Washington Post Polls

Jon Cohen, The Washington Post's polling director, will be online Monday, Nov. 6, at noon ET to discuss the latest round of pre-election polls.

A transcript follows.

____________________

Jon Cohen: Good afternoon, thanks for joining. So many polls, so little time, let's get right to your questions

_______________________

Richmond, Va.: The polls from Nov. 1-4 are bull! No way that close to half this country still supports the Republicans in this country. Your polling procedures are completely skewed, ergo, these polls are worthless.

Jon Cohen: I'll save an explanation of our polling methodology for another time, but one of the most remarkable things about the Democratic advantage that only recently narrowed was its stability in a country that's been, and in many ways remains, evenly divided politically.

Not only were the last two presidential elections nail-bitters, but also when we add up all the votes for House candidates across the country in each of the past five elections, there's little gap between the parties. For example, in 2004, Republican candidates got 49 percent of all votes cast, Democrats 47 percent. Neither party has had a double-digit win on the national House vote in 20 years.

And there's plenty of evidence in this new poll to show how divided the nation remains today. Take these four questions:

Which political party, the (Democrats) or the (Republicans), do you trust to do a better job handling the situation in Iraq? Democrats 42, Republicans 42

Which political party, the (Democrats) or the (Republicans), do you think better represents your own personal values? Democrats 48, Republicans 44

Whether or not you agree with them, do you think the Democrats are or are not offering the country a clear direction that's different from the Republicans? Yes, they are 49, No they are not 47

In making its case for war with Iraq, do you think the Bush administration (told the American public what it believed to be true), or (intentionally misled the American public)? Told the American public what it believed to be true 51; Intentionally misled the public 47

_______________________

Minneapolis: I believe that your poll and few other polls that were released yesterday reflected last week's Kerry effect at its height, when many for a day or two, having just heard his short sound bite out of context, and Bush's sharp attack, were convinced that Kerry insulted the troops. Could these polls have magnified this Kerry effect which we now know to have evaporated as an influence on the election?

The new CNN poll, which was released today and included sampling between Nov. 3-5, shows the Democrats back ahead. This is in stark contrast to the polls that sampled between Nov 1-3 including the Washington Post poll, which showed the Democratic lead significantly narrowing. Some of these Sunday polls showed surprising Republican gains in Rhode Island, Montana and Tennessee. But the USA Today poll, which sampled between Nov 2-5, has Tester over Burns by 9 percent, not dead even, as reported in the earlier poll.

I would appreciate your comments?

Jon Cohen: The Pew poll released yesterday does show that voters are tuned into the senator's 'botched joke,' six in 10 had heard a lot about it, but most said it wouldn't be a serious consideration for them.

That said, in a close election everything matters, and 18 percent of independents, the key swing voters, said Kerry's comments raised serious doubts about voting for Democratic candidates this year. Most likely, many of those independents who have serious doubts now weren't likely to vote Democratic regardless of Kerry's comments, so the impact is less certain.

A more direct gauge of a minimal 'Kerry effect' may be that there's no change in Post-ABC polling over the past two weeks on how military veterans plan to vote tomorrow.

_______________________

Washington, D.C.: So, am I really to believe these new polls that show Republicans managed to gain 10 points in the generic polls and have closed the gaps in so many races? Call me skeptical.

Jon Cohen: To me these polls show an expected narrowing. (Note there are now a few others showing bigger Dem advantages.) We're a divided country as evidenced by some of the data above. Also, we've seen some easing in the national mood and the president's approval rating, which have both been big drags on Republicans candidates nationally.

_______________________

Arlington, Va.: I never answer the phone unless I recognize who is calling me. Caller ID is one of man's great inventions. I suspect many, if not most, other people are like me. And then you have the huge cohort of young voters who only use cell phones. So, my question is, how accurate can these telephone polls possibly be? And at what point do traditional polls become totally meaningless?

Jon Cohen: Okay, lots of questions on methodology, understandably.

Technology, including caller ID and cell phones, poses a challenge to survey research, but so far not an insurmountable one. The gold standard in polling has been and remains random digit dialing of residential home phone numbers. Numerous studies have reported that the types of people, like yourself, who aren't likely to answer a telephone poll aren't as a group identifiably different either demographically or in their opinions from those who do. That may change, but to-date there's no clear bias from this.

Cell phones are both more and less of a problem. They're a bigger problem because there's some evidence that those who are 'cell phone only' differ on key demographics are not the same as those who have landlines, for example you mention that they're younger. However, on political questions, the young people who have home phones aren't clearly different from those who do. Also, it's possible, although more expensive to poll on cell phones. For example, a recent ABC News poll in New Orleans polled residents on both landlines and cell phones in the cases respondents didn't have home phones.

_______________________

Baltimore: There's a common perception that pollsters "blew" the 2004 elections, and I'm not confident that anyone has a very strong grasp on what's really going to happen tomorrow.

Is it true that pollsters are having a harder time predicting electoral outcomes and -- if so -- when (and who) was the last poll that very accurately predicted an election's results?

Jon Cohen: The final Post-ABC poll before the 04 election had Bush 49-Kerry 48; that was within sampling error of the final result, Bush 51-48.

There certainly was, and remains, some skepticism about the 2004 Exit Poll. That's based more on leaked, misinterpreted preliminary data than poor methods.

_______________________

JEP, Lindsborg, Kan.: John,

Can polls be manipulated by reframing the questions to better match the desired desired?

Jon Cohen: How questions are framed absolutely can alter some answers. That's why the Post publishes our complete questionnaires, so you can see how we asked each question and in what sequence. When I'm dubious about a poll result in another survey, I always turn to their questionnaire first.

_______________________

Montclair, N.J.: Dear Mr. Cohen:

I just logged on -- Sunday at 5:29 p.m. -- and I see that the Generic House Vote has narrowed among likely voters to 6 points. Pew puts it at 4 point. This is an incredible and fast narrowing from just two weeks ago. Are you embarrassed that the Washington Post didn't even see fit to make this a news item on its Web page on Sunday afternoon/evening, no less the headline it deserves? This is an incredible bias. It's similar to the way the Post and New York Times downplayed the half percentage point drop in the unemployment rate on Friday -- if it has gone up one-tenth of a point it would have been a headline. What gives? Are you the one responsible for this bias?

Jon Cohen: Thanks for your question. Our poll was featured prominently (front page above the fold) in our Sunday paper, and the lead graphic had the six-point margin. The headline in today's washingtonpost.com's PoliticsDaily (everyone should subscribe to this!) is "Polls Show Momentum for GOP."

On economic news, last week I wrote a piece about consumer confidence being at its highest level since early 2004. Our poll two weeks ago, showed ratings of the national economy at their best point since Bush took office.

How this translates into politics and tomorrow's election is a different matter.

_______________________

St. Louis Which House and Senate races will be bellwethers?

Jon Cohen: For control of the Senate, a key race to watch will be the Virginia senate race between George Allen and Jim Webb. Just one of the reasons: polls close at 7 p.m. in Virginia, so it'll be an early indicator about whether the Democrats can win the seats they need.

In the battle for the House, the competitive contests in Connecticut (CCC) will be an early sign. Polls close there at 8 p.m.

_______________________

Melbourne, Australia: Given the shocking revelations about Rev. Ted Haggard's behavior and his recent dismissal from the evangelical movement, do you see any impact in the polls likely to suggest that "Conservative Christian" turnout will diminish on Nov. 7?

Additionally, if the above mentioned scenario does eventuate, what individual House races would most likely be affected?

Jon Cohen: Great question from down under. What we've seen all year is somewhat lower support among white evangelical for Republican candidates compared to 2004. While this group remains the strongly Republican, in 2004 78 percent voted for Bush and 74 percent voted for Republican candidates for the House, but now that's 67 percent of likely voters.

Of course, lower turnout among this group would still tend to hurt Republican candidates. While we don't estimate their turnout, we do see that their level of interest in the campaign has increased substantially over the past two weeks.

_______________________

Pittsburgh: "And there's plenty of evidence in this new poll to show how divided the nation remains today."

If the nation is that evenly divided, why so much ebb and flow with the polling numbers?

Jon Cohen: the variation we see in this poll is on the "horse race" question; there's less change on the core questions such as Bush approval, support for the decision to go into Iraq, etc.

In many ways, how people sum up the range of their concerns (national and local alike) changes as they get closer to actually deciding on head-to-head political races.

_______________________

Dana Point, Calif.: What effect, if any, will the Saddam verdict and sentence have on the elections Tuesday? Given that evidence in the trial ended July 27, that sentencing was originally set for Oct. 16, and that on Oct. 29 the Chief Prosecutor Jaafar al-Mousaawi said verdicts needed to be delayed two weeks to allow for judicial "checks," does the timing of the sentence strike anyone as suspicious?

Jon Cohen: A majority has disapproved of the decision to go into Iraq since the fall of 2004. One reason for the, now narrowed, Democratic advantage heading into tomorrow's election is that sentiment about the war has been so negative for so long. Attitudes on the war, which 31 percent of likely voters call the single most important issue in the election, are unlikely to change significantly in the closing days.

_______________________

Jon Cohen: Thanks much for joining the conversation today, I enjoyed your questions and apologize for not getting to all of them. Enjoy tomorrow, if you can. And don't forget to check back on washingtonpost.com for full election results and coverage.

_______________________

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