We investigate trends in public opinion before and during the COVID-19 pandemic on concern about the virus, efforts to combat its spread, and support for various health care policy reforms. Using data from more than 400,000 interviews covering every state in the country over a 1-year period, we demonstrate initially high levels of concern about the virus and support for restrictions to combat it across geography and political party. Over time, however, these sentiments fade, and a partisan divide opens up as Republican concern and support drops more quickly. We find little evidence that the pandemic has shifted opinions on health care reforms such as Medicare-for-all. We argue that the differences in these trends have more to do with messages from party leaders about the pandemic and less to do with the virus’s state-level infection rates. While nearly everyone has been affected in some way by the pandemic, for most voters, the pandemic’s challenges have not translated into a desire for more sweeping change to the nation’s health care system.
Keywords: COVID-19, elections, health care, public opinion, polarization, pandemic
As of November 14, 2020, the novel coronavirus has infected at least 53.4 million people worldwide, killed more than one million people, and disrupted nearly every aspect of social, political, and cultural life. In the United States, the spread of COVID-19 has been particularly severe. The New York Times reports that on a per capita basis, the United States has seen more infections than many developed democracies (New York Times, 2020). In terms of total infections, the United States tops the list of all nations. Recent increases in infections and deaths in a number of U.S. states mean that the pandemic is far from over. Meanwhile, the U.S. presidential election is right around the corner.
The depth of the pandemic and the upcoming election raise two compelling questions: (1) How have Americans developed opinions about COVID-19? and (2) Has the pandemic altered opinion about public health policy more broadly? Answers to these questions will help us understand the ways in which the virus may be shaping the 2020 elections, but also potential avenues to stop its spread.
Americans’ views of the coronavirus may directly affect the pandemic in at least two ways. First, people’s concern about the pandemic and the choices they make affect their susceptibility to infection and the risk of infecting others. Second, people’s opinions about policies intended to combat the pandemic affect the ability of elected leaders to enact those policies, many of which impose substantial costs on citizens. People’s reactions to these policies may also play a role in which leaders they elect.
The pandemic has also exposed flaws in the U.S. health care system—not only in terms of specific measures needed to fight the pandemic, such as testing capacity and protective equipment for medical workers, but also in terms of basic features of health insurance provision. In particular, providing health insurance through employers seems increasingly problematic when unprecedented numbers of Americans have lost their jobs because of the pandemic.
Taking advantage of an extraordinarily large survey effort, the Democracy Fund + UCLA Nationscape Project, we map the contours of public opinion about the coronavirus and health care policy across geography and political party. We show initially high levels of concern about the virus and support for key measures to combat it. Over time, however, these sentiments have decreased, particularly among Republicans.
There has been much less overall change in views of health care policy, especially the proposals hotly debated during the Democratic primary, such as Medicare-for-all. The pandemic has not created a groundswell of support for larger government involvement in the health care system. This suggests that Americans have not yet heard messages from political leaders or others that would help them connect the pandemic to the need for larger scale policy reforms.
Our survey data come from the Democracy Fund + UCLA Nationscape Project (Tausanovitch & Vavreck, 2020). The project started in July of 2019 and is ongoing. It is fielded online to an average of 6,250 people each week by a team of researchers at UCLA with samples provided by Lucid, a market research firm. To date, Nationscape has completed over 400,000 interviews across the entire country. There are more interviews in big states like California and Ohio than in small states like New Hampshire, but there are enough interviews across both time and space to allow us to investigate different types of people in different places at different times.1 For example, these large weekly samples allow us to trace attitudes about COVID-19 as the infection rate increased on the East Coast (both for people living there and those living elsewhere) and compare what we observe to attitudes during the later period when infections subsided on the East Coast and rose in the South and West. The Nationscape data are particularly well suited to examining the impact of COVID-19, since its prevalence has varied across both time and space, as well as across time within states and localities.
The Nationscape sample is a nonprobability sample based on quotas and weights. Owing to the extremely large size of the Nationscape data collection, quotas are set across a large number of categories to ensure balance before weighting. Weights are also calculated to address intersections of categories that would not be possible with smaller sample sizes and to correct for variations in response probabilities. On items of interest, Nationscape data represent nationwide targets well prior to weighting. The representativeness of Nationscape was evaluated by replicating the process used by the Pew Research Center to evaluate the representativeness of online samples. Nationscape respondents were asked a set of questions also asked on reliable large-sample government surveys. Estimates were produced based on responses to these questions using the same procedures used to produce weekly Nationscape estimates. Nationscape returned estimates that were as close to the government survey targets as the other online samples tested by the Pew Research Center. The median difference between Nationscape estimates and the government survey targets was 3.5% while the Pew analysis found a median difference of 3.6% for other online samples. For more details on the representativeness of the sample, see Tausanovitch et al. (2019).
Early in the pandemic, the public’s concern about the coronavirus was quite high. In the first Nationscape survey to include questions about the pandemic, which was fielded March 18-26, the vast majority of the respondents—90%—said that they were somewhat or very concerned about the coronavirus (Figure 1). This was true among Republicans, Democrats, and Independents. The main variation among partisans was whether they were very concerned: at that point, 70% of Democrats said this compared to 51% of Republicans.
Notably, opinions did not vary much based on the state in which respondents lived. Here we use a simple categorization of U.S. states based on the extent and timing of their reported COVID infection rates. “Early peak” states reach a threshold of 30 confirmed cases per 100,000 residents before June 1. This includes states such as New York and Massachusetts. “Later peak” states fail to reach this threshold by June 1, but eventually reach at least 10 cases per 100,000. This includes states such as Arizona and Florida. The remaining states (“low rate”) have yet to reach 10 cases per 100,000 as of this writing.2 There is, of course, variation within states in the severity of the pandemic. But COVID statistics are so frequently presented at the state level that Americans could plausibly form attitudes based on state statistics. Moreover, research on the effects of COVID infections on political attitudes shows that those effects are similar whether infections are measured at the state or county levels (Warshaw et al., in press).
At this relatively early stage of the pandemic, respondents in each group of states express similar levels of concern. Views of the virus were fairly nationalized. The same pattern is visible in opinions about state and local policies intended to combat the virus, including restrictions on nonessential travel. Early in the pandemic, there was nearly universal support for these policies across parties and states. The main differences were simply in whether people “strongly” supported these policies. For example, Democrats were about 10 points more likely to strongly support canceling large meetings (71% vs. 61% for Republicans). And people in “early peak” states that were beginning to see a serious outbreak were about 8 points more likely to strongly support this policy than people in other states that had not seen as many cases.3
Over time, however, this consensus broke down. It did so largely along partisan lines, not based on the severity of the pandemic across states. Republicans became less likely to express concern, and especially strong concern, about the virus. Only with the upsurge in cases in June and July has their concern increased, although it is still not back to its level of mid-March.
Partisan divides on state and local COVID policies have become even starker. There are now 25-30-point gaps between Democrats and Republicans in their support for canceling large gatherings, restricting nonessential travel, and closing schools and universities—differences that persist even once factors like age, race, gender, and education are taken into account (Gadarian Kushner et al., 2020; Kam & Sides, 2020). The late summer and early fall upsurge in cases has done little to reduce this polarization. The upshot is that Democrats remain fairly unified in their support while Republicans are roughly evenly divided. Indeed, in states such as Texas, this divide plays out publicly, with some Republican voters and lawmakers opposing Republican leaders who try to implement stricter policies, such as mask requirements (Fernandez & Goodman, 2020).
This growing partisan polarization arguably reflects the messages of leaders in each party, following the “two-message model” described in John Zaller’s (1992) seminal work on attitude change. Some Republican leaders, and especially President Donald Trump, have repeatedly minimized the threat of the virus and advocated for a ‘reopening’ that would bring people back to stores, churches, schools, and the like. Meanwhile, Democratic leaders have tended to express continued concern about the virus and to support stricter restrictions. That these messages matter is evidenced by differentiating people based on how much attention they pay to politics. As Kam and Sides (2020) show, the partisan divide on COVID-19 is larger among people who pay more attention to politics and thus are more likely to hear messages from their party’s leaders.
Notably, this partisan polarization is present in each of the three groups of states (Figure 2). Regardless of the timing and severity of the virus, the large majority of Democrats express concern about the virus and support restrictions like canceling large gatherings. The surge of the virus in later peak states like Arizona and Florida may have led more Democrats in those states to support these restrictions—but any effect is modest at best.
Similarly, the decrease in concern and support among Republicans is visible in all three groups of states. Republicans appear to be responding to partisan messages more than the infection rate in their states. This is particularly evident in later peak states, where there has been at most a modest increase in Republican concern about the virus and little to no increase in their support for restrictions on large gatherings, travel, and schools despite increasing cases and deaths.
Although the COVID-19 pandemic has elicited extraordinary levels of concern and sacrifice by many Americans, there is one thing it has not done: build support for major reform to the health care system. In Nationscape surveys dating to July 2019, we have found no secular trends in policy preferences among Americans as a whole (Figure 3)—and no evidence of the partisan polarization visible in attitudes toward the coronavirus.
For example, just over half of the respondents have supported providing government-run health insurance to all Americans, while about a quarter oppose it, and a fifth have no opinion. Similar fractions of respondents have supported ‘Medicare-for-all’ and a larger fraction have supported subsidies for health care for lower income people who do not receive Medicare. A much smaller fraction (about 30%) supports government-run insurance when it is accompanied by the abolition of private insurance—suggesting that most respondents, while seeing some role for government in health insurance provision, are not willing to go as far as policymakers like Sen. Bernie Sanders.
The lack of change in opinion over the past year signals two things. First, the extensive discussion of health care reform during the 2020 Democratic primaries did not increase Americans’ support for reform, in the aggregate. Second, the COVID-19 pandemic has not yet shifted health care preferences either, even as it reveals shortcomings in how the United States provides health insurance and health care.
A global pandemic that hits the United States would seem like the exact kind of trigger to make people reexamine how health care is provided and paid for. Why hasn’t the pandemic changed Americans’ health care preferences? One possibility is that political leaders have not yet connected the pandemic to a broader agenda for health care reform—thereby helping ordinary voters, who are rightly focused on dealing with the day-to-day challenges of the pandemic, connect COVID-19 struggles to health care policies. Indeed, if anything, the pandemic may have reoriented political debates around the specific problems posed by the pandemic, rather than the more general problems with the U.S. health care system. As one Washington Post columnist put it, “We’re in a pandemic. Why have we stopped talking about Medicare-for-all?” (Olen, 2020).
It is commonly said that the COVID-19 pandemic is changing our politics. Certainly, it has changed the national political agenda and is arguably dragging down the political fortunes of President Trump and other Republican candidates (Warshaw et al., in press).
But at the same time, politics is also shaping the pandemic. Public attitudes about the virus reflect the increasingly polarized messages that Democrats and Republicans are receiving from their leaders. These messages appear to take precedence over local COVID infection rates, at least at the state level.
In turn, the absence of messages that explicitly link the pandemic to broader health reform mean that public attitudes toward proposals such as Medicare-for-all have not changed much. For the moment, most of the political energy is going to taming the virus, not to revolutionizing the health care system. If the pandemic is to have a lasting impact on the fundamentals of the U.S. health care system, this will need to change.
All authors contributed equally to the research and writing of this paper. C. Tausanovitch and L. Vavreck are the principal investigators of the Democracy Fund + UCLA Nationscape Project.
John Sides, Chris Tausanovitch, and Lynn Vavreck have no financial or non-financial disclosures to share for this article.
Fernandez, M., & Goodman, J. D. (2020, July 23). Red vs. red in Texas, with Republicans battling one another after mask order. New York Times. https://www.nytimes.com/2020/07/23/us/coronavirus-texas-abbott-republicans.html
Gadarian Kushner, S., Goodman, S. W., & Pepinsky, T. B. (2020). Partisanship, health behavior, and policy attitudes in the early stages of the COVID-19 pandemic. SSRN. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3562796
Kam, C. D., & Sides, J. (2020). Symptoms vary: Understanding Americans’ differing views on COVID-19, Ebola, and Zika. Democracy Fund Voter Study Group. https://www.voterstudygroup.org/publication/symptoms-vary
Kish, L. (1992). Weighting for Unequal Pi. Journal of Official Statistics, 8(2), 183–220.
New York Times. (2020, November 14). Covid word map: Tracking the global outbreak. https://www.nytimes.com/interactive/2020/world/coronavirus-maps.html?action=click&module=Top%20Stories&pgtype=Homepage#countries
New York Times. (2020). Covid-19 data Github repository. https://github.com/nytimes/covid-19-data
Olen, H. (2020, May 28). We’re in a pandemic. Why have we stopped talking about Medicare-for-all? Washington Post. https://www.washingtonpost.com/opinions/2020/05/28/were-pandemic-why-have-we-stopped-talking-about-medicare-for-all/
Tausanovitch, C., & Vavreck, L. (2020). The Democracy Fund + UCLA Nationscape Project. Data from the weeks of July 19, 2019-October 15, 2020, Los Angeles, CA.
Tausanovitch, C., Vavreck, L., Reny, T., Rossell Hayes, A., & Rudkin, A. (2019). The Democracy Fund + UCLA Methodology and representativeness assessment. Democracy Fund Voter Study Group. https://www.voterstudygroup.org/uploads/reports/Data/NS-Methodology-Representativeness-Assessment.pdf
Warshaw, C., Vavreck, L., & Baxter-King, R. (in press). The effect of local COVID-19 fatalities on Americans’ political preferences. Science Advances. http://chriswarshaw.com/papers/covid_elections.pdf
Zaller, J. (1992). The nature and origins of mass opinion. Cambridge University Press. https://doi.org/10.1017/CBO9780511818691
©2020 John Sides, Chris Tausanovitch, and Lynn Vavreck. This article is licensed under a Creative Commons Attribution (CC BY 4.0) International license, except where otherwise indicated with respect to particular material included in the article.