Graphics by Anna Wiederkehr
Americans have changed their behavior in ways that would have been unthinkable even a few months ago. Masks are an essential accessory. Social distancing is the norm. And even as states moved to reopen their economies in May and June, many Americans continued to think it was better for people to stay home.
But underneath that apparent consensus is a large — and growing — partisan divide. Even as cases and hospitalizations spike in red states that mostly escaped the early effects of the virus, Republicans and Democrats remain stubbornly split on the threat it poses. For instance, it was only in July that President Trump wore a mask in public for the first time. And perhaps thanks to Trump’s repeated downplaying of the threat that COVID-19 poses, Republicans are much less concerned than Democrats are about the virus.
On the one hand, according to surveys conducted by the Pew Research Center, Republicans have consistently been less likely than Democrats to say that they fear being hospitalized because of COVID-19 or that they might unknowingly spread the virus to others. But on the other hand, that partisan gap has widened significantly between April and June.
It’s hard to find a more extreme test of our tribal political attachments than the current pandemic, where Trump continues to downplay the risks of the virus in the face of near-universal opposition from medical experts. It also raises a thorny issue: In the midst of a pandemic, partisanship appears to be shaping people’s perceptions of their risk and personal behaviors — to the point that our divided politics actually affects our health. For Americans, that might mean that questions of whether to stay home, wear a mask or to see friends and family without social distancing are filtered through a partisan lens.
In other words, do our politics risk making us sick?
It’s pretty clear that at this point, Republicans’ and Democrats’ experiences of the pandemic have been steadily diverging for months. It’s much harder to say, though, what that means for transmission of the virus. Some surveys offer a glimmer of hope, suggesting that the partisan gaps in how people are actually behaving — whether they wear a mask, for example — are much narrower than the divides on questions about what they think the government should do in response to the virus or whether the worst is behind us. It’s possible, too, that some of the partisan divides we’re seeing now could start to narrow as outbreaks spiral out of control in states like Arizona, Florida and Texas.
These trends are cause for alarm among the small army of social scientists who have tried to figure out how Americans are responding to the virus since the beginning of the pandemic — from the conflicting signals they’ve received from Trump and other political leaders, to changing guidelines from public health experts.
“Some Republicans are much less freaked out by the virus than they were a few months ago,” said Marc Hetherington, a political scientist at the University of North Carolina who is tracking Americans’ perspectives of the coronavirus through a panel survey. “But things are changing so quickly — these new outbreaks could scare them and maybe some of that polarization disappears.”
That doesn’t mean the politicization of the virus isn’t having an impact, though. Take the political fighting around whether people should be required to wear masks or the timeline around when businesses should reopen. The virus is spiking in Georgia, with thousands of new cases each day, but the state’s Republican governor is suing the Democratic mayor of Atlanta over the city’s decision to revert to its most restrictive opening phase and mandate the wearing of masks. “The national conversation about how we behave during this pandemic has been so colored by the partisan divide that it’s becoming impossible to talk rationally about the risks we are and are not willing to tolerate,” said Dr. Sandro Galea, an epidemiologist and the dean of the Boston University School of Public Health who studies the politics of public health. “If both sides were pushed out of their corners, they would both have to concede quite a bit, and we’d frankly all be safer.”
Understanding how Americans are responding to the pandemic isn’t an easy task; there are essentially two methods at researchers’ disposal. The first is to use a survey. The second is to look at mobility trends, such as geolocation or credit card data, to see if people are actually behaving the way they say they are. And over the past few months, political scientists and economists have leaned on both methods to figure out how Americans are thinking about the COVID-19 pandemic and how that relates to their behavior. With the exception of a few studies conducted in late March and early April, when fear of the pandemic ground the economy to a complete halt, all of this research has uncovered an accelerating partisan divide, too.
For example, as early as March, a group of researchers found that Democrats in a large panel survey exhibited more worries than Republicans about the pandemic and were also likelier to embrace health behaviors like more frequent hand-washing or avoiding mass gatherings. The first round of Hetherington’s survey suggests a partisan divide in Americans’ support for some public health interventions, like widespread testing.
The problem with these surveys, of course, is that there’s no way to figure out, for example, whether someone who says they’re quarantining is actually doing so. So a number of other studies have tried to figure out what people were actually doing by using geolocation data to follow people’s movements. This research has found basically the same thing as the surveys: People in Republican-leaning counties, or counties that voted for Trump in 2016, didn’t reduce their activity as much as people in Democratic counties.
Another study that looked at individual-level smartphone data found a similar pattern. And one team of researchers examined both survey data and geolocation data and determined that the trend held up for both — people in more Republican areas were less likely to feel at risk because of COVID-19, and they were also less likely to stay at home.
But this mobility data has its own limitations, according to Rebecca Katz, a professor at Georgetown University Medical Center. It can only tell you whether people are leaving their homes, not where they’re going or whether they’re taking precautions. “We’re all using this data because it’s the data we have, but it’s imperfect,” she said. “Sometimes, I pack my kids in a car and we just drive for a little while so we can get out of the house — by my cell phone, we’re moving. But that doesn’t tell someone looking at that data if we are interacting with other people, or if we’re wearing masks.”
Geography is another confounding factor; people in rural areas are more likely to drive places, even if they’re otherwise following public health guidelines, and less densely populated parts of the country were also less hard hit by the virus in the beginning. The problem is that Republicans are more likely to live in those parts of the country — and the effects of political segregation and the virus’s trajectory are very difficult to untangle, especially for studies that were conducted a month or two into the pandemic.
The partisan split was hard to deny, though, so early on, a couple of research teams tried to figure out why Republicans and Democrats were responding to the pandemic differently. Two usual culprits — politicians and the media — emerged as possible factors in the divide.
One study conducted from late February through the end of March found that the partisan divide on risk perception and health behavior only narrowed after the White House issued federal social distancing guidelines, suggesting that Trump’s role as a national Republican leader could be quite significant. Several other studies dug into the impact of cable TV, with one survey finding that an MSNBC viewer’s response to the pandemic was quite different than that of a Fox News watcher. Another study focused only on the impact of Fox News and concluded that an increase in viewership did appear to result in less social distancing. The evidence for the effects of politicians and differing media sources is less robust because there aren’t as many studies, but it does suggest that even when there are serious health risks at stake, how both talk about the virus and the public health response may affect the way people behave.
Shana Gadarian, a professor of political science at Syracuse University who is helping to conduct one of the panel surveys, said she was surprised to see such enormous divides emerge as the pandemic wore on. According to other research she’s conducted, moments of extreme anxiety and uncertainty can actually make people more open to new sources of information — including public health experts and leaders from the opposing party. So at the beginning of the pandemic, she and her team expected that Americans would coalesce around public health experts’ recommendations, or that other demographic factors — like age — would turn into key dividing lines.
Scientists and doctors do still enjoy a high level of trust from most Americans, as Maggie Koerth wrote for FiveThirtyEight in May. But that doesn’t mean they are entirely immune to the winds of partisanship — for example, Democrats are likelier than Republicans to trust the Centers for Disease Control and Prevention.
Crucially, though, big divides haven’t emerged everywhere. According to the latest wave of the Democracy Fund’s Voter Study Group, conducted between July 2 to 8, the vast majority (88 percent) of Republicans said they wore a mask when going out in public, even though Republicans in greater numbers have said in other surveys that the government shouldn’t require people to wear masks. And according to Robert Griffin, research director of the Voter Study Group, that’s significantly higher than in any wave of the weekly data going back to May 28. There was more of a partisan gap in responses to other questions about coronavirus-related behavior, although it was still fairly modest.
So are these partisan splits actually driving the spread of the virus?
As it turns out, it’s hard to prove that Republicans’ resistance to mask mandates or social distancing is actually worsening the pandemic. One reason is that political scientists and economists don’t feel equipped to take on the epidemiological modeling that would be necessary to measure what, say, a partisan divide on hand-washing actually means for the spread of the disease. Yael Hochberg, an economist at Rice University, said that the lack of uniformity in testing data made her reluctant to wade into the public health data. “There are places where testing still isn’t widely available,” Hochberg said. “And if testing isn’t uniform, it’s hard to compare what you’re seeing in one county versus another.”
One study tried to pin down the effect of differing levels of compliance with social distancing policies among Republicans and Democrats using individual geolocation data. It concluded that a Trump voter who contracts COVID-19 infects 16 percent more people than a comparable Clinton voter. That’s a striking finding — but it’s also only one study, and several infectious disease experts who reviewed the paper at my request were a little skeptical of its conclusions.
Samuel Scarpino, a professor at Northeastern University who studies infectious diseases, said that it can be very difficult, even in a sophisticated model, to separate all of the confounding factors that could be at play, like geography. And Katz said that without information about whether people are wearing masks or engaging in social distancing, it’s hard to draw very solid conclusions about transmission from mobility data.
Scarpino was quick to add, though, that polarization can still be a serious problem, even if it’s hard to quantify its precise impact. “If politicians’ messaging is making people feel like they’re safe from COVID, those are people who are unnecessarily being put at risk,” he said. He’s also concerned that public health experts’ credibility will erode as certain health behaviors, like mask-wearing or social distancing, become associated with one party or another. “We’re kind of building the airplane as we fly it and we need to be able to change course when we get new evidence,” he said. “But it becomes harder to have those conversations and get buy-in from the public as the whole process becomes more politicized.”
There’s danger in exaggerating the extent of the partisan divide, though. Galea told me that he’s been struck by the fact that so many Americans — including nearly all Republicans — report they are going along with health experts’ recommendations, like wearing masks, at least to some degree. And it would be a mistake, Galea said, to gloss over this unusual level of partisan unity, because it’s a sign that health behaviors aren’t as divisive as they could be, given the strength of partisan loyalties.
“Nobody should ignore the fact that people on the political extremes are embracing polarizing positions on health behavior that should not be polarized,” Galea said. “But I think the evidence we have indicates that most people have tried to be responsible and adopt the recommended behaviors, even at a time of immense polarization and confusion and discomfort.”
That said, he still thinks some politicians — and in particular, Trump — need to do more to get on the same page as public health experts. “It’s not that politics is making it impossible to implement these health behaviors, because we see that many ordinary people are getting on board regardless of what political leadership is saying,” he said. “But that doesn’t mean we should give politicians a pass for turning these serious, serious health conversations into a political football, because that is very much to our detriment.”