Will The Polls Change In The Final Two Weeks?

  • Playlist
  • Download
  • Embed
    Embed <iframe src="http://www.npr.org/player/embed/95949903/95949874" width="100%" height="290" frameborder="0" scrolling="no">
  • Transcript

Current presidential polls show Barack Obama with a consistent lead over John McCain. How much volatility can we expect between now and Nov. 4?


This is All Things Considered from NPR News. I'm Melissa Block. Two weeks out from Election Day, and many of us are turning to polling Web sites to keep tabs on where the campaigns stand. Is Barack Obama's lead in the polls solid, or does John McCain show momentum? Are those blue and red lines going to converge as we head toward Election Day? And should we trust in the polls at all? Well, Mark Blumenthal is here to sort through some polling trends. He's a former Democratic pollster, now the editor and publisher of Pollster.com. It's a nonpartisan Web site that compiles poll results. Mark, thanks for coming in.

Mr. MARK BLUMENTHAL (Editor & Publisher, Pollster.com): Thanks for having me.

BLOCK: And at this point, your compilation right now has Barack Obama up over John McCain by 5.5 percentage points. What does the overall trend line show in this race at the end of the campaign?

Mr. BLUMENTHAL: Well, we've had really quite of a rollercoaster ride since the conventions. The McCain-Palin ticket narrowed the margin, moved slightly ahead right after the convention. And then during a very tumultuous September, we had the Obama campaign come back strong and at one point had about an eight-point advantage on our national trend chart. It's narrowed just a little bit over the last week. But as of right now, things look pretty good for the Democrats.

BLOCK: If you look at a candidate like Barack Obama, who's up, say, five, six percentage points now, is that determinative of a final outcome? Or have you - can you find plenty of cases in history where a candidate was up five or six points two weeks before Election Day, and lo and behold, they lost?

Mr. BLUMENTHAL: I think I've learned over the years that you never say never about pre-election polls. There are examples, 1996, where Bob Dole closed about five points of Bill Clinton's margin. There was 1980 where there was a very late reversal. Reagan had been leading by a little bit, ended up winning by a lot. And polls that were still in the field caught that over the final weekend.

I think the thing to try to do in looking at our situation right now is ask ourselves, you know, voters have been exposed to a lot of information now, three presidential debates, the conventions. They actually know quite a bit about these candidates. So you don't really have the conditions of 1980 where all that happened late. It's not impossible that the margins may close and that McCain could come back. It's looking pretty unlikely.

BLOCK: Do you think polling methodology has gotten any better? Are polls more reliable now than they used to be?

Mr. BLUMENTHAL: Strictly speaking, if you look at the sort of error rate in the last national poll as compared to the result on Election Day, they're about the same, and they've been pretty good since the mid '70s. They actually got a lot better from the '40s up until about 30 years ago. What's changed is that the conditions for doing polling have gotten much more challenging. It's harder to get respondents to stay on the phone. We have the cellphone problem. We have all sorts of other challenges thrown our way now. As of yet, they've not thrown off telephone poll accuracy. But then we get our final exam every two years, and we'll see what it looks like in two weeks.

BLOCK: Cellphone problem meaning voters who no longer have a landline. They have a cellphone and may not be getting polled at all.

Mr. BLUMENTHAL: Right. Pollster started using telephone sampling, started doing surveys by telephone, in the 1970s when the percentage of Americans who had a landline phone rose to, you know, over 90 percent. And it was there for a long time until about four years ago when a lot of Americans, particularly younger Americans, started disconnecting their landline service and using their cellphone exclusively. There are studies that show that 15 percent of adults live in households that can only be reached by a cellphone. And as that, sort of, uncovered portion gets bigger, the potential for pollsters to get something wrong is bigger.

BLOCK: Mark, as somebody who's been enmeshed in these polls for a long time, how prepared would you be on November 5 to realize that the polls got it way wrong one way or the other?

Mr. BLUMENTHAL: Well, we all should have a pretty recent memory of when that happened. It was just 10 months ago in New Hampshire where nine or ten public polls all had Barack Obama ahead by pretty healthy margins. And then on primary night, Hillary Clinton was the winner. I think the one sort of nightmare scenario for pollsters, and it may have played some role in New Hampshire, is the possibility that the people we can't reach or can't talk to might have a systematically different preference in the election than the ones we're able to interview. If the response rate ends up causing response bias, then it might be possible. And since we don't know for sure, because we don't know much about the people we can't interview, we should always be prepared for that possibility.

BLOCK: OK. Mark Blumenthal, editor and publisher of Pollster.com, thanks very much.

Mr. BLUMENTHAL: Thank you.

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

NPR transcripts are created on a rush deadline by a contractor for NPR, and accuracy and availability may vary. This text may not be in its final form and may be updated or revised in the future. Please be aware that the authoritative record of NPR’s programming is the audio.



Please keep your community civil. All comments must follow the NPR.org Community rules and terms of use, and will be moderated prior to posting. NPR reserves the right to use the comments we receive, in whole or in part, and to use the commenter's name and location, in any medium. See also the Terms of Use, Privacy Policy and Community FAQ.