What If the Polls Are Wrong?

What if the polls are wrong? The thought keeps nervous Democrats awake at night. The polls in crucial Midwest states were off by around five points in Trump’s favor in 2016, and Democrats have armed themselves with a don’t-get-fooled-again attitude toward the supposedly fickle numbers throughout the 2020 race. There have been just enough scary polls—like Saturday night’s Iowa poll from the respected Ann Selzer showing Biden trailing by seven points (the average before then had the race dead even)—to fuel nightmares of another late-breaking wave of working-class white voters. As of this morning, 538 has Biden at 90% to win, but last night, Nate Silver laid out an excruciating Trump victory scenario revolving around Pennsylvania polls being off.

But 2020 is not 2016 in a number of important ways. As the race closes, Joe Biden’s lead is significantly higher than Hillary Clinton’s was—he’s more than twice as far ahead—and while 2016 was a volatile race in which Clinton’s lead expanded and contracted right up until Election Day, Biden’s lead in 2020 has been a stable one, slowly expanding to his current nine points over the course of the last tumultuous nine months. And yet, the fear remains. What if? What if the polls are wrong?

Well, one thing that might happen if the polls are wrong is that Joe Biden will win by a lot. No, more than that. Like, a lot a lot. It’s important to remember that 2016 was not the only election in which the polls were imprecise. You only have to go back to 2012 to find an example of a time when polls missed in favor of the Democrats—national polls gave Barack Obama a slim edge before the election, but he eventually won by four points. If election night is like 2012, Biden’s nine-point lead would turn into 13, complete with comfortable margins in more than enough states to have the presidency sewn up (with a margin big enough to nullify Republicans' efforts to litigate the result) by bedtime.

A 2012-level polling miss in Biden’s favor would mean that he notches comfortable wins of between three and five points in Florida and North Carolina, which have the practical and legal infrastructure in place to count ballots quickly—meaning that as long as they aren’t super close, we’ll know their results on election night. A Biden win in these two states effectively ends Trump’s presidency. And a comfortable Biden win by margins unforeseen in this year’s polling means there’s probably a lot of other good news in store for Democrats as well.

In addition to ensuring wins across the upper Midwest, where Biden’s polling already gives him robust leads in Michigan and Wisconsin and a solid lead in Pennsylvania, a miss in his direction would put Arizona firmly in his column, along with more closely contested states like Georgia and Iowa. Depending on the magnitude of the miss, it would also give him a fighting chance at Ohio and, stunningly, a real chance to win the Democrats’ great white whale, Texas. Right now, 538 gives Biden a 36% chance of winning the state, or a 7% higher chance than it gave Trump of beating Hillary Clinton four years ago.

But it’s not just that Biden can run up the score in the Electoral College. A polling miss in Democrats’ favor could also have massive implications for which party controls the Senate, and whether a President Biden gets to enact his agenda. Right now, there are a number of extremely tight Senate races, with Iowa a toss-up and Democrats narrowly favored to flip two seats in Maine and North Carolina, and to win comfortably in Colorado and Arizona, while only losing Alabama. That would give them a narrow 51-49 majority. However, 538 gives the Democrats at least a 20% chance of winning in five more races. A polling error that misses Democratic strength means that instead of sweating the presidential race, we should keep in mind that Senate seats in Georgia, South Carolina, Kansas, Montana, and even Alaska will come right down to the wire (a second Georgia Senate seat is almost certain to end up in a run-off in January as well).

Most of the time, polls are likely to be equally off the mark in either direction. In general, they are imprecise tools, so the best bet is to average the polls and think about that number as the middle of a range where the election might end up. This year, however, there are at least a handful of reasons to think variance might run in Biden’s favor. This isn’t to say that he will perform better than his polling, only to say that it might be a little more likely that he will than that Trump will.

For starters, there’s the composition of the polling average. As Charlie Cook of the Cook Political Report argues, Biden’s polling in higher-quality polls actually looks closer to the 9–10-point lead range than where 538 has it, at just under 9, or where Real Clear Politics has it, at between 7 and 8. That doesn’t necessarily mean the higher-quality polls are right, but you’d rather be on Biden’s side of the equation.

Then there are the district-level polls, both public and private. In addition to all the polling done by media companies, there is a host of congressional district–level polling that most people never see. Parties use it to make decisions about where to spend resources and when, and campaigns use it to look for trends that might be unique to their districts. According to Dave Wasserman, also of the Cook Political Report, those polls show Trump faring poorly, specifically in Pennsylvania, where they have him running somewhere between seven and 12 points behind 2016. That would be consistent with a race where he is down at least nine points. And it’s also worth noting that these were the kinds of polls that, in 2016, led Wasserman to suggest that despite national and state-level polling giving Clinton a substantial edge, she might in fact be in a very close race.

Finally, there’s the question of turnout. You may have noticed over the past couple of weeks that a lot of people have already been voting. Thanks to the pandemic, most states have made it easier to vote by mail or vote early than ever before. And people are voting before Election Day in historic numbers. Also, while mail/early voting did not use to be a partisan issue—Democrats and Republicans availed themselves of the option roughly equally—this year the president has spent much of the summer and fall railing against it. Thus, in 2020 it has become a method favored by Democrats, with Republicans more likely to show up to vote in person on Election Day.

This is very likely to be the case. But at the current moment, what we definitely know is that early voting has shown extremely high turnout—over 91 million people have already voted—while we only expect Election Day turnout to be very high. So while we already know that predictions about record Democratic turnout have been proven true, it’s still possible that predictions about Republican turnout could end up being overstated. Again, it’s important to emphasize that this is not likely to happen, but you’d rather be Biden than Trump in this situation.

Realistically, there’s nothing that will calm nervous Democrats down until they see Joe Biden getting sworn into office with their own eyes. The polls say that’s likely to happen. And for all the focus on the possibility that it doesn’t, it’s just as likely that he’ll be swept in on the strength of a margin of victory north of 12 points—the kind of historic result that hasn’t been seen in a century. So, yes, the polls can be wrong. But that’s not always a bad thing.

Originally Appeared on GQ