RACHEL MARTIN, HOST:
Pollsters and media organizations thought that they had learned what they needed to from the 2016 election, but as Tuesday night's results rumbled in, the vote tallies often diverged quite sharply from what the polls and analysts had suggested - again. NPR's media correspondent David Folkenflik has been tracking this and joins us now. Good morning, David.
DAVID FOLKENFLIK, BYLINE: Good morning, Rachel.
MARTIN: So as you looked at all those polls running up to the election and then you saw the results flood in, what was your takeaway?
FOLKENFLIK: Hard not to conclude that it's another black eye for polling and for the news organizations that rely on them. And why do I say that? Well, most of these polls, you read them in newspapers, you see them on TV, you might hear them on our network. They have media branding on them, even those done by colleges or private firms. The polls and the medias are intertwined on this stuff. So when you see a lot of wrong results, that reflects on us, too, regardless of whether Biden or Trump ultimately wins. And it looks like Biden - things are going in his direction. That huge blue wave, I don't think we're seeing it. The Dems were projected pick up seats in the House. They're losing seats in the House.
The idea that Democrats would win back the Senate, I believe that one of the major aggregators of polls said with 65% confidence the Dems will win back the Senate. That sure didn't happen. It doesn't appear to be happening. The Washington Post said that Biden was up by 17 in Wisconsin - 17. The New York Times' survey of reputable polls there had him up by about 10. An NPR poll nationally in mid-October gave Biden a double-digit lead nationally among likely voters. I know we're still waiting on some results, but there's currently maybe about a 2 1/2% difference between the two candidates. It's nowhere near double digits. People thought 2016 was a hot mess, and this may end up being worse for polling.
MARTIN: I mean, to be fair, these polls are usually given some context by news anchors - right? - as snapshots. They're usually offered with margins of error, right?
FOLKENFLIK: They are. And that's true. They're saying they're snapshots in time. There's margins of error, folks. This is where we are now. And yet when people talk about it on TV, they seem to indicate this is where things actually are. They're being used to predict - they're being used to project things that people think they have a fuller understanding of where things are and where things are going to be than they probably do. One of the things that really became popular in the last few years is these aggregation of polls.
If you think of the two Nates, as I think of them, there's a really smart guy named Nate Cohn at The New York Times who does this. There's Nate Silver of ABC's famous FiveThirtyEight site. And these guys blend combinations of the most reputable firms on the national and state levels for rolling averages that they say would smooth out imperfections in polls. On election night, you saw them online basically chiding people for doubting the strength of that model. Silver at one point in particular urged people to filter out distractions and bad information at one point on Twitter as people were doubting the strength of the polls. He wrote, the narrative here is pretty dumb overall in one of his tweets. And here's the thing - these aggregators are meant to provide great clarity, and maybe there wasn't so much clarity to be had.
MARTIN: All right. President Trump has been attacking the polls. He says they're biased, especially those of Fox News that show him trailing Biden. I mean, is there any truth to that?
FOLKENFLIK: Right. Well, he attacks Fox News because he expects a certain kind of loyalty from Fox. And yet Fox, whatever its ideology in its most popular shows, actually has some pretty rigorous polling operations. But he attacks them generally when they don't say what he wants them to say. I don't think it's bias. I think there are flaws in the design. People have assumptions built into their models, and they've tried very hard to overcome them. In each election, they're trying to essentially correct for the mistakes of the last. And yet we're finding that there are new problems. So even as they get more and more refined and perhaps smarter methodologically, there are problems with both the fact that they've got to make up for cellphones instead of landlines and they've got to make up for a lot of voters who have lost faith in polling in the media, whether because we haven't lived up to their expectations or because of years of beating up on the media, particularly by the right.
MARTIN: So we've outlined the flaws of polls. I mean, they are fraught. What's the argument for keeping them?
FOLKENFLIK: Well, without polls, we're flying blind. It's a tool we can use to help figure out what issues to illuminate, which people to talk to. A news organization can't interview 1,500 people over a single weekend. A poll can yield their thinking. But if the media is going to treat it like a Ouija board and confer magical powers on it, that's a mistake. You know, my sense is is think about polls as suggesting what's out there. Think about polls as helping you understand what's on people's mind. But don't use it in the same way to predict the future. I think tonally, you've seen that on the air too much where anchors and analysts seem to give it scientific precision that it simply doesn't seem to have, at least not right now.
MARTIN: NPR media correspondent David Folkenflik, thank you.
FOLKENFLIK: You bet.
NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.