NPR logo

Deciphering Presidential Polls

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript
Deciphering Presidential Polls

Election 2008

Deciphering Presidential Polls

Deciphering Presidential Polls

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript

With less than two weeks to go before Election Day, the polls suggest that Democrat Barack Obama leads Republican John McCain. It's a little tougher to tell whether Obama is ahead by a lot, or a little.


And with the finish line less than two weeks away, the polls suggest that Barack Obama is ahead. It's less clear whether he's ahead by a lot or a little. Here to explain how to interpret all the polls is NPR's Washington editor, Ron Elving. Good Morning.

RON ELVING: Good morning, Renee.

MONTAGNE: Now, how are these very different numbers possible? I mean, you have to wonder if pollsters are following the same campaign on some days.

ELVING: Yes. Maybe we should ask whether they're all polling the same country. Each of these polls probably is legitimate and useful in its own way, but you have to say they're describing slightly different parts of the human landscape. So you need to look at multiple polls and know a little bit about how they work if you're going to get the whole picture.

MONTAGNE: Well, break that down for us. Actually starting with multiple polls, you kind of think they are all the same in a way, even though they've got different names and come from different organizations?

ELVING: Yes, that's right, but they do have somewhat different methods, and they are looking for different determiners about who's going to vote. The biggest difference is between polls of registered voters and likely voters. A pollster calls you. He's looking for citizens 18 or over. He gets one. He asks you, are you registered to vote? If you say no, the caller hangs up, he's not interested. You know, you might be a perfectly nice person, but you're not relevant to the task at hand. Then pollsters will ask screening questions to determine how likely you are to vote.

MONTAGNE: And those would be what?

ELVING: Kind of questions such as, have you voted before? How much are you intent on voting? How important is it? How certain are you you're going to vote? But those questions about whether or not you voted before have been the best indicator of voting in the future. And that's where some question marks arise, you know, because people who have voted before tend to have certain characteristics. They tend to be over 30, better educated, better paid, married, and white.

MONTAGNE: Well, likely voters - as they are defined at this moment in time - are they thought to be more accurate as an indicator of the outcome than other voters?

ELVING: In the past they have been. If you look at 2004, polls of all registered voters, the broader group, had John Kerry ahead just a little bit. But the final Gallup poll showed George Bush ahead just a little bit, and that was right. And we know that only roughly half the country votes even in presidential elections, and so the pollsters who go after just the "likelys" are trying to get down to the harder core. It costs a little more, but it can be important for that final degree of accuracy in most past elections.

MONTAGNE: Although, Ron, how much has that changed in this particular election? And let's start with how complicating it might be when you're trying to find a likely voter given the suggestion that there are lots and lots of new voters and young voters.

ELVING: Yes. That's the great question of 2008. Will there be more first-time voters than ever? And will younger, less-affluent people, singles, and people of color vote in comparable proportion to the rest of the population? If they do, then the polls of all registered voters that we've been seeing will be more accurate, and that would mean that the wider Obama leads that we've been seeing in some of the newspaper polls as opposed to some of the tracking polls would be more predictive.

You know, Gallup now has several different numbers that they put out. They have a Gallup traditional, a little bit like Coke Classic. And they also have a Gallup expanded which takes in a few more of the nontraditional voters. And then they have their regular tracking poll of all registered voters, and that shows the widest Obama lead of all.

MONTAGNE: So you're saying Gallup traditional, that would be voters who are the most likely voters. And you're saying the range goes all the way to just more or less registered voters. And in this campaign, one could look at either one and make different choices.

ELVING: And on certain days there has been a great deal of difference between those numbers. That difference however has closed in the last several days, and even Gallup traditional is now showing a widening Obama lead.

MONTAGNE: OK. Well, watch this space, I guess. Thanks, Ron.

ELVING: Thank you, Renee.

MONTAGNE: NPR's Washington editor, Ron Elving. He writes the "Watching Washington" column at, and this week it's on how to decipher the polls.

Copyright © 2008 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

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.