What exactly are the polls telling us about the presidential election

With less than twenty-four hours until the presidential election, all eyes are on the polls. But maybe they shouldn’t be.

WKAR’s Michelle Jokisch Polo spoke with Michigan State University Political Scientific Eric Gonzalez Juenke to learn why polls may not accurately predict the behavior of voters or the outcome of the all-important electoral college.

Interview Highlights

On why the polls are predicting a tight race for the presidential election

On why the polls aren't built to be precise

On whether we should trust the polls

Interview Transcript
--

Melorie Begay: With less than twenty four hours until the presidential election all eyes are on the polls. But maybe they shouldn’t be. That’s according to Michigan State University Political Scientist Eric Gonzalez Juenke.

WKAR’s Michelle Jokisch Polo spoke with him and learned why the polls may not accurately predict the decisions of voters or the outcome of the all-important electoral college.

Michelle Jokisch Polo: In the weeks leading up to the 2016 election, polls across the country predicted an easy win for then-Democratic presidential nominee Hillary Clinton. But we know that’s not what happened.

The polls were right about the popular vote, but they missed the mark on the Electoral College. Has anything changed since then?

Eric Gonzalez Juenke: Political scientists don't think that they missed that polling. This wasn't like a historic polling miss. It was a normal polling miss. What's interesting about 2024 is we're saying the same thing the polls are telling us: the race is close, but we're a normal polling miss away, and by normal this means like three to five points in each of these battleground states. That’s what happened in 2016 and so voters kind of don't trust the polls because of that. But if you look historically, it's just a very normal polling miss, because pollsters can't really tell who's going to show up on election day, and so they have to make guesses about who's going to show up on election day. Those guesses are how they weight the people in their polls to say, I didn't get enough Republicans in my poll, and so I'm going to have to weigh them a little more. I didn't get enough men in my poll, so I need to weigh them a little more. The group that is probably driving a lot of this right now is folks without a college degree, and so if they're not getting a lot of those folks able to answer their surveys, then they're weighing them, and that weight really can drive some of these polls. So it can make a tied race look like a four point Trump win or a four point Harris win. So we're relying on polls, but polls at this point, because the race is so close, can't they're just not built to be that precise.

Jokisch Polo: So, are you saying there’s an error in the polls?

Juenke: Yes, but a normal error, meaning like this. The polls aren't built to be that precise. They can't tell you exactly what's going to happen, but they can say, and this is that margin of error that we always hear about, all the polls are close. That means these are all coin flips. So, flip seven coins, three of them are going to come up for Harris and four of them are going to come up for Trump, or four for Harris and three for Trump and that could happen. So we could be in a world where one of the candidates barely gets over the 270 because those coin flips were all coming up differently. But what’s probably more likely is because the polls don't know who's going to show up on election day, that is probably common or systematic across all the states. I’ll give you some examples. Let's say, because of Dobbs and because of the abortion issue, more women show up than the pollsters are expecting, and women who do show up vote for Harris more than Trump than they're expecting, then that systematically is going to tip all those coin flips in Harris's direction, and she could very well sweep all or most of the states. Or let's look at the flip side. Let's say Trump successfully turns out more of non-college educated, maybe more rural, or people who are kind of lower income, who don't necessarily turn out to a lot of elections historically, let's say he somehow has figured out a way to get them to turn out this time and that's a surprise, because the pollsters are like, well, we don't know how many of those folks are going to we can guess, and if they get that wrong, then that's also going to be common across all those states. And so then all the coins flip to his direction. So, what looks like a 50/50, race ends up being something that looks like a blowout.

Jokisch Polo: So at this point, should we trust the polls at all?

Juenke: Yes, yes, yes, yes, yes. What we try is to imagine a world without polls and if we had a world without polls. We would go back to and we know what that world looks like, because there were times most throughout American history, we didn't have polling, and it's just people guessing. It's just people in the news, people who had a lot of readers, people who people trust, just kind of talking about yard signs or talking about what they heard from insiders on the campaigns and all this stuff, and it's not useful. And the polling industry has gotten really, really good at what they do.

Jokisch Polo: Thanks for being here, Eric.

Juenke: You’re welcome. Good luck with everything.

Jokisch Polo: That was Eric Juenke, political scientist at Michigan State University.

Begay: And that was WKAR’s Latinx Stories Reporter Michelle Jokisch Polo. This is WKAR.

 This conversation has been edited for clarity and conciseness.