U.S. Markets closed
  • S&P Futures

    +8.00 (+0.22%)
  • Dow Futures

    +63.00 (+0.21%)
  • Nasdaq Futures

    +40.00 (+0.33%)
  • Russell 2000 Futures

    +7.50 (+0.41%)
  • Crude Oil

    +0.30 (+0.66%)
  • Gold

    +4.00 (+0.22%)
  • Silver

    +0.09 (+0.38%)

    +0.0013 (+0.1074%)
  • 10-Yr Bond

    -0.0040 (-0.45%)
  • Vix

    -0.39 (-1.80%)

    +0.0008 (+0.0576%)

    -0.1160 (-0.1111%)

    -105.02 (-0.56%)
  • CMC Crypto 200

    -2.31 (-0.62%)
  • FTSE 100

    -41.08 (-0.64%)
  • Nikkei 225

    +148.84 (+0.57%)

‘Expectations set the narrative:’ Marist poll director on election polls

Yahoo Finance’s Alexis Christoforous and Lee Miringoff, Director of the Marist Poll at Marist College, discuss election poll expectations.

Video Transcript

ALEXIS CHRISTOFOROUS: Well, the pollsters, they needed to shine this election day, especially since the industry needed a big night to regain the public's trust after what happened during the 2016 election. Did the pollsters get it wrong? Here to talk about it is Lee Miringoff, director of the Marist poll at Marist College.

Boy, oh boy, those critics can really be tough, Lee. Thanks for being with us and agreeing to do the interview. So I mean, is the criticism warranted, do you think? Did the pollsters get it wrong?

LEE MIRINGOFF: Well, I think it's an expectations problem again. The expectations contributed to by the forecasters that there's a 90% chance that Joe Biden would win, just like they were talking when Hillary Clinton would four years ago. And that expectation really sets the narrative in place.

Our numbers, we did four battleground states. Two were actually very good. The other two were within the margin of error. But there's a lot of pollsters who had double digit misses this time. And I think that has to do with it was a very, very tough polling environment because we had people voting, unprecedented numbers. We had people voting in-person early. And we had people voting on Election Day itself.

And those three groups had very different preferences. And we didn't know what the denominator was. How big would each one be in our recipes? So it was a very hard environment. Some of the polls did well, but some of them touted it in the wrong direction. And that is open for criticism.

ALEXIS CHRISTOFOROUS: Do you think the time has come for pollsters like yourself to do a little-- is to step back and do a little self evaluation? I mean, because demographics cannot always be divided into neat slices of pie. People don't always vote in tidy groups. So what do you do to sort of build the public's trust for the next election? We're not even through with this election, but for the next election.

LEE MIRINGOFF: Well, I think it's fair to say we should really wait until all the counting is completed to see where the errors were, how great errors different pollsters had. There's a lot of pollsters who were not using scientific approaches. And science is messy. I mean, you know, it's hard when you have this difficult environment.

What we're doing, because we're always want to improve what we do, is we're recalling last night and today, the people we spoke to in the four battleground states that we polled, find out, did they, in fact, vote, how they voted-- very crucial-- and who they voted for. And so we're really doing a recall method to try to see, did what we-- well, did what we thought was going to happen really happened, based on the people we talked to?

So it's sort of an exercise in trying to improve down the road. 2018 was good. 2016 was a mixed bag. This time, there were some real outliers. Some people did well. [INAUDIBLE] in Iowa nailed it at 7 points for Trump, which turned out to be exactly right. But you saw some polls in Wisconsin, 16, 17 points. That was really way off the target.

ALEXIS CHRISTOFOROUS: Mm-hmm. I mean, I get that, you know, early voting, absentee ballots, a record number of them, definitely impacted the ability to exit poll more accurately. But did bad sampling at the end of the day skew the findings? And how do you deal with that going forward?

LEE MIRINGOFF: Well, as I say, you have to really make sure that the approach you're taking does get the right number of cell phone people, does get the right number of landline people. You want to call at different times. You want to get callbacks. You want to have all the opportunities to avail themselves.

Because although, statistically, the differences aren't really that enormous, editorially, they really are. So if you have someone winning by two points in Florida and the other person wins by two points, that's, like, a big deal in terms of news, even though, statistically, it's not. But we have to finetune it down.

And I think sometimes, you know, pollsters and the media, we like to talk about momentum and slippage and the horse race and who's ahead and who's behind and all those kinds of things.

And maybe if we talked in terms of what the error margin really was, which is both directions, so if you have a four-point error margin, it's really an eight-point range, then you don't necessarily have everybody ahead if they're three or four points ahead. Because they're not really ahead. They're just maybe tilting that way.

And I think we try to do that. But everybody gets caught up in, you know, who's ahead and by how much? And that kind of precision gets us into trouble as an industry. That's why you're asking me to take one for the team today.

ALEXIS CHRISTOFOROUS: Based on your polling at Marist, were you surprised to see the razor thin margins we are seeing in Georgia and Arizona?

LEE MIRINGOFF: Well, Arizona, we had dead even. I think it was 48-48. So that looked like an outlier. But today's outlier is tomorrow's conventional wisdom. And it turned out that one was one of the best-- actually, that was the best state we have, in Arizona.

We're seeing a big difference between the Rust Belt and the Sun Belt. And what Biden did and what Trump did in those two regions really speaks to the future demography of what the strategies are going to be in terms of the polling, in terms of the campaigns. Because we really have two different groupings of Americans. And clearly, Biden was able to kind of resurrect the Rust Belt, just barely. But he made a few inroads into the Sun Belt areas in Arizona, for example.

We'll see in the future. I mean, this is why it's always a moving target, which makes it-- I've been doing this a long, long, long time. And it's always exciting to try to kind of figure out where the electorate is going and try to be able to identify it. But it's a science, but it's an art, too. And when you get close to Election Day, you've got all kinds of folks joining in with all kinds of numbers. And that can get understandably confusing for the consumer of all this.

ALEXIS CHRISTOFOROUS: You know, [INAUDIBLE] were feeling that we might see him run again in 2024. Could he ever win the party's nomination again after this debacle?

LEE MIRINGOFF: You never say no about Donald Trump. I mean, he wasn't supposed to win the first time. And he certainly wasn't supposed to be a president if you use the conventional wisdom. He defies that. I mean, you could see with the turnout this week, you know, it was huge, but on both sides.

Some people thought maybe Trump was getting a little boring. And, you know, the Supreme Court was already in his direction, like the conservatives wanted to. And maybe he didn't have the temperament and et cetera, et cetera, and the coronavirus, of course.

But his folks showed up, as well as Biden's folks. So we had a-- you know, the water kept all the boats rising proportionately. So the answer to that question, long answer, he will be four years older. I like people in their 70s who do run for office because it makes me feel younger. I will be in my 70s the next presidential campaign. So if it ends up Biden and Trump, I'm still the kid.