FiveThirtyEight

Welcome to Pollapalooza, our weekly polling roundup.

Poll(s) of the fortnight

Because of the coronavirus pandemic, many states have taken steps to make it easier for people to vote by mail. (We’re tracking them all here.) Absentee ballots have constituted a majority of the votes cast in most primaries since the pandemic struck, and polling suggests a record number of people will vote by mail in the general election as well.

However, amid saturation coverage of problems with the U.S. Postal Service, new polling from CNBC/Change Research suggests that the number of Americans planning to vote by mail has ticked down. In early August, 38 percent of voters in six battleground states said they planned to vote by mail. But in the pollster’s just-released Aug. 21-23 poll, the number of voters in those states saying they planned to vote by mail was down to 33 percent. Among all voters nationwide, the share planning to vote by mail went from 36 percent to 33 percent — although that drop was within the poll’s margin of error.

Other recent polls agree that about a third of voters intend to vote by mail this year. But not all voters plan to do so in equal numbers. Democrats are much likelier than Republicans to say they will vote by mail — which makes sense given that Democrats also tend to be more supportive of mail voting. (By contrast, the Republican standard bearer, President Trump, has repeatedly and inaccurately assailed mail voting as ripe for fraud.)

According to a recent NBC News/Wall Street Journal poll, 30 percent of registered voters said they planned to vote by mail, and 43 percent said they planned to vote in person on Election Day. But among Trump supporters, only 11 percent said they planned to vote by mail, and 66 percent said they planned to vote in person on Election Day. Among Joe Biden backers, 47 percent said they planned to vote by mail, while only 26 percent said they planned to vote in person on Election Day. (The share who said they would vote early in person was consistently 20-21 percent among all three groups: Trump supporters, Biden supporters and voters overall.)

If this holds, it would mean votes cast on Election Day would skew heavily toward Trump, and votes cast by mail would skew heavily toward Biden. This has serious implications for … well, democracy. First, Trump could argue the mail ballots (which, remember, could account for most of Biden’s votes) were fraudulent and thus should not be counted. Although it’s unlikely they’d actually be thrown out, this would damage the credibility of the election in the eyes of many Trump supporters. Second, it could mean the first votes counted on election night will be disproportionately good for Trump, who might claim victory based on incomplete returns. It might not be until days later, after a good chunk of the Democratic-leaning mail vote is counted, that Biden pulls ahead.

Let’s do a quick-and-dirty exercise to show what I mean. In the NBC News/Wall Street Journal poll overall, Biden led Trump by 9 percentage points among registered voters. But Biden led Trump by 63 points (!) among voters who planned to vote by mail, and Trump led Biden by 33 points among voters who planned to vote in person on Election Day. If this kind of partisan split occurred in every state, Biden would win the mail vote in all 50 states — from Alabama to Wyoming — and Trump would win the Election Day vote in all 50.

To be clear, I don’t expect Trump to be leading in all 50 states when the first results are reported on Nov. 3. For one thing, the partisan split between Election Day and mail votes will surely vary from state to state; I don’t literally think Trump will win the Election Day vote in every state, nor Biden the mail vote in every one. For another, Election Day votes aren’t the only results we’ll get on election night: We’ll also get early in-person votes in states that have early voting, which should make things look better for Biden. (Biden led Trump by 12 points among early in-person voters in the NBC News/Wall Street Journal poll.) And some mail votes will probably be counted by the time we go to bed as well.

Nevertheless, the gap between the Election Day and mail vote could be huge this year, which should serve as a stark reminder not to overreact to election-night returns. We could — not necessarily will, but could — have absurd initial results like Trump winning Massachusetts or Illinois. That would not mean he has won the election; it takes days, sometimes even weeks, to count every mail ballot, so we may not know the identity of the next president until mid-November.

Other polling bites

Trump approval

According to FiveThirtyEight’s presidential approval tracker, 42.2 percent of Americans approve of the job Trump is doing as president, while 54.3 percent disapprove (a net approval rating of -12.1 points). At this time last week, 41.8 percent approved and 54.2 percent disapproved (a net approval rating of -12.4 points). One month ago, Trump had an approval rating of 40.2 percent and a disapproval rating of 55.8 percent, for a net approval rating of -15.6 points.

Generic ballot

In our average of polls of the generic congressional ballot, Democrats currently lead by 7.3 percentage points (48.3 percent to 41.0 percent). A week ago, Democrats led Republicans by 7.4 points (48.4 percent to 41.0 percent). At this time last month, voters preferred Democrats by 8.5 points (49.5 percent to 41.0 percent).

Make sure to check out FiveThirtyEight’s 2020 presidential forecast, as well as all the polls we’ve been collecting ahead of the 2020 elections.


Filed under