Subscribe to What Next on Apple Podcasts for the full episode.
The 2020 election was not supposed to be this close. Thousands of votes still need to be counted before a winner can be declared, but this much is already clear: Many of the projections for battleground states were wrong. Again. Like they were wrong in 2016, though perhaps for different reasons. On Thursday’s episode of What Next, recorded midnight Wednesday, I talked to Slate politics writer Will Saletan about what the polls missed and what Democrats can learn from it. A portion of our conversation is transcribed below, condensed and edited for clarity.
William Saletan: OK, I was the person at Slate who was assigned to write the “Don’t Panic” piece. Everyone else at Slate was panicking. I was not because I knew, of course, that I could trust the polls. I told myself, and I told everyone else, that after 2016 these pollsters looked at the data, they looked at their mistakes, they weren’t going to let another Hillary Clinton implosion happen to them again. So, going into this election, Joe Biden had leads of 5 points in Pennsylvania, of more than 8 points in Michigan and Wisconsin. And I told everybody: don’t worry about Michigan and Wisconsin. Those leads are just way too big for a polling error to cost Biden the states. And darn if he didn’t come close to it. [laughs]
Mary Harris: Given how close it was, or is, I wonder if you have shifted your opinions about polling.
I can’t defend how badly the polls appear at this point to have missed the state-by-state results in some of these states. I mean, if the numbers turn out the way it appears they will turn out, the ballot numbers in Michigan and Wisconsin, then that’s a miss of novel proportions. So I can’t defend the miss, but I will defend the science of polling, because what pollsters do is they come in with a theory about what the electorate is going to look like, and then they encounter the data of the election, and often they’re surprised. And what you do when you’re surprised in science is you adjust your theory to fit the data and then you try again with a slightly different hypothesis about who’s going to show up in the election. So the polls will get better, and they have gotten better. They got better in 2018 as a result of adjusting for their errors in 2016. And we still do not know and will not know for some time in what way the pollsters screwed up calculating who was going to show up in this election.
There’s been some chatter that there’s just something about Donald Trump being on the ballot that messes with these polls, because the polls were off in 2016, and then they seemed to recover and be a little more on target in 2018 with the midterm elections, where you could kind of vote your heart and try to send a message to Trump but Trump himself wasn’t on the ballot. And now here we are in 2020 and it seems like the polls are off again. I wonder if you think about it like this could be a Trump problem somehow, not a polling problem.
That is possible, and it is certainly interesting and worth thinking about, that it happened when Trump was on the ballot and didn’t happen when he wasn’t. But I’m sort of interested in the theory that what we are experiencing right now is an ongoing political realignment and that Donald Trump simply attracts a different kind of voter from what the Republican Party has traditionally attracted. And perhaps pollsters are trying to learn how to anticipate and measure this kind of voter.
Looking at the exit polls, you see the president putting together one of the most diverse collections of voters and supporters in many years for the Republicans. The Republican Party that voted for Trump seems to be really the party of the working class, which, of course, was the traditionally Democratic voter. And then you look at who voted for Biden and you’re looking at people who are wealthier, more college educated. I wonder if you see it the same way.
Well, the exit poll data that we have so far are confusing. In some ways, Biden did better than Hillary Clinton at attracting working people, getting non-college white voters, getting voters from households that made less than $100,000, getting union households. That was sort of Joe Biden’s job in the Midwest, and he did it. But clearly he got swamped by a whole lot of other people who showed up and just drowned out those gains.
The early data does seem to show demographic shifts. In 2016, Trump’s biggest base of support was white—in particular white women, who delivered for him in a major way. The gains the president has made with communities of color are real, even if they aren’t big enough to be decisive.
I think Democrats were surprised by how much Trump was able to add to that base with support from Blacks and Latinos. And when I say Latino, I mean with an O, I mean men. Trump does better with men in general, but he did better than Hillary Clinton in both categories, Black and Latino. He ended up getting, according to the preliminary exit poll numbers, 18 percent of Black men and 36 percent of Latino men. I certainly thought, and I think a lot of my friends and colleagues thought, look, Donald Trump is a bigot. You can argue about whether he is personally racist, but he is certainly a racial demagogue. He is certainly out there talking about Mexican immigrants and Somalis and how Cory Booker, the Black senator from New Jersey, is going to bring low-income housing projects to your suburb. So the dog whistles are all over the place with this guy, and you’re thinking to yourself, therefore, Black people and Latinos are going to vote against him. But in fact he did surprisingly well with them, and that is certainly a result that a lot of Democrats need to reflect on and ask themselves. You know, they just cannot count on how repellent the opponent is, how overtly racist he is, to drive Black and brown people, people of color, to vote Democratic. Democrats need to figure out how to deliver a more affirmative message that attracts people of color to the party and its candidates.
You’ve written that, looking at the exit polls, the electorate seems to be more conservative than it was in 2016. And given that you’re actually kind of surprised that Biden seems to be winning the popular vote by as much as he is, what does that mean to you?
The data show from the exit polls that there was a greater ratio of conservative voters to liberal voters in terms of self-definition. Joe Biden did better than Hillary Clinton, and the way he did that, the way he exceeded her results in a more difficult electorate, was by attracting independent voters. Hillary Clinton lost independent voters—narrowly, but she lost them—to Donald Trump in 2016. Joe Biden comes along, faces the same opponent in 2020—who’s now an incumbent, so it’s in some ways more difficult—and Joe Biden wins independent voters by, according to the initial numbers, about 14 percent. And that was decisive. That was absolutely decisive, because if he hadn’t won them nationally that way, he would not have pulled out these states. He would not have pulled out states like Michigan and Wisconsin. So it was hugely important that Joe Biden reached voters in the middle.
It’s kind of a chicken-and-egg thing, though, because Joe Biden was meant to appeal to conservatives, and of course so is Trump. So in some ways, maybe it shouldn’t surprise us that the electorate is more conservative. That’s exactly who these candidates were speaking to.
Yeah. And if you are on the left of the Democratic Party and you’ve just been watching this result and you thought it was going to be a cakewalk—you went along and nominated Joe Biden, the candidate of the middle, because he was going to attract Republican crossovers, he was going to attract independents, he was going to win the election handily—and then you see this absolute nail-biter, you’ve got to be thinking to yourself, man, will you try us next time? Will you try Bernie Sanders or Elizabeth Warren or somebody who really fires up the left? Because then maybe, instead of scratching and clawing for every voter in the middle, every independent voter, every potentially disaffected Republican, maybe next time we could turn out our people, people on the left, and get a more left-leaning electorate, in which case it would be easier. So we don’t really have an election yet to test that theory, but we might get one next time or after that.
Because there’s going to be so much soul-searching among the Democrats after this.
If you are a leftist, if you are a progressive, and you feel like the Democratic Party has been spending too much time in the middle, this nail-biter of an election is kind of your opportunity to say, look, we just tried it your way, and we darn near died when we shouldn’t have. So let’s try it our way.
Subscribe to What Next on Apple Podcasts
Get more news from Mary Harris every weekday.