. One might imagine that those in higher income brackets—as Republicans were until relatively recently—would be more likely to embrace scientific expertise. It is also noteworthy that Republicans of that era purged conspiracy-minded groups like the John Birch Society from their ranks at the behest of conservative leaders such as William F. Buckley and Ronald Reagan (Mulloy 2014). Conservatism then is different from conservatism now.

Rather than the parties being divided over opinions on the size and reach of government, today's divide is centrally organized according to stances on the wisdom of existing traditions and hierarchies (Hetherington and Weiler 2018). This is at least partially the result of the rise of a range of culture issues in the last decades of the 20th century. On issues including race, ethnicity, gender, or sexual orientation, Democrats embraced groups that had been disadvantaged by existing traditions and occupied the bottom of social hierarchies. Republicans took the opposite position, often basing their opposition to change on their belief in traditional religious tenets. In addition to creating more racial, ethnic, and gender polarization in voting than ever before, these disputes between the parties created a divide anchored more in education than in income.

This change coincided with evolving views about the value of scientific expertise. From the 1970s (when survey organizations first started to ask Americans about their confidence in science) through the 1990s, Republicans consistently expressed more confidence than Democrats did. Those differences reversed over time, as the basis for party identification shifted from beliefs about size of government to beliefs about race and culture. Our survey research shows that in 2020, 58% of Democrats believed the benefits of scientific research outweighed any harmful effects, compared to only 40% of Republicans. This makes sense, given the recent changes in party identification; scientific advances over centuries have often run counter to traditional understandings of how the world should be ordered. We suspect the same impulse that fuels the Republican anti-elitist, anti-university rhetoric that is so common around election time is precisely what led Republicans in the electorate to follow party leadership in questioning the validity of scientific assessments and expert recommendations made during the pandemic. Put more directly, President Trump's skepticism about the seriousness of COVID-19 tapped into a tendency that already existed among his copartisans.

Consensus, Then Dissensus in Pandemic Perceptions

It is hard to remember now, several years past the beginning of the pandemic, but the political environment when the crisis first took hold was much different from how it was in the months and years that followed. Revisiting that time suggests that the fracturing of the country and the resulting carnage the United States experienced was not inevitable. President Trump reluctantly championed the national lockdown in March 2020, only later to refrain from embracing the steps that scientists and public health professionals recommended. In fact, both Republican and Democratic leaders were initially sober about the dangers posed by COVID-19. Given the stakes for the economy, neither side was enthusiastic about instituting lockdowns, shutting nonessential businesses, or upending their constituents’ lives. However, the grisly outcomes that occurred first in China, then in Italy, and next in New York City caused leaders to react forcefully. Reflecting the bipartisanship of the moment, Congress quickly passed—and President Trump signed—a massive aid package (the CARES Act) to help soon-to-be struggling businesses and citizens.

By the time infections had begun to spread widely across the country, however, the moment of elite cooperation had passed. President Trump was a central reason why. Whether he was trying to quell panic or truly believed his rhetoric, his public statements were often at odds with reality and aimed at downplaying risk. During the shutdown in March, he expressed his belief that suicide deaths “definitely would be far greater in numbers than the numbers that we are talking about” relative to COVID-19 (Cathey 2020). In May, he said COVID-19 would “go away without a vaccine,” and in June he claimed the pandemic was “fading away” (Wingrove 2020). On Independence Day, he pronounced 99% of cases “totally harmless” (Rabin and Cameron 2020).

The president has the country's biggest megaphone, but Trump was not alone among Republican officeholders in minimizing the threat. Adolph and colleagues (2021) found that Republican governors were several times less likely than Democratic governors to impose policies such as mask mandates, and even when they did, the mitigation efforts tended to come a month later on average than in Democratic-run states. Instead, Republican governors seemed to race one another to reopen their states. In mid-April, Governor Greg Abbott of Texas said he was “working on very aggressive strategies to make sure Texas was first at getting back to work” (CBS News Texas 2020). Governor Brian Kemp of Georgia was even more aggressive than Trump himself in reopening his state's economy in the last week of April (Higgins-Dunn and Feuer 2020). Not all Republican governors were similarly invested in “getting back to normal”; Charlie Baker of Massachusetts and Larry Hogan of Maryland approached mitigation more like Democratic governors did. It is noteworthy that their states generally tend to vote for Democrats.

With Democratic officeholders at least publicly taking the pandemic seriously throughout this same period, Americans’ information environments began to diverge. This is important, especially in the present polarized environment, because citizens tend to seek out news and information that align with their political predispositions. As evidence that this process was at work during the COVID-19 crisis, citizens who paid more attention to politics—and were thus best positioned to reflect cues from party leaders—were more divided when it came to attitudes toward COVID-19 than those who paid less attention (Kam and Sides 2020).

To evaluate the evolution of COVID-19 attitudes and behaviors over time, we contracted with Qualtrics in April 2020 to collect survey data on American adults, with the sample calibrated to meet US Census Bureau benchmarks on race, gender, education level, and income. We fielded additional survey waves—each featuring more than 2,000 respondents—throughout the following two years. Our data show a consistent pattern: early waves reveal general agreement in the mass public, similar to that observed among elites. But as spring turned to summer in 2020, Republicans’ opinions and health behaviors began to reflect their party elites’ message that public health professionals had overblown the pandemic's seriousness. The result was sharp disagreement about how the government should address the pandemic, how to protect oneself from illness, and even whether COVID-19 represented a legitimate health threat.

Consider, for example, partisans’ attitudes toward mitigation measures, displayed over five waves in figure 1. Overwhelming majorities endorsed a range of extraordinary restrictions that government might need to impose to limit viral spread. In April 2020, 96% of Democrats and 88% of Republicans endorsed stay-at-home orders, indicating that regardless of partisanship, Americans supported one of the greatest (and most invasive) tools that government could use to decrease infections. By March 2021, that percentage had dropped to 42% among Republicans, with the difference between Democrats and Republicans increasing from 9 to 40 percentage points. The same basic pattern of change emerged for support for business closures and mask mandates, which we also display in figure 1. Support for mask mandates is particularly noteworthy: Republicans’ endorsement of such restrictions decreased from 80% in April 2020 to 66% in March 2021, while Democrats’ support for them actually increased slightly during the same time frame, from 92% to 94%.

These sharp party-line differences and declines in support for pandemic mitigation policies among Republicans paint a stark picture of the politicized response to COVID-19. It is important to note, however, that Republicans did not respond to party cues as a monolithic bloc; there were important variations in their proclivity to downplay the pandemic. Age was likely one factor that led some Republicans to support mitigation measures they otherwise would have been reluctant to embrace. COVID-19 tended to hold the most severe consequences for the elderly. This heightened risk encouraged Americans older than the age of 65 to wear face coverings at higher rates than younger Americans, even in rural areas (Haischer et al. 2020), which skew heavily Republican.

Anxiety also had a particularly dramatic effect. Fear of becoming seriously ill from COVID-19 led Americans of all partisan colors—but especially Republicans—to seek out new information about the pandemic, hold more accurate information about COVID-19, and support policies to mitigate its impact (Mehlhaff et al. 2024). In fact, Republicans who reported being “very concerned” about COVID-19 looked nearly identical to Democrats in their support for mitigation policies, a position they could only take by resisting the influence of party cues.

In seven waves, we asked respondents to report how concerned they were about becoming seriously ill from COVID-19. Figure 2 plots the percentage of Democrats and Republicans in each wave attesting that they were “somewhat” or “very concerned” about becoming seriously ill. Here, too, the distance between parties increased over time. In April 2020, about 63% of Republicans expressed being fearful of an illness brought on by COVID-19, only 16 percentage points fewer than Democrats. Because the virus had not yet spread to red states or the redder parts of blue states, the difference had some basis in fact. However, as the pandemic worsened and the virus spread into red states and red areas within blue states, Republicans actually expressed less concern about getting seriously ill. Approximately one year later, in March 2022, the gap between parties had increased to 28 percentage points.

In general, both Democrats’ and Republicans’ fears of becoming seriously ill declined over the course of our panel survey, but one exception is notable. Both parties saw a brief spike in concern in October 2020. In fact, this wave represents the highest degree of concern among Democrats, higher even than at the beginning of the pandemic in April, when uncertainty and anxiety were especially high. We can rule out increased caseloads as the cause of this temporary increase in fear; records from the CDC show that around 7,000 people died from COVID-19 on a weekly basis in October 2020. This is a massive loss of life, to be sure. But it is dwarfed by the 17,000 weekly deaths seen in April of the same year. If fear of COVID-19 was exclusively a function of pandemic severity, partisans’ reported fear in October would be much lower. Instead, we attribute this spike in anxiety to the combination of two different factors. First, the presidential election campaign reached a fever pitch in the same month—an election in which COVID-19 was the major issue. Although our data cannot speak directly to this hypothesis, it is noteworthy that the election season and the fear of COVID-19 appear to track each other closely. Second, President Trump tested positive for COVID-19 and fell ill at the beginning of October. That likely acted as a sign indicating cause for alarm, even to Republicans who might not otherwise have taken the risk of illness seriously.

One reason Republicans expressed less concern about getting seriously ill is that they increasingly evaluated the virus as being more like a flu rather than as a potentially deadly disease. We asked respondents to assess, on a 10-point scale, the degree to which they thought people suffered as a result of catching COVID-19. Figure 3 depicts the percentages of Republicans and Democrats responding that they thought COVID-19 did not represent a serious threat (response options 1–4), was more “like a bad flu” (5–7), or portended “serious illness, with risk of death” (8–10). A noticeable partisan gap is already visible in the June 2020 wave. In this survey, 61% of Democrats perceived COVID-19 as causing a risk of death, about 20 percentage points greater than the percentage among Republicans. By October of that same year, the proportion of Democrats perceiving COVID-19 as life-threatening had risen to 71%, whereas the proportion of Republicans with that perception had decreased to 36%, resulting a gap of 35 percentage points. It is remarkable that, after more than 200,000 Americans had died, the percentage of Republicans who viewed COVID-19 as potentially life-threatening actually fell by 5 percentage points. At the other end of the scale, a vanishingly small subset of Democrats believed that COVID-19 was “not serious”—only 3% in October. Meanwhile, five times as many Republicans—15%—expressed this belief.

One reasonable explanation for these diverging patterns in policy support, anxiety, and severity perceptions is that Republicans and Democrats simply experienced different pandemics. Republicans tend to live in lower-density, more rural areas, while Democrats are more likely to live in high-density cities, where an airborne virus is more likely to spread unabated. Our data suggest, however, that this explanation holds little water. In fact, Democrats and Republicans were strikingly similar in their pandemic experiences. We began asking our respondents in April 2020 whether they personally knew someone who had been infected by COVID-19 in the past month, and we continued to do so throughout the year. The party differences were 2, 6, 6, and 3 percentage points across the four surveys. While the divide between Republicans’ and Democrats’ attitudes about the pandemic widened, their experiences remained quite similar.

Taking (In)Action to Prevent Infections

As Republicans adopted partisan cues and came to see COVID-19 as less threatening, their personal behavior began to reflect those attitudes. In June and October 2020, we asked Americans the degree to which they were engaging in certain activities, with the questions focusing on behaviors that might increase the chances of contracting and transmitting COVID-19. Some of the activities were essential tasks, such as going to the grocery store and going to work. Others involved more personal choices, such as going to a restaurant or shopping in a mall. Figure 4 reveals that, in June, Democrats were about 10 percentage points more likely than Republicans to say they were doing these things “a lot less” than they were before the pandemic.

By October, despite skyrocketing infection rates, partisan differences more than doubled when it came to personal choices such as avoiding restaurants and shopping malls. Both Democrats and Republicans became wearier of staying home. Only about 60% of Democrats reported they were now going to restaurants and shopping malls “a lot less” than before the pandemic; those percentages had been almost 15 percentage points higher in June. However, the drop among Republicans was even more pronounced. For example, only 40% of Republicans reported going to restaurants “a lot less” in October. Accordingly, the difference between parties more than doubled in four months, largely because of Republicans returning to their prepandemic routines.

Not only did Republicans start to return to their normal lives more than Democrats did, their willingness to take actions to fight COVID-19 also declined sharply. In six waves extending from April 2020 to August 2021, we asked respondents to consider three health-related behaviors that could decrease the transmission of illness, especially COVID-19: staying six feet away from other people, wearing a mask when indoors in public, and washing one's hands frequently. Figure 5 displays the percentages of Republicans and Democrats in each wave reporting that they had engaged in these behaviors “very often” during the previous two weeks. Similar to the results presented in figures 1 through 3, partisans differed little in their behaviors in the early days of the pandemic; April 2020 estimates for all three behaviors are never more than 10 percentage points apart. Democrats always engaged in mitigation behaviors more than Republicans did, but this difference was not especially large at first.

By late spring, differences between Democrats and Republicans in the electorate began to develop. Similar again to results in previous analyses, growing partisan differences were largely attributable to especially steep decreases in mitigation actions among Republicans. When the pandemic started, Democrats stated that they engaged in social distancing at a rate of 80%; Republicans were just behind at 73%. By August 2021, the percentage of Democrats practicing social distancing had decreased only 9 percentage points, to 71%, but the percentage of Republicans doing so had taken a sharp nosedive to 39%. The gap between the parties widened from 7 to 32 percentage points during that period, almost entirely the result of changes in Republicans’ willingness to stay six feet apart. The data on mask-wearing tells a similar story, with the gap between parties doubling from 10 to 20 percentage points. The increase was hardly the result of movement on the Democratic side. In fact, Democrats wore masks at a higher rate in August 2021—almost a year and a half into the pandemic—than they did at the outset.

These differing levels of commitment to mitigation strategies likely opened the door to a deepening of the crisis. Figure 6A illustrates the relationship that developed between the percentage of people in a state who voted for President Trump in the 2016 election and the percentage of people wearing a mask in that state. The relationship is incredibly strong, with a correlation of −0.85. More than 90% of people wore masks in dark blue states such as Massachusetts, Hawaii, and Maryland. In contrast, fewer than 75% wore them in dark red states such as Wyoming, North Dakota, and Idaho. Mask-wearing in swing states such as North Carolina, Georgia, and Pennsylvania clusters around 85%.

Because less mask-wearing results in more viral spread, America's bifurcated response also reveals itself in the number of confirmed cases at the state level. States that voted more heavily for Trump not only masked less but also tended to lead the nation in infections after June 1, a date we chose because states struggled to identify best practices before then. As figure 6B shows, the relationship is again strong, with a correlation of 0.65. Blue states as regionally diverse as Vermont, Maryland, and Washington had fewer than 5,000 infections per 100,000 residents. Regionally diverse red states, such as the Dakotas, Tennessee, and Utah all had more than 8,000 infections per 100,000 residents.

This analysis is purely correlational, and caseloads and mask-wearing habits might be explained in part by a wide variety of factors. One such factor might simply be the average age of a state's residents. Older Americans were more likely to contract COVID-19 and to develop serious illness as a result, which could partially account for the higher caseloads in Republican-leaning states, which also tend to have older residents. In addition, as we previously explained, the antimitigation rhetoric of many Republican governors likely influenced the attitudes and behaviors of copartisans in their respective states. Moreover, Republican governors’ tendency to open their states sooner and spurn mask mandates throughout the pandemic likely affected the health of their states’ residents. Caseloads and mask-wearing are themselves related; lower masking rates produce higher caseloads. Even so, it is striking that each variable tracks closely with the relative division of Republican and Democratic votes. Elite messaging is likely to have played a large, if not singular, role (Gadarian et al. 2022).

A Shot in the Partisan Arm

By the time infections were spreading largely unchecked in the United States, it became clear that developing a safe, reliable vaccine was likely the only feasible means to bring an end to the pandemic. Thanks to advances in mRNA technology and a scientific community working overtime for months on end, the United States produced such a vaccine a mere nine months after the first lockdowns were instituted—a record-setting development timeline. The Food and Drug Administration issued the first emergency use authorization for a COVID-19 vaccine in December 2020, and the general public gradually began to get vaccinated over the ensuing months.

At first, it seemed as if Americans might be enthusiastic about vaccination. In June 2020, still early in the pandemic and months before a vaccine was developed, 63% of Democrats and 48% of Republicans stated that they believed others should get vaccinated if a vaccine was successfully developed. Democrats were more supportive of vaccination, but with about half of Republicans similarly expressing support, these might have been interpreted as encouraging numbers at the time. At the height of election season that same year, however, both partisan groups were less supportive of vaccination than they had been over the summer. In October, 55% of Democrats and only 37% of Republicans believed others should get vaccinated if a vaccine was developed.

This decrease in support across the board might reflect the changing information environment. In March, the idea of a vaccine was still rather abstract. Once multiple vaccines began final clinical trials later in the year, concerns about potential side effects and the speed at which the vaccine was developed were starting to become more prominent. However, the decrease in support for vaccination likely also reflects the rhetoric unleashed by both presidential candidates throughout the campaign, in which COVID-19 was the most salient issue.

Absent a blanket legal mandate, public messaging was perhaps the most valuable tool the federal government could use to encourage Americans to get vaccinated. In fact, Larsen and colleagues (2023) used a field experiment to show that provaccine messaging from President Trump increased vaccine uptake in Trump-voting counties. Although his delivery was not always the most convincing or reassuring, President Trump actually maintained a supportive stance toward vaccination, calling the vaccines “one of the greatest achievements of mankind” and insisting that “if you take the vaccine, you're protected” (Finn 2021). At the same time, however, he was loath to impose any sort of vaccine mandate during his final months in office or to support it after President Biden's inauguration, urging Americans to “forget about the mandates; people have to have their freedom” (Oppenheim 2021).

Whether because of Trump's rhetoric or in spite of it, the partisan gap in vaccine uptake was substantial. Democrat-leaning counties consistently had higher vaccination rates than Republican-leaning ones, and this gap widened over time (Ye 2023). Once individuals completed their initial vaccine series, they were equally likely to get a booster shot, regardless of their county's partisan lean (Kates, Tolbert, and Orgera 2021). Nevertheless, the gap in baseline vaccine uptake likely had devastating effects on health outcomes. Data beginning in May 2021, the month in which vaccines became widely available, show that counties that voted for Trump in 2020 had a death rate 2.73 times that of counties that voted for Biden. This disparity in death rates increases as counties’ Trump vote share increases, and the effect persists even after controlling for age (Wood and Brumfiel 2021).

To whom did Americans look when seeking advice on vaccination? Could the source of partisans’ information explain the difference in vaccination rates? We presented respondents with an assortment of entities—President Trump, President Biden, Anthony Fauci, their own doctor, and the scientific community in general—and asked how much confidence they had in each entity's vaccine-related advice. Figure 7 displays the percentages of Democrats and Republicans feeling “a great deal” of confidence in advice from each entity. We first asked this question in March 2021, just as vaccines were starting to become available to the general public, and again in August 2021, when anyone who wanted a vaccine could easily get one. Notably, responses do not differ much across the two waves.

Republicans put the most stock in advice from two sources: President Trump and their own doctor. Forty-four percent of Republicans expressed “a great deal” of trust in President Trump's vaccine-related advice, and about the same proportion said they trusted their own doctor. Democrats trusted their own doctor at rates about 15 percentage points higher than Republicans, at about the same rate at which they trusted President Biden. They expressed even more trust in Anthony Fauci, a key member of the White House COVID-19 task forces under both Trump and Biden. Perhaps unsurprisingly, Republicans expressed little trust in these individuals, with their support never surpassing 20%. Similarly, a majority of Democrats expressed trust in the scientific community in general, while Republicans’ support lagged behind. This may reflect more fundamental disagreements than merely COVID-19 messaging. Our data show that 58% of Democrats believe the benefits of scientific research have outweighed its harmful effects, while only 40% of Republicans believe the same. Twenty-nine percent of Republicans also believe “we depend too much on science and not enough on faith,” again revealing a preference for using traditional religious tenets to guide society rather than the views of scientific experts.

We also asked our respondents a series of questions about how they perceive the efficacy of the vaccines, regardless of whether they got vaccinated themselves. Figure 8 shows how Republicans and Democrats answered two of these questions in March 2022. Even a year after vaccines became available, 16% of Republicans still believed the vaccines were more dangerous than COVID-19 itself. Perhaps more astonishing, 37% of Republicans interpreted the continued infection of vaccinated Americans as evidence that vaccines were ineffective, rather than as evidence that vaccines can make illness less severe. Only 17% of Democrats agreed with this statement, representing a gap of 20 percentage points between parties.

To add insult to injury, evidence is emerging to suggest that Republican hesitancy surrounding the COVID-19 vaccines may be spilling over into general beliefs about vaccination. Take the seasonal flu vaccine, for example. Two sets of survey data gathered in late 2021 suggest an enormous partisan gap in flu vaccine uptake has developed since the pandemic began (see Enten 2021). Data collected by IPSOS revealed that 68% of Democrats said they had either received the flu vaccine or were very likely to get it, compared with just 44% of Republicans. A poll from the Kaiser Family Foundation showed the same pattern: 65% of Democrats but only 40% of Republicans said they had already gotten or would definitely get a flu shot. These numbers represent a staggering departure from data gathered before the pandemic took hold in the United States. In February 2020, an Associated Press–NORC survey found only a 4-percentage-point difference in flu vaccine uptake (58% vs. 54%). In 2016, the difference was only 2 percentage points, according to a Princeton Survey Research Associates poll (Enten 2021).

A partisan gap is also widening with regard to childhood vaccines. In surveys conducted in 2016 and 2019, Pew Research Center found only modest differences between Republicans and Democrats in their support for children being required to receive the MMR vaccines to attend public school. Whereas Republican support was 79% in the two prepandemic surveys, it plummeted to 57% in a March 2023 survey that asked the exact same question. Democrats’ support, in contrast, remained steady at 85% (Funk et al. 2023). Party-line differences also appear to be widening when it comes to reports of adverse vaccine effects. Motta (2023) finds a significant jump in parents reporting health concerns after having their children receive the MMR vaccines starting in 2021, but only in states that tend to vote Republican. There is no corresponding jump in reporting concerns in states that tend to vote Democratic. This development probably reflects the findings of a recent YouGov/Economist poll that finds fewer than half (47%) of Republicans believe that vaccinations in general are “very safe.” By contrast, nearly 75% of Democrats believe vaccinations are safe (Frankovic 2021).

Discussion

The COVID-19 pandemic reminds us that recommendations from public health experts and life-saving vaccines from the pharmaceutical industry can only take a country so far. Information and expertise must flow from the scientific community to the mass public. Especially in the case of a pandemic in which people must rely on how other people behave to remain safe, public buy-in must be nearly universal. In the United States, at least, politics short-circuited that process. From the perspective of political science, this development was hardly surprising. A successful pandemic response would require public faith in media and science. Unfortunately, decades-long Republican efforts to win votes by running against these institutions has undermined faith in them. As a result, Republicans naturally turned to President Trump and Fox News for advice about the pandemic rather than Anthony Fauci and the majority of news media that broadcast information that would have provided better guidance.

The change in confidence in institutions over time is stark. In responses to the 1973 General Social Survey (GSS), Republicans actually expressed more confidence in the media than Democrats did. Indeed, only 13% of Republicans said they had “hardly any confidence” in the media. By 2018, that percentage had shot up to 65%. In 1991, when the GSS asked about confidence in the scientific community, nearly half of Republicans (47%) and about a third of Democrats (32%) expressed “a great deal of confidence.” By 2018, Republican confidence had dropped to 39%, while Democratic confidence surged to 50% (Hetherington and Ladd 2020). The pandemic only polarized assessments further. In the 2022 GSS, high confidence in the scientific community among Republicans fell to a mere 26%. It is tough to defeat a pandemic when huge swaths of a major political party have little faith in scientists (the group best positioned to inform best practices) and the media (the group best positioned to transmit those best practices to the masses).

Even with this troubling context in mind, our survey data collected over the course of the pandemic provide a major takeaway time and again: the partisan division over matters of public health, scientific integrity, and even ground truth was not inevitable. It was spurred in large part through deliberate public messaging undertaken by Republican political leaders, President Trump in particular. In an environment of uncertainty, anxiety, and risk, partisans coped by looking to their preferred elites for guidance on what to believe, where to get information, and how to approach their daily lives. At the beginning of the pandemic, political elites worked together to pass legislation and communicate the seriousness of COVID-19. This unified messaging could have continued throughout the ensuing months and years. Instead, messaging diverged, and the mass public responded by becoming divided on attitudes toward mitigation policies and behaviors. This is a reminder that public health and individual well-being are not divorced from politics; that elections matter; and that responsible governance is as important as ever.

Acknowledgments

We thank the editor and an anonymous reviewer for their valuable input. We also thank Leah Christiani, Emily Wager, Amanda Aziz, and Abby Cassario for their research assistance. The Raymond Dawson Chair and the UNC Policy Collaboratory provided most of the funds used to carry out the surveys. Pam Conover, Mike MacKuen, and Tim Ryan were also invaluable collaborators on the survey work.

Note

1.

Galileo was put to death by the Catholic Church for the heresy of arguing that the Earth revolved around the sun, a belief that challenged the traditional understanding that the Earth sat at the center of the universe.

References

Adolph, Christopher, Amano, Kenya, Bang-Jensen, Bree, Fullman, Nancy, and Wilkerson, John.
2021
. “
Pandemic Politics: Timing State-Level Social Distancing Responses to COVID-19
.”
Journal of Health Politics, Policy and Law
46
, no.
2
:
211
33
.
Campbell, Angus, Converse, Philip E., Miller, Warren E., and Stokes, Donald E.
1960
.
The American Voter
.
Chicago
:
University of Chicago Press
.
Carmines, Edward G., and Stimson, James A.
1980
. “
The Two Faces of Issue Voting
.”
American Political Science Review
74
, no.
1
:
78
91
.
Cathey, Libby.
2020
. “
Fact Checking Trump's Claim about Suicides if the Economic Shutdown Continues
.” ABC News,
March
25
. https://abcnews.go.com/Politics/fact-checking-trumps-claim-suicide-thousands-economic-shutdown/story?id=69790273.
CBS News Texas
.
2020
. “
Coronavirus in Texas: Governor Abbott: ‘Aggressive Strategies to Make Sure Texas Is First at Getting Back to Work.’
April
9
. https://www.cbsnews.com/texas/news/coronavirus-texas-governor-abbott-back-to-work/.
Converse, Philip E.
1964
. “
The Nature of Belief Systems in Mass Publics
.” In
Ideology and Discontent
, edited by Apter, David E.,
206
61
.
New York
:
Free Press of Glencoe
.
Druckman, James N., Peterson, Erik, and Slothuus, Rune.
2013
. “
How Elite Partisan Polarization Affects Public Opinion Formation
.”
American Political Science Review
107
, no.
1
:
57
79
.
Ellis, Christopher, and Stimson, James A.
2012
.
Ideology in America
.
New York
:
Cambridge University Press
.
Enten, Harry.
2021
.
“Flu Shots Uptake Is Now Partisan
.
It Didn't Use to Be.” CNN
,
November
14
. https://www.cnn.com/2021/11/14/politics/flu-partisan-divide-analysis/index.html.
Finn, Teaganne.
2021
. “
Trump Renews Praise for Covid Vaccines, ‘One of the Greatest Achievements of Mankind.’
” NBC News,
December
23
. https://www.nbcnews.com/politics/donald-trump/trump-renews-praise-covid-vaccines-one-greatest-achievements-mankind-n1286551.
Frankovic, Kathy.
2021
. “
Fewer than Half of Republicans Now Support Requiring Childhood Vaccinations for Infectious Diseases
.” YouGov,
October
13
. https://today.yougov.com/topics/politics/articles-reports/2021/10/13/support-requiring-child-vaccinations.
Funk, Cary, Tyson, Alec, Kennedy, Brian, and Pasquini, Giancarlo.
2023
. “
Americans’ Largely Positive Views of Childhood Vaccines Hold Steady
.” Pew Research Center,
May
16
. https://www.pewresearch.org/science/2023/05/16/americans-largely-positive-views-of-childhood-vaccines-hold-steady/#:∼:text=Americans%20remain%20steadfast%20in%20their,new%20Pew%20Research%20Center%20survey.
Gadarian, Shana Kushner, Goodman, Sara Wallace, and Pepinsky, Thomas B.
2022
.
Pandemic Politics: The Deadly Toll of Partisanship in the Age of COVID
.
Princeton, NJ
:
Princeton University Press
.
Gerber, Alan S., and Huber, Gregory A.
2010
. “
Partisanship, Political Control, and Economic Assessments
.”
American Journal of Political Science
54
, no.
1
:
153
73
.
Haischer, Michael H., Beilfuss, Rachel, Hart, Meggie Rose, Opielinski, Lauren, Wrucke, David, Zirgaitis, Gretchen, Uhrich, Toni D., and Hunter, Sandra K.
2020
. “
Who Is Wearing a Mask? Gender-, Age-, and Location-Related Differences during the COVID-19 Pandemic
.”
PLoS One
15
, no.
10
: article ID e0240785.
Hetherington, Marc, and Ladd, Jonathan M.
2020
. “
Destroying Trust in the Media, Science, and Government Has Left America Vulnerable to Disaster
.” Brookings Institution,
May
1
. https://www.brookings.edu/articles/destroying-trust-in-the-media-science-and-government-has-left-america-vulnerable-to-disaster/.
Hetherington, Marc, and Weiler, Jonathan.
2018
.
Prius or Pickup? How the Answers to Four Simple Questions Explain America's Great Divide
.
New York
:
Houghton Mifflin Harcourt
.
Higgins-Dunn, Noah, and Feuer, Will.
2020
. “
Georgia Gov. Kemp Urges People to Wear Masks despite Suing Atlanta over Mask Mandate
.” CNBC,
July
17
. https://www.cnbc.com/2020/07/17/georgia-gov-kemp-urges-people-to-wear-masks-after-suing-atlanta-over-mask-mandate.html.
Hofstadter, Richard.
1963
.
Anti-Intellectualism in American Life.
New York
:
Vintage Books
.
Johnston, Christopher D., and Wronski, Julie.
2015
. “
Personality Dispositions and Political Preferences across Hard and Easy Issues
.”
Political Psychology
36
, no.
1
:
35
53
.
Kam, Cindy D., and Sides, John.
2020
. “
Symptoms Vary: Understanding Americans’ Differing Views on COVID-19, Ebola, and Zika
.” Democracy Fund Voter Study Group, July. https://www.voterstudygroup.org/publication/symptoms-vary.
Kates, Jennifer, Tolbert, Jennifer, and Orgera, Kendal.
2021
. “
The Red/Blue Divide in COVID-19 Vaccination Rates
.” Kaiser Family Foundation,
September
14
. https://www.kff.org/policy-watch/the-red-blue-divide-in-covid-19-vaccination-rates/.
Larsen, Bradley J., Ryan, Timothy J., Greene, Steven, Hetherington, Marc J., Maxwell, Rahsaan, and Tadelis, Steven.
2023
. “
Counter-Stereotypical Messaging and Partisan Cues: Moving the Needle on Vaccines in a Polarized United States
.”
Science Advances
9
, no.
29
: article ID eadg9434.
Leeper, Thomas J., and Slothuus, Rune.
2014
. “
Political Parties, Motivated Reasoning, and Public Opinion Formation
.”
Political Psychology
35
,
suppl. 1
:
129
56
.
Lenz, Gabriel S.
2012
.
Follow the Leader? How Voters Respond to Politicians’ Policies and Performance
.
Chicago
:
University of Chicago Press
.
Lupia, Arthur, and McCubbins, Mathew D.
1998
.
The Democratic Dilemma: Can Citizens Learn What They Need to Know?
New York
:
Cambridge University Press
.
Mason, Lilliana.
2018
.
Uncivil Agreement: How Politics Became Our Identity
.
Chicago
:
University of Chicago Press
.
Mehlhaff, Isaac D., Ryan, Timothy J., Hetherington, Marc J., and MacKuen, Michael B.
2024
. “
Where Motivated Reasoning Withers and Looms Large: Fear and Partisan Reactions to the COVID-19 Pandemic
.”
American Journal of Political Science
68
, no.
1
:
5
23
.
Mordecai, Mara, and Connaughton, Aidan.
2020
. “
Public Opinion about Coronavirus Is More Politically Divided in US than in Other Advanced Economies
.” Pew Research Center,
October
28
. https://www.pewresearch.org/short-reads/2020/10/28/public-opinion-about-coronavirus-is-more-politically-divided-in-u-s-than-in-other-advanced-economies/.
Motta, Matt.
2023
. “
Is Partisan Conflict over COVID-19 Vaccination Eroding Support for Childhood Vaccine Mandates?
NPJ Vaccines
8
, article no. 5.
Mulloy, D. J.
2014
.
The World of the John Birch Society: Conspiracy, Conservatism, and the Cold War.
Nashville, TN
:
Vanderbilt University Press
.
NTI/JHCHS (Nuclear Threat Initiative and Johns Hopkins Center for Health Security)
.
2019
.
Global Health Security Index: Building Collective Action and Accountability
.
October
. https://www.ghsindex.org/wp-content/uploads/2020/04/2019-Global-Health-Security-Index.pdf.
Oppenheim, Oren.
2021
. “
Trump Again Publicly Defends COVID-19 Vaccines, but Slams Mandates
.” ABC News,
December
23
. https://abcnews.go.com/Politics/trump-publicly-defends-covid-19-vaccines-slams-mandates/story?id=81914671.
ORSSCC (Oversight and Reform Select Subcommittee on the Coronavirus Crisis)
.
2022
. “
A Hearing with Trump White House Coronavirus Response Coordinator Dr. Deborah Birx
.”
June
23
. https://www.congress.gov/event/117th-congress/house-event/114940/text?s=1&r=90.
Rabin, Roni Caryn, and Cameron, Chris.
2020
. “
Trump Falsely Claims ‘99 Percent’ of Virus Cases Are ‘Totally Harmless.’
New York Times,
July
7
. https://www.nytimes.com/2020/07/05/us/politics/trump-coronavirus-factcheck.html.
Wingrove, Josh.
2020
. “
President Trump Claims Coronavirus Will ‘Fade Away’—As US Sees 20,000 New Cases a Day
.”
Time
,
June
18
. https://time.com/5855541/trump-coronavirus-fade-away/.
Wood, Daniel, and Brumfiel, Geoff.
2021
.
“Pro-Trump Counties Now Have Far Higher COVID Death Rates
.
Misinformation Is to Blame.” NPR
,
December
5
. https://www.npr.org/sections/health-shots/2021/12/05/1059828993/data-vaccine-misinformation-trump-counties-covid-death-rate.
Ye, Xinyuan.
2023
. “
Exploring the Relationship between Political Partisanship and COVID-19 Vaccination Rate
.”
Journal of Public Health
45
, no.
1
:
91
98
.

Supplementary data