Ask Cokie: Exit Polling

Nov 8, 2018
Originally published on November 9, 2018 11:42 am
Copyright 2018 NPR. To see more, visit http://www.npr.org/.

STEVE INSKEEP, HOST:

This week, news anchors have been making much use of this two word phrase.

(SOUNDBITE OF MONTAGE)

UNIDENTIFIED REPORTER #1: We've got a little bit more news out of the exit polls.

UNIDENTIFIED REPORTER #2: Those first exit polls.

UNIDENTIFIED REPORTER #3: We've got more exit polls for you.

UNIDENTIFIED REPORTER #4: We are now getting some information from exit polls here.

INSKEEP: Exit polls - questions put to voters as they leave polling places after casting their ballots. They can give us our first glimpse of election outcomes, tell us who voted, how and why. And exit polls raise new questions, too. David Greene put some of those questions to NPR's Cokie Roberts for our regular segment Ask Cokie.

DAVID GREENE, BYLINE: Cokie, good to have you back.

COKIE ROBERTS, BYLINE: Nice to talk to you on this post-election, David.

GREENE: Yes, it is. We made it. Now we head to 2020, I guess...

ROBERTS: (Laughter) I guess.

GREENE: ...Which is getting already. All right. Let's get right to our first question.

ERIC MENA: Hi. This is Eric Mena from Oswego, N.Y. And my question is, I have been voting in various elections for about 30 years now in both small towns and large metropolitan cities, but I have yet to experience an exit poll. Is it a random occurrence or is it dependent on volunteer availability or otherwise?

GREENE: Cokie, I think Eric is feeling left out.

ROBERTS: Poor Eric. Well, actually, the people who hand out the questionnaires, most of them are paid for the day, some are volunteers. But where they're based is on a couple of things. How competitive are the races in that state? So, for instance, this year, they were conducted for governors' races and Senate races in 21 states, and then around the country, testing basic voter attitudes in an election year. They're conducted in about a thousand of the hundred-thousand-plus polling places. Now, also, with early voting and more absentees, the exit pollsters also do some telephone surveys to try to get to those voters.

GREENE: OK. So our next listener wants to know how good exit polls actually are. Karen Lund writes, how accurate is it? She also wants to know, has accuracy improved over time?

ROBERTS: There's been a lot of debate over the years about this. In general, exit polls of overcounted college-educated people and undercounted rural voters. Again, that's the people who self-select to answer these pollsters. I've always rolled my eyes when I see the percentage of voters who say they have graduate degrees because it's so much higher than the population as a whole. The effect of that is to tilt towards the Democrats. There was a concerted effort this year to correct that. And, in fact, the percent who said they had college and graduate degrees was much closer to the national average.

But still, in the early exits this year before the actual vote totals were added in - and they do get added in - it showed Democrats leading in Senate races that they actually lost. So it isn't totally fixed. But look, David, the important thing to keep in mind is that these polls are much more useful for analysis than for prediction because of the sheer volume of the number of people surveyed.

GREENE: That's amazing because I feel like we're all glued to these exit polls as the day is going on, thinking they're going to predict something. You're saying that that's just not the case.

ROBERTS: Well, they might. But it's a mistake to look at them for that because what they're really wonderful at is telling us what issues the voters care about, what the demographics of the voting population are, things like the very substantial gap between the way men and women voted this year. Those go far beyond informing just who's going to win. And then we journalists use that data to interpret the elections. But they're also a very key source of information for how Congress sets priorities. So, for instance, after Tuesday, even if members thought they knew that health care was driving many House votes, now they know it, and they are very likely to respond to it.

GREENE: All right. Interesting stuff, Cokie. Thanks, as always.

ROBERTS: Good to talk to you, David.

GREENE: That is commentator Cokie Roberts. And you can ask Cokie your questions about how politics and government work, just send us an email - askcokie@npr.org. That's the address. Or you can tweet us. Just use the hashtag #askcokie. Transcript provided by NPR, Copyright NPR.