Identifying potential voters is a challenge for pollsters in every election. This year, the coronavirus, voting by mail, and an increase in political commitment may make it even more difficult than usual.
For now, Joe Biden’s nine-point advantage in critical battlefield states is so significant that he is essentially invulnerable to assumptions about participation, according to New York Times / Siena College surveys of states most likely to decide the choice. But Biden’s supporters are much more likely to be concerned about voting in person during the pandemic, and his broad electoral leadership among registered voters could be reduced if his concerns persist in the election.
Overall, a quarter of registered voters in the battlefield states said they would be uncomfortable voting in person. Perhaps surprisingly, these voters are fairly demographically representative of the country. The groups most affected by the coronavirus so far (older people, blacks, Latinos, people in densely populated areas) are generally more likely to say they would feel uncomfortable voting, but not much.
Instead, discomfort with voting is primarily a function of political views. People were asked if they would be uncomfortable voting in person if the election was held during the week they were interviewed in June. About 40 percent of Biden’s supporters said they would be uncomfortable, compared to just 6 percent of Trump’s supporters. This political division transcends demography. A young Biden supporter in a rural area, for example, would be more likely to feel uncomfortable voting than an old Trump supporter in a city, even though the health risk is likely quite low for the Biden voter and potentially quite significant for the Trump supporter.
Most of these voters would go to the polls anyway. But about a quarter of uncomfortable voters, or about 6 percent of the general electorate, said they would feel too uncomfortable to vote in person if the elections were held during the week they were interviewed. This includes 8 percent of all Biden supporters in the battlefield states, compared to less than 2 percent of Trump supporters.
Above all, voters who said they would be too uncomfortable to vote in person for Mr. Biden, 63 percent to 9 percent. Mr. Biden’s leadership among registered voters would be reduced by three net points if these voters stayed home and did not vote by other means.
It is important to emphasize that unexcused absentee voting, in which any voter can request a ballot by mail, is available in all six battlefield states included in the Times / Siena data. Many of these voters would likely manage to navigate the absentee voting process and successfully vote by mail, although it is impossible to say how many.
However, these voters seem less likely to vote under any circumstances. Only 40 percent of those who said they would feel too uncomfortable to vote participated in the 2018 midterm elections, compared to 65 percent of those who said they would still vote. Similarly, only 35 percent of those who said they would be too uncomfortable to vote said they were “almost sure” to participate, compared to 66 percent of all other voters.
Their weaker voting record indicates that many may not vote anyway, coronavirus or not. And less political participation may mean that ballots are less likely to be searched and mailed in absentia.
For pollsters, the coronavirus makes the task of modeling the likely electorate even more challenging than usual.
Most political polls estimate the composition of the electorate in one of two ways: they ask respondents if they will vote in the next election or if respondents voted in the past, according to their voter registration record.
Both approaches have disadvantages. A model based on voting history could be biased if participation patterns change considerably from a previous election. Simply asking respondents if they plan to vote has several problems of its own: Voters tend to exaggerate their probability of voting, and respondents tend to be more politically engaged than demographically similar voters who do not take polls.
At this early stage, neither approach could account for any eventual coronavirus effect on participation. A model voting history certainly would not, although absentee requests and returns could be useful closer to the election. It also seems unlikely that voters in June will give a lot of thought to the coronavirus when evaluating their intention to vote in November, although here again it seems more likely that voters will do so as the vote approaches.
Either way, Mr. Biden would maintain significant leadership among likely voters in the current wave of Times / Siena state battlefield polls using these standard practices. Your advantage is reduced by about half a percentage point, whether you use a model based on voting history or a probabilistic probable voter screen based on a voter’s stated intention (for example, a 90 percent probability if they are “almost certain,” but only a 20 percent chance if they say “not likely.” This doesn’t factor in whether respondents said they would be too uncomfortable to vote in person.
One factor that is not included in the most likely voter screens is enthusiasm. It is certainly possible that the candidate with an enthusiasm advantage has a participation advantage, but an unenthusiastic vote counts as much as an enthusiastic one, and the majority of registered voters appear in the presidential elections anyway.
Still, Biden doesn’t appear to be at a significant disadvantage of enthusiasm: 65 percent of his supporters said they were “very excited” about voting in November, compared to 66 percent of Trump supporters. Importantly, the survey question asked if respondents were excited about the vote, not if they were excited about the candidate they supported, where Mr. Biden appears to be at a more significant disadvantage.
Perhaps more surprisingly, Biden also enjoys a nine-point lead in a model based on the likely electorate’s vote history, despite the fact that this electorate is whiter, older, and more Republican than the electorate on the battlefield in Overall, reflecting surprising strength among voters with a strong voting record. You also enjoy a nine-point lead using the standard Times / Siena likely voter approach, which combines voting history and self-reported intent.
The lack of a general gap between registered and likely voters masks a modest underlying regional variation. Mr. Biden is at a modest disadvantage in participation in the Sun Belt states, where he depends on young, non-white voters, who are two low-participation groups. Trump, on the other hand, appears to be at a slight disadvantage among potential voters in the Rust Belt, where he depends on the support of white voters without a college degree, another low-turnout group.
Of course, all of this could change before the elections. Mr. Biden’s position among likely voters may deteriorate if he loses ground among older voters with a high voter turnout or, less likely, voters with a higher level of higher education. And anyone could be more or less likely to vote, even if nobody’s opinion of a candidate changes. For these reasons, many public pollsters do not report the results to prospective voters until after Labor Day.
The coronavirus may require even greater caution in this cycle.
