« WaPo/Univision/TRPI Poll: Kerry Ahead Nearly 2-1 Among Latino RV's | Main | Quinnipiac Poll: Kerry, Bush tied Among FL RV's »

Analysis of Major National Polls

By Alan Abramowitz

Here's where the presidential race stands right now based on an analysis of the most recent national polls conducted by 10 leading polling organizations. For this analysis I have excluded all partisan polls, internet polls, and robo-dial polls. That leaves out Democracy Corps, Harris, Economist/YouGov, and Rasmussen. Some of these polls, especially Democracy Corps, are in my opinion very reputable, but for the sake of fairness I'm excluding them. Included in the analysis are the following polls: Gallup, CBS/New York Times, NBC/Wall Street Journal, ABC/Washington Post, Zogby, Time, Newsweek, Pew, AP/Ipsos, and LA Times. All of these polls except Zogby report results for registered as well as likely voters.

Among likely voters, Bush was leading in 5 polls, Kerry in 2, and 3 were tied. The average level of support for the candidates was Bush 48.2, Kerry 47.0, Nader 1.3.

Among registered voters, Bush was leading in 3 polls, Kerry in 2, and 4 were tied. The average level of support for the candidates was Bush 47.0, Kerry 46.0, Nader 1.9.

It is clear from these results that heading into the final weekend of the campaign, the presidential race right now is extremely close. George Bush appears to hold a very slight lead nationally, but his support remains below the 50 percent level that is generally considered necessary for an incumbent since undecided voters generally break toward the challenger by a wide margin.

Comments

Alan:

I understand leaving out the more "partisan" polls. Unfortunately you kept Gallup in ..who I believe you criticized for being heavy on the Rep's in sampling!! and you eliminated Harris...?? why??

Addtionally, Alan, can you give us the dates these polls were conducted, whether they are all RV's or Lv's, or LV's with leaners.

As you know, a poll is only as good as the time it was taken. I also question the accuracy of "averaging polls" ?

Other wise , Alan ,you do good analytical work on here... keep it up!!

Thanks

Ted Sheedy

i understand that the undecideds break 2-1 for the challenger. But even if that is true this time, how can Kerry win with that. For example, lets say the latest polls average out to Bush 47, Kerry 46. That leaves 7% undecideds. That means Kerry gets 2/3 of 7, or 4.666 and bush gets 2.33. That means that kerry is at 50.6 and bush is at 49.333. Is that enough? That is with assumptions. If these numbers are off at all (say bush48 to kerry 46) this changes. And it doesnt compute in the Nader or Badnarik votes.

Can someone please clarify how the undecideds can save Kerry (Lord knows i want him to win)

PS> i know all of this is based on these piece-of-crap inaccurate polls...

I'm starting to worry that we are setting up here for a major disappointment here regarding the undecided rule. What is the empirical evidence that undecided voters break heavily for the challenger in presidential elections? This link is to a page that purports to show (in quite some detail) what has happened in past presidential elections, and it isn't good news for Kerry.

http://myelectionanalysis.blogspot.com/2004/09/whither-undecideds.html

I hate to say it but I don't think we can assume that Kerry is going to get undecided voters by 2:1 on election day.

I'd love to see the same analysis on the swing state polls.

Fairness? Fairness? I don't know what meaning that word might have in the present context.

Your analyses should strive above all for *accuracy*. Fairness is a thing of the distant past.

Hey Ruy... I don't think it's "fair" to rely on the right-wing money people represented in such polls as Gallup, CBS/New York Times, NBC/Wall Street Journal, ABC/Washington Post, Zogby, Time, Newsweek.

They have been proved very biased towards Republican candidates and so I think you play into "their" hands.

What do you think of Mickey Kaus's analysis concluding undecideds can't be counted on to break for the challenger?

http://myelectionanalysis.blogspot.com/2004/09/whither-undecideds.html

Interesting that you leave out YouGov.

Here in the UK, the internet polling by this company is reckoned to be the most reliable of all methods.