Play Live Radio
Next Up:
0:00
0:00
0:00 0:00
Available On Air Stations
March is Women's History Month!

Exit Polls Sample Precincts To Predict Election Victor

RENEE MONTAGNE, HOST:

This is MORNING EDITION from NPR News. I'm Renee Montagne.

STEVE INSKEEP, HOST:

And I'm Steve Inskeep.

When voters go to sleep on election night, they have usually heard reports on who's won the election. Many people are devoutly hoping that that will be the case on Tuesday night. But not all the results are actual results. Some are vote tallies, but most are projections based on exit polls and other data collected by pollsters.

Andy Kohut - a guru of political polling - is here to explain how exit polls work. He is president of the Pew Research Center.

When we think about exit polls, the exit polls that we'll be hearing about on election night, who's doing them?

ANDY KOHUT: They're being done by a consortium of networks. The three television networks, AP, CNN come together, they hire a firm called Edison-Mitofsky. Then, along with the networks, other news organizations, such as NPR, or major newspapers, buy a subscription, but they are not full partners. So this is a media-funded, sponsored, collective effort.

INSKEEP: So I imagine there's a pollster standing outside of a polling place somewhere asking people as they come out how they just voted. How many people are doing that across the country?

KOHUT: Oh, I would say it's probably in the neighborhood of 600 precincts are being sampled across the country. There are 350 precincts being sampled for the national exit poll, the poll that projects the popular vote. And then there are exit precincts in the key states or the states in which surveys will be produced and independent estimates will be made for Ohio or Virginia. And not only the swing states, but for the approximately 31 of the 50 states.

The exit pollsters organization will report on opinion in these states and make projections. The exit poll has two purposes. It's to explain the vote, but secondly to provide information for the news organizations, the networks specifically, to make projections. Projections of the winner before all the votes are counted.

INSKEEP: So they're used in projections of who wins. We know they can go wrong, because people - many people will remember 2000 when there were incorrect calls of who had won Florida. It turned out to be a disputed election.

But let me ask about that other thing. Explaining the vote. Explaining why America voted the way it did. How effective are the exit polls in doing that?

KOHUT: Very effective.

INSKEEP: Really?

KOHUT: They give us a sense of the mood of the electorate. They tell us why voters voted as they did. What were the important candidate traits that drove choice? What were the issues that drove choice? When did people make up their mind? Who voted for whom? What's the nature of the gender gap or the racial gap? So it tells us so many things about how America came to make a decision about this choice.

INSKEEP: Does that end up being politically significant, going forward, because you'll end up having a bunch of pundits seizing on this poll and saying this is what the election means. This is what America wants the next president to do?

KOHUT: Absolutely. For example, in 2004, there was a poorly worded exit poll question about morality. And there was a misunderstanding of what that response meant. And the first reactions were Bush was chosen on the basis of the morality factor. Well, as it turns out, that wasn't the case.

INSKEEP: It wasn't conservative social issues, say, that drove the election. It was about something else.

KOHUT: Perceptions of national security and more confidence in President Bush than in Senator John Kerry.

INSKEEP: But you said 31 out of 50 states are involved in this? What happens in the other 19?

KOHUT: In the other 19 states, there will be no reporting in detail about what happens in Kansas or some place, but Kansas will be represented in the national exit poll.

INSKEEP: And also, if we're trying to get a sense of what America thinks, are exit pollsters finding some way to talk to people who have already voted early?

KOHUT: Absolutely. Part of the exit polling now is not really polling at the election sites. They're based on telephone surveys with people who have already voted. And we're guesstimating that anywhere from 35 to 40 percent of the electorate will vote before Election Day.

INSKEEP: Do you think that anybody who has access to the exit poll results will know who won well before the networks actually call it?

KOHUT: If it is as close as it appears to be, maybe not. But generally we do. And certainly in 2008 we did. In 2004, we had an exceptional situation. The exit poll looked wrong. In fact, here at NPR, we decided at 6:30 that there was a problem with this exit poll and we weren't going to rely on it. And ultimately, the exit pollsters came back to us and said, yes, we have a problem. We're correcting it. And at eight or nine o'clock, we had better set of data.

INSKEEP: So there might be some drama behind the scenes as people are waiting at home to hear what the networks have to say.

KOHUT: Potentially, I mean this election has a high potential for drama.

INSKEEP: Andrew Kohut of Pew Research Center, thanks for coming by.

KOHUT: Happy to be here. Transcript provided by NPR, Copyright NPR.