Douglas Schwartz has been the director of the Quinnipiac University Poll since 1995. He began his career as CBS News survey associate election night analyst for the late 60 Minutes correspondent Ed Bradley. We spoke with him in late summer about the upcoming election and what polling got right and wrong about the 2016 election.
Let’s start off with the most important question: According to the polls, who will be our next president?
It’s important to emphasize that polls are just a snapshot in time and a lot can change between now and the election. Right now if you look at the national and battleground-state polls, Biden would have to be considered the favorite. But you can’t count Trump out; there’s plenty of time between now and the election.
Were the polls similar in 2016?
You’d probably have to say that Biden’s numbers are stronger at this point. Whether you’re looking at the matchup or his favorability rating versus Hillary Clinton’s or his honesty rating versus Hillary Clinton’s, you would say that Biden is in a stronger position than Hillary Clinton was in 2016 at a similar point in the campaign.
Speaking of 2016, how has polling changed since that election?
There is one methodological change some state pollsters have made. Pollsters in some of the key battleground states underestimated the share of the vote from white non-college-educated voters and that was a big part of Trump’s base. Some of those state pollsters are now weighting by education to make sure they get the correct percentage of white non-college-educated voters as well as white college-educated voters in their samples. We’ve always weighted by education, and most of the major national polls have always weighted by education, but there were some state pollsters that didn’t.
So polls are a snapshot and not always predictive...
There’s a big difference between what a poll is and what the forecasters do. Forecasters predict a couple of months before the election this is the percentage chance that Joe Biden will win the election, this is the percentage chance that Donald Trump will win the election. But that’s not what polls do; polls just measure at a very specific moment in time what people are thinking. So it’s really a different function. It’s also important to remember that in 2016 the voters actually did change in the final week and final few days of the campaign. Exit polls showed that those voters that were the late deciders, the bulk went for Trump, especially in some of the key swing states. The national polls were pretty much dead on — they had Hillary Clinton winning the popular vote by three points on average. She ended up winning by two points.
So in some ways was it a perception of what the polls were showing versus what they actually were showing?
Some of it was also a case of misinterpreting polling. Some of the polls were showing a very close range, but the reporting of it sometimes would forget there’s a margin of error with polls. For example, we had in Florida that Hillary Clinton was ahead by one point in our final poll; she ended up losing by one point. We were on the wrong side but the way it was getting reported was that Hillary Clinton has a lead in Florida, but she didn’t. We never said she had a lead; our headline was “too close to call.” There’s a margin of error with every poll.
Yeah, one point means it can go either way.
Exactly! It was a toss-up.
What is the most surprising result you’ve seen from a poll in your time at Quinnipiac?
I don’t know of one specific surprising thing that I can say, but overall we’ve seen President Trump’s job-approval numbers be remarkably consistent, hovering around the 40 percent mark for almost his entire term in office, and that’s pretty rare. Usually you’ll see some movement, but in his case, it has been a very tight band since he’s been in office no matter what’s happened. No matter what event happens, no matter what he does, it pretty much stays right around 40 percent.
If you’re a Biden supporter, what should you be worried about from the polls this year? And if you’re a Trump supporter, is there any reason for optimism?
If you’re a Biden supporter and you’re looking at the numbers showing he’s ahead, you don’t want to be overconfident, and if you’re a Trump supporter you don’t want to think all is lost. Trump gained ground against Hillary Clinton; we’ll have to see if he’s able to gain ground against Joe Biden. In 2016 one of the reasons people had the perception that the polls were wrong is that Hillary Clinton won the popular vote but lost the Electoral College. So that potential exists this year, and when looking at the national polls, keep that in mind.