Arsh Raziuddin / Vishakha Darbha / The Atlantic

As the two parties offer dueling interpretations of Donald Trump’s response to the coronavirus outbreak, new polling suggests that the pandemic’s ultimate political consequences may be determined by the substantial group of voters who accept the central premise of each side’s case.

The principal line of attack from Democrats, amplified in super-PAC ads already running across the battleground states, is that Trump downplayed the virus’s risk and fumbled the government’s response, especially as the threat grew from late January through early March. The principal defense offered by Republicans is that Trump has responded effectively since he declared a national emergency in mid-March.

Republicans expect that if voters conclude Trump avoided the worst, they will forgive any mistakes at the outset. “Voters are saying they don’t love that first phase, but a lot of people feel personally off guard too. So that’s why you see that generous ‘we were all caught off guard’ sentiment,” the GOP pollster Kristen Soltis Anderson told me recently. The final verdict, she believes, will turn on whether “people perceive the administration as having been uniquely flat-footed, versus all of us being caught off guard by a lack of information.”

Democrats believe that whatever the total cost, Americans will conclude that casualties were higher because of Trump’s early inaction. If voters decide that “we are in a situation where the U.S. had a worse impact than the rest of the world, then the way he handled it early on, and let it get out of control, is pretty damn relevant,” says the Democratic pollster Nick Gourevitch of the Global Strategy Group.

This week’s new national NBC/Wall Street Journal poll offered an intriguing measure of how these arguments are resonating by giving voters a four-way choice to describe their assessment of Trump’s performance. Its results suggest that both sides have made inroads with their claims, but that Trump faces an imposing bloc of entrenched critics who believe his response has been poor from the start.

The poll allowed voters to combine their assessments of Trump’s initial response with their views about his actions now. The largest group, 45 percent, gave Trump negative marks on both measures: They said that “he did not take the threat seriously enough at the beginning and is still not handling it well.” The next-largest group, 30 percent, gave Trump positive marks on both fronts. A negligible 2 percent took the (somewhat eccentric) position that Trump did take the crisis seriously initially, but hasn’t handled it well since. The final group, at exactly 20 percent, is the one that reflected each side’s core arguments, saying that Trump didn’t take the threat seriously enough at the start, but is handling it well now.

From one direction, these results point to an ominous vulnerability for the president. They mean that 65 percent of Americans believe that Trump did not take the threat seriously enough at first—a result in line with other recent polls, as well as the daily Navigator survey run by the Global Strategy Group and GBAO Strategies, another Democratic firm.

From the other direction, the results mean that 50 percent of Americans say Trump is doing a good job now. That’s higher than the 44 percent in the survey who said they approved of his handling of the outbreak when asked directly, but it’s not much different from the results of other recent national polls, which have found the country divided close to 50-50 on his current performance in the crisis. That split in public opinion puts the president in a much more comfortable position than the two-to-one negative verdict over his initial response.

But the most consequential group at this moment may be the one that straddles both these results—the overlapping segment of the Venn diagram. Who is that ambivalent group? It includes many of the constituencies already considered the swing groups in the electorate, according to detailed results provided by Public Opinion Strategies, the Republican firm that co-directs the NBC/WSJ poll with the Democratic firm Hart Research Associates.

The constituencies reflected most in this group include white independents, Republicans who don’t regularly watch Fox News (or who consider themselves “soft” partisans), voters in Sun Belt swing states, young people, white women without a college degree, and voters who live in outer suburbs and rural areas.

Attitudes on this question about Trump’s response powerfully link to opinions about his reelection. Among the 30 percent who said Trump has been strong on the virus throughout, the president leads the presumptive Democratic nominee, Joe Biden, by 94 percent to 4 percent.

But Trump faces a comparable deficit among the much larger group of 45 percent, who say he’s mishandled the outbreak from the beginning. Among those respondents, Biden leads 88 percent to 2 percent. The fact that more of these voters are undecided (some 10 percent) only compounds the president’s challenge: Biden has more room to grow among Trump’s consistent critics than Trump does among his consistent supporters.

The implication is clear. If Trump can’t reduce the large number of voters who think he has mishandled the crisis all along, he’ll need to win a disproportionate share of the 20 percent of voters who are conflicted about his performance. In the NBC/WSJ survey, Trump does lead among those ambivalent voters. But his margin of 54 percent to 33 percent isn’t nearly enough to overcome his deficit with his big bloc of critics. Among all voters, Biden leads him in the poll by a solid 49 percent to 42 percent.

Embedded in all these numbers is evidence that the pandemic could harden the geographic divide that has characterized the Trump era, with the GOP losing ground in major metropolitan areas but consolidating its hold on rural places.

So far, the largest metropolitan areas in almost every state, as I recently reported, are facing much higher caseloads than less populated communities, even when adjusted on a per-person basis. And in those areas, the verdict on Trump’s response is clear and harsh. In the NBC/WSJ poll, a solid majority of residents of both urban communities (56 percent) and inner suburbs (54 percent) said Trump has mishandled the outbreak from the beginning. In both types of communities, only about one in four respondents gives Trump positive marks on both fronts, and only about one in six falls into the divided category.

The picture is very different in the outer suburban and rural communities, where the disease generally hasn’t hit as hard yet. In the NBC/WSJ poll, about two-fifths say he’s handling it well and has from the outset, and only one-third or less take the opposite view on both counts. But in both communities, about one in four respondents falls into the conflicted group, a higher share than in the big population centers. That means that while about two-thirds of respondents in rural communities and outer suburbs say Trump is handling the crisis well now, a clear majority in both also say he did not act quickly enough at the outset.

The obvious risk for Trump is that the latter verdict could overshadow the former if the disease reaches more deeply into those less-populated communities.

That’s an especially huge danger for Trump because, as another set of recent polls shows, his precarious public support on the virus heavily depends on preponderant backing from voters in the places that have been least affected. The Pew Research Center divided respondents in its mid-April poll into three groups: those living in counties that faced high, medium, and low incidences of the disease as of early in the month. The high- and medium-impact counties on one side and the low-impact counties on the other each accounted for almost exactly half of the nation’s population.

These areas diverged strikingly in their assessments of Trump’s response. And ominously for the president, assessments in the medium-impact counties were closer to the (mostly negative) high-impact group.

For instance, while half of those in the low-impact counties said he has done a good job of responding to the needs of hospitals, doctors, and nurses, more than three-fifths of respondents in both the high- and medium-impact counties said he has not, according to detailed results provided to me by Pew. Likewise, while almost exactly half of those in low-impact counties said he’s provided accurate information to the public, more than three-fifths of those in the high- and mid-impact counties disagreed.

It's not hard to imagine that the next phase of the political struggle over the virus might further erode Trump’s position in the largest places that have been both the most heavily affected and the most critical of his response.

Trump and several Republican governors are now pushing aggressively to reopen as much of the economy as possible. But these moves are sparking resistance from many mayors and other local officials, especially those in the big metro areas of Republican-led states. In Atlanta, for example, Mayor Keisha Lance Bottoms immediately denounced Governor Brian Kemp’s order to begin reopening businesses, including gyms and nail salons, on Friday. In Texas, the Dallas County commissioners responded to signals that Governor Greg Abbott intends to reopen the economy soon by defiantly voting this week to extend the county’s stay-at-home order through May 15.

As I’ve written before, in seeking to quickly reopen the economy, Trump and the GOP governors are taking the risk of signaling to metro voters that they are placing a higher priority on the economic needs of their small-town base than on the public-health concerns of those in the largest communities.

Trump and many other Republicans would probably accept that risk so long as they can maintain their rock-solid support in the smaller places. The big question is whether they can sustain such backing if the disease reaches those communities in greater force, especially since even there, according to the NBC/WSJ poll, so many voters already are either openly critical or ambivalent about his response.

Michael T. Osterholm, the director of the Center for Infectious Disease Research and Policy at the University of Minnesota, believes that the disease’s diffusion remains inevitable. Even without an early end to social distancing, it is highly likely that the virus will spread until it infects about two-thirds of all Americans, he told me, estimating that roughly 5 percent have been infected so far. Osterholm said he anticipates a second round of disease, most likely in the fall, when the outbreak could not only resurge in the places that have already been hit, but also reach more broadly into places that have been spared.

A new study from the Brookings Institution demographer William Frey already shows that the disease is extending somewhat deeper into Trump country over time. While Trump won only about one-third of the vote in the counties that reached a critical mass of cases in early April, Frey found, that number rose slightly past 40 percent in the counties that have joined them since.

The communities beyond the metro core that generally have suffered the least damage from the outbreak contain the most voters who describe themselves as ambivalent about Trump’s response. If at any point before November, the disease does reach further into those culturally conservative smaller places at the core of Trump’s electoral coalition, it might threaten his last line of defense—especially because Trump and other Republicans are now betting so heavily on pushing to unwind social restrictions. To Osterholm, the question is not if, but when, the coronavirus reaches all corners of the country: “This virus’s movement is a lot like gravity: You can count on it.”

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.