It’s not even election day, though Ten million of Americans have already gone through the process of waiting in line to cast a vote. Even for those with the options for mail-in and drop-off ballots, early voting lines to have stretched for blocks in several states.
In theory, voting early means fewer people will try to take the polls on the actual election day – a good thing during a pandemic. But when an early vote itself leads to a large crowd – and voters keep coming back – it is worth asking if we are really avoiding as much virus transmission as we hoped.
How big is the COVID-19 risk of voting in person? Here at FiveThirtyEight, we usually turn to scientific research to answer such a question. So we did that. And … well …
There are two studies that attempted to quantify how the vote in Wisconsin Elementary School in early April contributed to the spread of COVID-19. One of them concludes that it wasn’t. The other says the opposite: the busier and more crowded the polling station, the more local COVID-19 cases have occurred afterwards.
With the millions of people voting in person this election season, the conclusions that the study makes is extremely logical. And experts told me that one study is probably closer to the truth than the other. But neither of the two studies really answers the question completely – a fact that is also of great importance for understanding the risks of COVID-19. Studying a new virus in real time is frustratingly inaccurate, especially when our country has missed important opportunities for thorough testing and tracking.
If you can still think back on the long eons leading up to April 2020, you will remember that the first few weeks of the pandemic were a chaotic time for the primary voting. In Wisconsin, it wasn’t clear until almost the day of the election whether and when a personal vote would take place.
The turbulent nature of this elementary school has drawn researchers to study it. “I saw in the papers that they were forcing people to vote. I hated it and expected to find many infections, ”said Lawrence Wein, professor of management science at Stanford University School of Business. Because of his focus on health care, he wanted to know how mismanagement of an election could contribute to the spread of COVID-19.
Wein’s study calculated the R-zero value of the state – a measure of how quickly the virus is spreading – in the weeks before and after the April 7th primaries. During this time nothing seemed to change in the R-Nothing. And his data showed that hospital stays in the state actually fell in the weeks following the vote, suggesting that his calculations weren’t just a fluke caused by inadequate testing. It wasn’t what Wein expected, but his data suggested that the steps election officials had taken – wearing masks and gloves, providing hand sanitizer to voters, and facilitating social distancing – helped that COVID-19 did not spread in the elections.
(FiveThirtyEight published separate research by two sociologists who came to a similar conclusion.)
But wait a second. Because while Wein’s approach tells us somethingIt’s not the whole picture. There is no counterfactual as there is no way to dress up what would have happened in the election had not was held. Otherwise, you won’t know if a static R-Nothing means the choice didn’t change anything, or if it means cases would have like otherwise and the choice just kept them on par.
This is where the other study comes in. Chad Cotti, professor of economics at the University of Wisconsin-Oshkosh, found that counties across the state did not approach elementary school in the same way. Polling stations would be closed and consolidated, said Cotti, but it was not uniform. The result was large differences in the voter density per polling station from district to district. “I live in Appleton and don’t think we’ve had any consolidation,” he said. “But Green Bay has had significant consolidation, and these counties are contiguous.”
Cottis paper used these differences to produce some kind of counterfactual result. We’ll never know what would have happened to the total number of COVID-19 states had the primaries never happened. However, we can know whether counties that had more voters in fewer locations had a greater effect than counties that did not.
And here there was One difference: a 10 percent increase in personal voters in a given location was linked to an 18.4 percent increase in COVID-19 cases in that county. More voters, more viruses.
Cotti’s paper may provide some insight into Wein’s shortcomings, but none of these papers are perfect. Cotti’s study, for example, is still a preprint that will be revised during the peer review process. One thing he told me the reviewers wanted to approach is the fact that the relationship between the increase in voters per location and the increase in local COVID cases is likely to be non-linear. “Adding a person to a polling station does not mean incremental increases. There is likely a relatively safe density and after a certain point it increases rapidly. So a group twice that size can be more than twice as dangerous,” he said.
But more importantly, other experts told me, none of these studies can really tell us how risky choices are compared to other behaviors. To do that, you’d have to randomly assign people to voting and non-voting groups and figure out which group is more likely to get COVID-19, said Tariq Cheema, pulmonologist and intensive care doctor for the Allegheny Health Network Who has published research on the transmission of COVID-19. And that’s not really ethical.
According to Jennifer Dowd, Professor of Demographics and Population Health at Oxford University in England, this is a challenge that makes sociological research a real bear in even the best of situations. For example, let’s say you want to know how a change in policy – like increasing the minimum wage – changes results in a particular city or state. You will run into the same problems, she said. How do you know the two places you’re comparing would have been on the same trail if either of them hadn’t raised the minimum wage? What if places that are more inclined to raise wages initially had something different about them? How do you separate cause and effect?
“It’s a very persistent problem in social science. And with COVID it was really difficult,” said Dowd. For example, we don’t know if the data from the Cotti paper tells us – or if it tells us – the risks of voting That people who vote are generally more likely to socialize in groups. The data will be messy, Dowd told me, and any research purporting to show the impact of a particular event on transmission will be imperfect.
This problem is compounded by the fact that the United States has never implemented a consistent and widespread test and tracing program. “If you had a really big program and could look at clusters, you could get very descriptive data. You don’t know for sure, but you could see the patterns more clearly,” said Dowd.
If, instead, we want to know how risky voting is, we basically have to draw conclusions about other research results. For example, We know that group situations are riskier. We know that breathing aerosols is more important for transmission than touching surfaces. We know from studies of cases in places with more intensive tracking programs that it is much safer to be outdoors than in. We know that mask use and distancing reduce transmission.
And taking it to the vote? “My concerns are the lines,” said Cheema. The actual voting time is short and there aren’t many other people around. But long queues allow large groups of people plenty of time to be close to each other, to eat and drink, and to take off their masks. When these lines are inside rather than outside, the risks simply increase. He expects cases of COVID-19 to emerge across the country sometime around November 15. He compared it to the flight risks, where it doesn’t seem to be high to sit in a well-ventilated airplane with a mask on. Risk activity – but sitting around a crowded airport is.
Dowd was less concerned about a vote spike, but agreed that the wait is the hardest part. It’s also the part that voters have the least control over. You can wear a mask. You can choose to stand at a safe distance. You can go away and come back later when the line is shorter. In general, however, you cannot force your polling station to wait outside.
So Cotti is glad that the general election had more choices than the primaries offered. “More choice, more time to vote – that should reduce the number of people who can vote,” he said. “It should reduce the risk.”