Americans Want to Return to Normal. But Also They Don’t.

COVID surveys provide a muddy picture. Is it the polls or the American people who are confused?

A spectrum of coronaviruses with sad and happy faces on them.
The Atlantic

Recent opinion surveys give mixed messages about how Americans perceive the current state of the pandemic, and what they think we should do about it. In a February Washington Post/ABC News poll, for example, 58 percent of Americans said that controlling the spread of the coronavirus is more important than loosening restrictions on normal activities. In a Yahoo News/YouGov poll conducted the same week, 51 percent said we need to learn to live with COVID-19 and get back to normal.

These are two of several examples that show Americans have seemingly conflicting views about the pandemic. A natural question to ask is why—is it the polls or the American people who are confused? And what do Americans really think?

Trust in public-opinion data has been damaged by the past two presidential elections; in each, the Democratic candidate’s popularity was overestimated in preelection polls, resulting in a surprise victory for Donald Trump in 2016 and a narrower-than-expected victory for Joe Biden in 2020. Despite the cycle of soul-searching that followed the 2016 errors, which were mostly confined to a few states, the problems got worse in 2020. Some pollsters pinned the blame on “nonresponse error,” meaning Trump supporters who wouldn’t answer surveys—perhaps because they’re mistrustful of institutions—but did turn out to vote. These nonrespondents could have skewed the results, making support for Trump seem less robust than it really was.

There are substantial differences between preelection polls and issue-specific COVID polls, however. The biggest distinction is that preelection pollsters want to include in their surveys only people who are going to vote in the upcoming election—at most about two-thirds of eligible voters. That is a difficult task, prone to error. COVID pollsters want to talk with all American adults—a known population. If you’re an American and an adult, your opinion counts; there’s no added complexity of registration status or motivation to cast a ballot on Election Day.

Additionally, in an election, a margin of error of a few percentage points could easily be larger than the ultimate spread. In issue polling, whether 48 or 52 percent of Americans hold a belief is not that consequential. It’s about half either way.

What about nonresponse error? Might COVID polls be, for instance, overestimating support for mitigation measures because some people who dislike them simply won’t pick up the phone to say so? Most surveys outside the election context do not have a real-world trial to test the accuracy of their results. In the case of COVID, however, vaccine uptake has offered an unusual reality check.

As vaccines became available, several polling firms (including the Public Religion Research Institute, where I work) began asking about vaccine uptake in their surveys. Because the CDC was also tracking the number of vaccines given, this provided a population number with which to measure the surveys’ accuracy. The news was positive. In September, the Pew Research Center analyzed 98 polls conducted by 19 polling firms and found that survey results matched the CDC rates quite closely. High-quality polls have done a reasonable job of measuring COVID-vaccination behavior.

Nonetheless, COVID surveys provide a muddy picture of how people are thinking about the pandemic. That’s probably a result of how pollsters phrase the questions they’re asking—always a big factor in opinion surveys. Good pollsters can disagree on how to word a question, and they can have different purposes for a question that lead to different wording. The disparity in wording is not a bad thing—it is simply information you need to know about a poll to put it in proper context.

Going back to the earlier examples, the Post/ABC poll asked respondents to choose between two options: “What do you think is more important—trying to control the spread of the coronavirus, even if it means having some restrictions on normal activities, or having no restrictions on normal activities, even if it hurts efforts to control the spread of the coronavirus?” Compare that with what the Yahoo News/YouGov poll asked: “Which is closer to your opinion of COVID-19?” with the answer options of “We need to learn to live with it and get back to normal” and “We need to do more to vaccinate, wear masks, and test.”

These questions seem similar. Both ask respondents to choose between “normal” and restrictions. But some key differences likely affected how people answered. First, the Post/ABC poll asked what is “more important,” but did not specify to whom, while the Yahoo News/YouGov poll asked which option was “closer to your opinion.” The latter framing implies that all that matters is your opinion, whereas “more important” leaves the door open for thinking more broadly about society.

Second, the answer options about restrictions varied considerably. The Post/ABC poll referred to “some restrictions on normal activities” (58 percent support), while the Yahoo News/YouGov poll specifically mentioned vaccines, masks, and tests (36 percent support). Perhaps it’s not surprising that Americans are more supportive of the vague “some restrictions” than masks and so forth in particular.

Other polls show similar juxtapositions. The March Axios/Ipsos coronavirus-index poll found that 51 percent of respondents support their state or local government requiring masks in public places, but 54 percent support federal, state, and local governments lifting all COVID restrictions. That’s right: Majorities in the same poll supported mask mandates and lifting all restrictions—which presumably include mask mandates, although the latter question, notably, did not use the word masks.

Looking at another item from the same poll, with four answer options, helps clarify this apparent contradiction. Some 27 percent of Americans said the country should “open up and get back to life as usual with no coronavirus mandates or requirements,” and another 44 percent said the country should “move toward opening up, but still take some precautions.” Only 15 percent said we should “mostly keep coronavirus precautions and requirements in place,” and 8 percent wanted to “increase mask mandates and coronavirus vaccine requirements.” If the plurality of Americans are in the cautious-but-move-on category, that leaves a lot of space for different opinions on particular mandates and restrictions—and perhaps even a willingness to go along with whatever the government chooses. That is to say, if the government decides to require masks, these people are okay with it; if the government decides not to require masks, they’re okay with that too.

February’s Kaiser Family Foundation COVID-19 Vaccine Monitor poll offered further insight by asking a series of questions about how Americans would feel if restrictions were lifted and if they were not lifted. Nearly half of respondents said they would be worried about increased deaths (49 percent) and overwhelmed hospitals (48 percent) if restrictions were lifted, and a majority (61 percent) worried that people at high risk would be left behind. At the same time, more than six in 10 worried that kids’ and teenagers’ mental health would suffer (65 percent) and that businesses would lose revenue (63 percent) if restrictions were not lifted.

Together, what these polls show is that no single question can capture public opinion on COVID-19. Americans express a strong desire to get on with life, but also believe that COVID remains a real problem. Americans’ attitudes are complex because the issues involved are complex. If Americans can agree about anything, it’s that this complexity is here to stay. According to the Kaiser Family Foundation poll, 78 percent of Americans believe that “normal” life will look different than it did pre-pandemic.