Election 2008: The Latest Polling

Jon Cohen and Jennifer Agiesta
Washington Post Polling Department
Wednesday, May 14, 2008; 2:00 PM

Washington Post polling director Jon Cohen and polling analyst Jennifer Agiesta were online Wednesday, May 14 at 2 p.m. ET to discuss the findings the latest Washington-Post ABC poll -- including some early general election polling -- and any other questions you might have.

U.S. Outlook Is Worst Since '92, Poll Finds | Burdened by the Weight of Inflation | Data (Post, May 13-14)

The transcript follows.


Jennifer Agiesta: Hi everyone -- welcome to the chat. We should have links to coverage of our most recent poll posted shortly. Lots to cover, so let's get started.


Anonymous: Any chance of an Obama-Clinton ticket now? And do you think whoever wins will accept the pseudo-apology of McCain supporter Rev. Hagee (saying he regrets any remarks Catholics may have found hurtful, not that he acknowledges what he said was wrong)?

Jennifer Agiesta: Thanks for the question. As of our latest poll, Clinton is the favorite choice of Democrats to be Obama's running mate, should he win the nomination. About four in 10 named her in an open-ended question, but about as many either had no opinion or said the choice was up to Obama. Among those who'd like to see Obama win, 32 percent said he should choose Clinton; it was 47 percent among Clinton supporters.

Other mentions: 10 percent named previous vice presidential nominee John Edwards, 3 percent Bill Richardson and 2 percent Al Gore. But at this early stage, much of the response to this question is name recognition.

Another measure found that more than half of all adults said having Clinton on the ticket would make a difference in their vote -- a quarter said they'd be more likely to consider a Democratic ticket including Clinton, and 18 percent said they'd be more apt to vote Republican if Obama and Clinton lead the Democratic line.


johnkwhite1: When more than 80 percent of the nation's voters think the nation is seriously off-track, I keep wondering how it is that not one presidential candidate has said what needs to be said: We need to regulate in the same way we regulate utilities those corporations engaged in providing necessary commodities, e.g. gas and oil, and we need to ensure balanced trade with all nations, e.g. Communist China, and we need to tax the living hell out of American corporations, e.g. drug manufacturers, that choose to offshore production and manufacturing.

Yeah, things would be wild for a while until the American-based production and manufacturing was back on track, but after that, America would be strong again. All that needs to happen is getting rid of the corporate lickspittle politicians in Washington. ... Right now, they all are corporate lickspittles -- Democrat and Republican.

Jon Cohen: You're certainly right about the context. In our new poll, 82 percent of Americans said the country is pretty seriously off on the wrong track, and that was true of majorities of Democrats, Republicans and independents. The percentage saying we're going in the wrong direction is now within a point of its all time high in Post-ABC polls.

But the reasons for pessimistic outlook are varied (the downturn in economy and the unpopular war being two primary causes), so you're unlikely to see the kind of major overhaul of the role of government you envision. Nevertheless, trade policy has popped up as an issue in the Democratic primary, and it may play a key role in some swing states this fall.


New York: Chris Cillizza wrote on his blog that McCain was winning white women by 50 percent 43 percent over Obama, while Obama was winning women overall 54 percent to 40 percent. It seems impossible, given the composition of the electorate, for Obama to be winning 54 percent of women, but only 43 percent of white women, who must make up somewhere in the neighborhood of 80 percent of the female electorate. Is my math wrong, or is Cillizza's? The Fix: McCain vs Obama: By the Numbers (, May 13)

Jennifer Agiesta: Hi New York. Obama holds a lead among women (54 to 40, as reported by The Fix) because of his overwhelming lead among black women. White women make up about three-quarters of all women in this poll, and they do break 50 percent for McCain and 43 percent for Obama.

Among African American women, Obama leads McCain 94 percent to 4 percent; among all nonwhite women, he leads 88 percent to 9 percent.


Bellingham, Wash.: I've asked this of a number of chatters at The Washington Post and haven't gotten a response so I hope you will take a crack. Here goes: Clinton claims she is more electable come November. One could argue, however, that one of the few factors that has kept her alive in the Democratic Primary is the skew toward females in the Democratic electorate -- it has skewed 55 to 60 percent female in most states.

Even given this demographic advantage, Clinton finds herself behind in states won, votes received and delegates allotted (both super and non-super). So how can Clinton argue (and should superdelegates take seriously) that she is more electable, given that when we get to November the gender split will break a lot closer to 51-49 women rather than the 58-42 seen during the primaries? Seems that all things being equal, Clinton will get her lunch handed to her in the general, given a more even gender split. Thanks for addressing my question...

Jon Cohen: Hopefully will have time to return to the implications, but for starters here are the numbers from 2004 and 2000: According to network exit polls, 54 percent of the national electorate was female in 2004; in 2000, women made up 53 percent of all voters. Fifty-three percent of Ohio voters in '04 were women, as were 54 percent of those casting ballots in Florida in '00.


Seattle: Why should we believe polling data now as a view to the future? Given the presidential campaigns' record with polling accuracy, aren't polls becoming outdated?

Jon Cohen: I'd suggest that you not see polls as predictive -- they're designed to measure opinions as they stand today, not magically presage attitudes at some future point in time.

I'd also submit that high-quality polls conducted toward the end of campaigns are generally on-the-mark, with some obvious recent exceptions.


Washington: I just saw something on TV that said that the numbers for Clinton vs. McCain were at 56 percent to 40 percent, and for Obama vs. McCain they were at 48 percent to 41 percent. First, how has Clinton's lead over McCain improved so much? Second, what happened to the 11 percent in the Obama/McCain matchup? Why are people willing to commit to Hillary but not Barack?

Jennifer Agiesta: Hi Washington, thanks for the question. There are a staggering number of polls out there at the moment, of varying degrees of quality and with varying results. Looking at national results, in our latest poll and in a number of reputable recent polls, Clinton and Obama fare similarly against McCain.

The Post-ABC poll released this week had Obama leading McCain 51 to 44 percent, while Clinton had 49 percent to McCain's 46. Since March, we've found Obama consistently ahead of McCain, while the Clinton matchup has varied somewhat -- and with six months before Election Day, those variations will continue to happen as the ongoing campaign impacts public opinion on the candidates.

Both candidates, however, would bring a strong base of supporters to the general election: 96 percent of Democrats who said they'd prefer to see Obama win the nomination say they'll vote for him over McCain, and 94 percent of Clinton's backers would choose her in November.


Arlington, Va.: Please explain to this polling novice why your sample population regularly has 10 percent-15 percent more respondents who lean Democratic than Republican (50 percent to 29 percent in the most recent poll) when the past two presidential elections were just about evenly split? Why does this imbalance in your sample not affect your polling results?

Jon Cohen: A good question, and one that pops up almost every time we release a poll. We and all other polling organizations that release data on "party identification" show a Democratic advantage. This varies from poll to poll (in our poll this week, 34 percent of respondents considered themselves Democratic, 28 percent Republican and 34 percent independent).

You're highlighting the "leaned party identification" numbers we post along with each survey. That's the result after we ask those who don't generally think of themselves as partisans whether they "lean" more toward the Democrats or the GOP.

We wrote a bit more about the underlying question as part of our major project on political independents, conducted with Kaiser and the Harvard School of Public Health; here's the link.

There's a ton more to write on party identification. I'll post some other links soon.


Accountability: One problem I have with pundits and polling is that they lack accountability. If a poll is well-done but ends up being 100 percent off-the-mark, is anyone held accountable (fired, blacklisted, etc.)?

Jon Cohen: Thanks for the question. First, I put pundits and pollsters in different camps (but then again I would). "Well-done" polls, as you call them, are the single best available way to gauge public opinion. But without a discourse on why it's important to have a good read on what people think, I will say that we do take a pollster's track record into account when we consider writing about new findings. Ultimately, though, the single best way to judge polls is to look at how they were conducted, not their results.


sgw99: The media hammers a propaganda line (recession, inflation, economic collapse) and then does a poll that shows people think just what the media has told them. Ironic that this Washington Post story is printed the day after economic data is released that shows inflation remains very low. "Journalistic ethics" at work. Burdened by the Weight of Inflation (Post, May 14)

Jennifer Agiesta: Regardless of what the media say, people know their own pocketbooks, and our questions focused on the personal impact of recent price increases. Prices for food and fuel have climbed dramatically (in fact, the government today reported the highest monthly spike in food prices in nearly 20 years, and average gas prices have risen 11 cents in just the past week), and our results reflect those changes.

As one poll respondent told me yesterday: "I'm very concerned. It's just getting harder and harder to make ends meet, and it's the stuff you have to have, and it's getting to the point where we're having to make choices."


darrynfoley: Let's draw the only accurate conclusion possible: 1,000 adults out of about 300 million U.S. citizens believe this. I'm so sick of these polls that sample a mere 0.000001 percent of U.S. citizens being extrapolated into "what America thinks." The only accurate conclusion that can be drawn is that "out of 1,000 people surveyed, X percent of them believe Y." Please stop the faulty logic of extrapolating this upon America at large. The real poll will come in November, and that's the only one that matters.

Jon Cohen: If you don't believe in random samples, next time you go to the doctor's office and she asks for a vile of blood, say "no, no -- take it all."

Pollsters have few jokes, but that's one...

Speaking of which, this weekend is the annual conference of the American Association for Public Opinion Research (AAPOR); I'll let you know next time if there are any new, good jokes.


New York: Why do the press titles only state that Clinton won in West Virginia when at the same time Obama won in Nebraska? The Nebraska delegates virtually cancel out Clinton's delegates in West Virginia. Why isn't there more coverage on that?

Jon Cohen: Last night's Democratic voting in Nebraska was a "beauty contest"; the state's delegates were allocated according to the results of a Democratic caucus held Feb. 9. Obama won 68 percent of caucusgoers. Last night he won the symbolic primary 49 to 47 percent.


Washington: Although the elections are not for another five months, how are things trending, and what would your early predictions be for the White House winner?

Jon Cohen: Going nowhere near a predication, but the "atmospherics" are clearly Democratic-leaning now. In our poll this week, Democrats have a 21-point advantage as the party better-suited to cope with the country's main problems. That ties their biggest lead on that question in Post-ABC polls going back to 1982. The only other time the gap was so large in the Democrats' favor was in mid-January 1993, on the eve of the Clinton inaugural.


Chicago: Mentioned by many readers (and some writers) of the news after Sen. Clinton's West Virginia win was that hers may be more a coalition of Appalachian geography than demographics. What are your thoughts from the poll? How does this affect the race in November, and is there a vice presidential choice for Obama that would shore up some of this?

Jennifer Agiesta: It's true that Clinton has had broad advantages in Appalachian regions, and in general, geography and demographics go hand in hand. The demographic fault lines in the primary campaign have been well established, but people who vote in primaries are very different from those who vote in general elections, and there is little to suggest that these divides necessarily will remain the same in November.

In our poll, Clinton and Obama each trail McCain by nine points among white voters, and by similar numbers among white non-college voters, and both do extremely well among African American voters and win among women. But some differences remain -- Obama outperforms Clinton among young voters, while she bests him among seniors. Obama holds a greater advantage in the Midwest, but Clinton does a little better in the South.

So the contours of the general election contest still are being shaped, and may not be the same as those we've seen in the primaries.


Fairfax County, Va.: I was angry after North Carolina and Indiana that the exit polls did not ask about the gas tax holiday, which was the really live issue in the final few days and was a fascinating exercise in two totally different approaches to the voters. Instead, they asked once again about Rev. Wright, who by then had fallen off the radar screen (somewhat). No problems with asking about both, but it's annoying now to have to infer whether or not Obama's rather daring truth-telling strategy worked, or had no effect, or even lost him Indiana (which I doubt). Who decides things like this? If only one of the two topics was polled, it should have been the gas tax holiday.

Jon Cohen: There was a gas tax question on last night's West Virginia exit poll, but you're right it would have been nice to have a week earlier.

We subscribe to the exit poll, but we have no input into its questions. -- so while I have no information about the decisions of what questions to put on each exit poll, it's possible that the gas tax issue simply popped up too late to adjust the questionnaires (the exit poll is conducted on paper, and questionnaires have to be printed and mailed to exit pollsters in advance).


Jennifer Agiesta: Thanks everyone for your great questions today, and be sure to look for further poll analysis on Behind the Numbers. Thanks!


Editor's Note: moderators retain editorial control over Discussions and choose the most relevant questions for guests and hosts; guests and hosts can decline to answer questions. is not responsible for any content posted by third parties.

© 2008 The Washington Post Company