As Election Day Nears, Poll Data Questioned Some recent polls show Sen. Barack Obama with a double digit lead in the race for the White House, while others show a much narrower gap between him and his opponent Sen. John McCain. Many wonder whether poll data is truly an accurate portrayal of voter sentiment. Celinda Lake, a veteran Democratic with Lake Research Partners, and Pamela Gentry, a political analyst with BET Networks, discuss the value of poll data among voters.

As Election Day Nears, Poll Data Questioned

As Election Day Nears, Poll Data Questioned

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

Some recent polls show Sen. Barack Obama with a double digit lead in the race for the White House, while others show a much narrower gap between him and his opponent Sen. John McCain. Many wonder whether poll data is truly an accurate portrayal of voter sentiment. Celinda Lake, a veteran Democratic with Lake Research Partners, and Pamela Gentry, a political analyst with BET Networks, discuss the value of poll data among voters.


I'm Michel Martin, and this is Tell Me More from NPR News. Coming up in our international briefing, we get an update on the fighting in the Congo. The latest on the power-sharing deal in Zimbabwe and stories from the frontlines of the AIDS epidemic in India.

But first, back in the U.S., following the poll numbers, hard to believe Election Day is just five days away. On November 4th we'll find out if our next president is John McCain or Barack Obama. We'll also learn which of the dozens of public-opinion polls were right and which ones were way off base. But why is there so much variation to begin with? Why are some surveys telling us the election's all but over and others that it's a squeaker going down to the wire?

To find out we're going to turn to Celinda Lake. She's a veteran Democratic pollster. She's the president of Lake Research Partners. She's not affiliated with either of the presidential campaigns this year. Also joining us is Pamela Gentry. She's the Washington bureau chief and senior political producer for BET Networks. They're both here with me. Thank you both so much for joining us.

Ms. CELINDA LAKE (Democratic Pollster;President, Lake Research Partners): Thanks for having us.

Ms. PAMELA GENTRY (Washington Bureau Chief and Senior Political Producer, BET Networks): It's a pleasure.

MARTIN: Celinda, just looking at the recent polls, some show Obama with a double-digit lead, others say just three percentage points separating the candidates. How do you account for this disparity?

(Soundbite of laughter)

Ms. LAKE: It is hard, and I think it must be really hard on observers now. One of the biggest differences is no one knows who's actually going to vote. And the turnout models are very, very different in all these polls. So the ones with the biggest margin for Obama tend to be ones that have the most new models, that think there will be the most new voters. And the ones that have the nearest margin are the ones that are the most conservative in their turnout models. They tend to think of traditional turnout patterns among people with color, traditional patterns among young people. Frankly, I think, you know, we want to see Obama ahead in all of them, that's the key, which he is right now. So I think he's clearly going to win. But it means that there is no vote to be taken for granted, that turnout really will make a difference in this election.

MARTIN: Pamela, you've been writing about this. Do you think that there are people or voting blacks that are routinely missed by the pollsters, and how do you think this affects the data?

Ms. GENTRY: Well, I think she just answered that in the sense that, you know, you all have these newly-registered voters. They have no patterns so you don't know what they're going to do. And likely voters has always been the base polling, so that means that they have shown up somewhere before. So if you have new voters, they're not there. And if you have these young voters that are 18 to 35, they haven't gotten a phone call and they're not on anyone's list because they never had a landline telephone. They're all cell phone - they're the cell-phone generation. So the chances of them ever being asked in any one of these polls, I'm sure will be corrected in the next four years.

MARTIN: How do you think - do you think that some of the confusion comes from the polling or is it the way the journalists - we as journalists write about and report on the polls?

Ms. GENTRY: I think it's probably a combination, but I think the pollsters have a good system, and so their system we've relied on for a very long time. What's confusing to us, I know for me in particular, is how to report these polls because if you go to a national poll, you see a two or three percent margin, and then you go state by state, it gets up to seven to twelve percent. And I also think another variable that we haven't thought about is the early voting. If you're polling a likely voter that's already voted, who knows what they're going to tell you when you call them three or four days in? And so these campaigns must be pulling their hair out in states that are swing states with early voting for two weeks out, and they're still trying to poll.

MARTIN: Celinda, I should mention that you team up with a Republican pollster, Ed Goeas, for some of your national polls. Why do you do that?

Ms. LAKE: We do that because we think it's a great - first of all, it's a great partnership and a great way to learn from each other. And there's a lot of bipartisan polling out there. What's different about the GW battleground poll that we do with Ed is that we look at the same data but we don't write joint analysis. We each write separate analysis, and it's been a great way to see what's the different perspective of the two parties.

And actually, right now, one of the biggest differences in perspective between the two parties is the turnout model. We've been reporting from the battleground survey the Republican turnout model, but frankly, our internal work and our internal model would have the Democratic vote - would have the margin more like five, six points, seven points because we believe more new voters, more people of color turning out to vote.

By the way, pollsters have developed one technique - and you're right, the early vote is hard to figure out. And so we ask people, have you already voted or let them volunteer because we were getting such high refusals. And now, on Election Day, in many states over half the voters will have already voted.

MARTIN: When you say high refusals, what do you mean?

Ms. LAKE: Well, when you ask people, how are you going to vote? People didn't know where to put themselves because they already had, and so now we let people volunteer. Did you already vote, and then we say, OK, this is just for research purposes. Because interestingly, when people haven't voted yet, they're very willing to tell you how they're going to vote. Once they have voted, they tend to say, wait a minute, this is the secret ballot box and I'm not going to tell you how I actually voted.

MARTIN: That's so interesting. I didn't know that.

Ms. LAKE: So we have to say for research purposes just tell us how you voted.

MARTIN: That is really interesting. Pamela, the public complains - it's interesting, you know, we're having this conversation, and then I can tell you exactly what some of the email is that we're going to get as soon as we finish, which is that you people in the media spend so much time talking about the polls, you spend too much time talking about the polls. Do you think that's a fair criticism? Do people say that to you and what do you think of that?

Ms. GENTRY: Well, they - the criticism I get is that I don't believe the polls, you know. I mean, that's the first thing out of someone's mouth is that. Do you really believe those polls are accurate? And then, you know, I try and do the explanation of why we look at polls and why polls are important. I said, is it because a lot of things - maybe it's not just about how you vote, but you like polls. I said, you want to know what the most important issue is and you want candidates to address those things, so polls are good. But I think a lot of times, just people don't believe it because they say, well, I've never been polled. It's sort of like the Nielsen ratings, you know, how do they know what I watched? So they don't believe it.

MARTIN: Do the people who say that tend to be people who are on the winning side or the losing side?

Ms. GENTRY: I hear it from both sides, but again, I deal with a large African-American community when I'm talking to most of my - the folks that respond to me and who write on my blog, and they don't feel like they've ever been polled. So their perception of their accuracy is a little bit suspect.

MARTIN: If you're just joining us, you're listening to Tell Me More from NPR News. We are talking about the presidential polls. How accurate are they? Why is there so much variation in those polls? And we're speaking with political analyst Pamela Gentry from BET and pollster Celinda Lake, who is a Democrat but who routinely teams up with a Republican pollster for her current sort of national polling.

Celinda, do you and Ed ever disagree, your Republican partner, about how to ask a question? I mean, one of the things that people say to us is, you know, the way you ask a question is going to dictate the answer. Some people feel that way. So do you ever disagree about that and do you think that's true?

Ms. LAKE: We do actually, and that's part of the learning, too. And in fact, what we do is we each write our own questionnaire. Unfortunately, we would love it if we could just ask separate questionnaires and then come together, but obviously we can't. So then we bring our separate questionnaires together and compare notes. And frankly, the wording of the questionnaire, the order of the wording of the questionnaire makes a big difference.

If I could address your last question, too. Frankly, a lot of pollsters, myself included, think the media spends too much time on polling, and we think that the media spends too much time - not NPR. You guys are an exception. But on the least important aspect of the polling. For us, the horse race question is the least important one because it's the one that's most likely to change. What we're much more interested in are the underlying patterns, the underlying attitudes. How are people feeling about the economy? Are they connecting it to their vote? How do people feel about change? Is change risk or is change good? Those are the far more interesting questions, and yet the media just tends to report the horse race. So I think the much more thoughtful analysis that you see on NPR, that you see on BET is great because pollsters think that there is too much time spent on the polls.

MARTIN: What about - what are some of the underlying trends that you think are most interesting, that you're gleaning from the work that you've done in the cycle?

Ms. LAKE: Well, there are a couple that are really fascinating. First of all, we have young voters turning out in record numbers, and I think there's no question that there will be record turnout. We have, for the first time, undecided voters older. Usually at this point in time, undecided voters are younger. But this year, young voters made up their minds early on, and the astronomical levels of support for Obama made a big difference in the primaries and make a big difference in the general election.

Secondly, we see that men really are from Mars and women really are from Venus, and there's a big gender gap in the data with women solidly voting Democratic, men voting Republican. Women still more undecided. And both parties going after the women, and it will be women who will determine who's president.

MARTIN: And how does race play into this?

Ms. LAKE: Race is huge. First of all, we see African-Americans turning out in record numbers and voting at the highest possible levels for Barack Obama. We also see African-American voters are voting early, which they never do. So for example, in Georgia, there may be a huge upset in the Senate race because African-American voters came out early to vote. We also see the last trend - we really do live in an unmarried America now. Half of American households are unmarried, and as big as the gap is between male and female voters, the gap between married and unmarried voters is even greater. And so you see married voters voting - splitting their vote, married men voting more Republican, unmarried voters, both men and women, voting overwhelmingly Democratic.

MARTIN: That's really interesting. Pamela, finally, I know African-Americans often complain that they are kind of lumped together as a group, that the nuances and differences of opinion among - within the community are not fully explored. So at BET, it is your job to explore all those differences. Given what Celinda said about how - and all of the surveys show African-Americans very solidly behind Barack Obama and the Democratic ticket this year - are there any nuances to explore in the African-American vote that you think will be interesting to talk about after the election?

Ms. LAKE: I think that the best part about the - him being able to capture the vote was that he had to earn the vote because Hillary Clinton really had the African-American vote when this contest started. And he made it an asserted effort to convince African-Americans to come to his team, and that was important simply because 85 percent of African-Americans always vote Democratic. So even the fact that he had to make that effort shows that they were not just one group of people who automatically were just going to follow a candidate. They - originally Hillary Clinton was their candidate, and they made a choice to go towards Barack Obama.

MARTIN: And when did that consolidation happen behind Obama?

Ms. LAKE: I think it was late in the primaries, and I think it was after Iowa because I think prior to Iowa, African-Americans didn't believe that white America would vote for an African-American. And once they saw that white America saw his viability and he was campaigning very well and doing well in states like New Hampshire, I think that they started giving him a second look, and I think that's what made the difference.

MARTIN: OK. What do think is going to make the difference, Celinda,very briefly. You said women are going to decide this election. Which women? How many women and what states?

Ms. LAKE: Well, I think, first of all, the battleground states, and the polls are looking better high in the battleground states than they are nationwide. That's the other misnomer. Secondly, young people. If they turn out in record numbers, they'll elect a president. And a lot of people down ballot. One of the most important things is not just the election of Barack Obama, but to send a whole team of change and make sure you don't just drop off at the presidential level.

MARTIN: Celinda Lake is a Democratic pollster - she made that clear - and president of Lake Research Partners. Also important to mention that she does team up with a Republican pollster in arriving at her data. Pamela Gentry is the Washington bureau chief and senior political producer for BET Networks. They were both kind enough to join us here in our Washington studios. Thank you both so much.

Ms. GENTRY: Thank you.

Ms. LAKE: Thank you.

MARTIN: And we also want to hear from you on Election Day, Tuesday, November 4th. We'd like to know about your trip to the voting booth. We want to know how long you waited in line, whether there was tension between Barack Obama and John McCain's supporters, was it a lovefest, if your experience was smooth sailing or rough waters. Did you vote early? You can call our comment line at 202-842-3522. Once again, that number is 202-842-3522.

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.