« Follow-up on Gallup Poll | Main | Bad Gallup! No Biscuit! »

You Know, I Think We're Finally Getting to 'Em

At least over at Newsweek, where their latest poll has Bush up by 5 points (50-45) in their 2-way LV matchup (50-44 in their 3-way). But here's the headline and lead of their polling release:

Bush/Cheney in Dead Heat With Kerry/Edwards in Two-Way and Three-Way Matchups Among Registered Voters

In a two-way matchup, the presidential race remains in a dead heat in the latest Newsweek Poll. Among registered voters, Bush/Cheney gets 48 percent and Kerry/Edwards 47 percent of the vote....

And here's the headline and lead of the polling story they posted on their website:

Too Close to Call: With the debates behind them, the contenders in the race for the White House remain locked in a dead heat in the latest NEWSWEEK poll

With just 17 days remaining in the race to the White House, President George W. Bush and Sen. John Kerry remain locked in a dead heat, according to the NEWSWEEK poll, taken after Wednesday’s final debate in Arizona. In a three-way race with Independent candidate Ralph Nader, 48 percent of all voters say they would reelect Bush while 46 percent prefer Kerry.

Good job, Newsweek! They don't even get to their LV results until the third paragraph of the above story.

As to the data themselves--leaving aside the LV issue--there are some rather odd things about it. One is that their RV sample--where the race is close to even--shows a strong swing to the Republicans in party ID, compared to Newsweek's last survey, conducted right after the first debate.

While Newsweek rather unhelpfully only provides their (demographically) unweighted party ID distributions for RVs, one can infer from other data they provide that there has been about a 6 point swing toward the Republicans in their demographically weighted party ID distributions between the two polls. That means that, if one party-weighted their current poll to match their previous party ID distribution (which was pretty close to that of the 2000 exits), the horse race results between the two polls would look remarkably similar: a 2-3 point Kerry lead.

Newsweek provides a number of interesting subgroup horse race numbers for their 3-way RV matchup that are worth taking a look at:

1. Independents favor Kerry/Edwards by 11 points, 51-40.

2. First-time voters favor Kerry/Edwards by 21, 57-36.

3. Early voters favor Kerry/Edwards by 9, 52-43.

4. Young voters (18-29) favor Kerry/Edwards by 9, 50-40, and seniors (65+) favor Kerry/Edwards by 15, 54-39.

5. Men favor K/E by 50-46 and women favor Bush-Cheney by 49-43.

Huh? Boy, I had to look over those gender breakdown data several times to make sure I wasn't seeing things. And I'm still not sure they didn't somehow mislabel their categories. But if they didn't, it's certainly a head-scratcher. Perhaps they not only oversampled Republicans in general but Republican women in particular. Who knows.

It's also interesting to note that, even with a Republican-leaning sample, the poll still gives Bush only 47 percent approval rating, basically the same as he received in their last poll. And his re-elect number, at 47 percent, is also about what he received previously.

As for the rest of the poll, it generally shows Kerry and Bush with advantages where you'd expect them to have them. However, in contrast to the Time poll I posted on earlier, Kerry's margins on given issues and attributes are smaller and Bush's larger than they were in Newsweek's previous poll. But it is difficult to assess how much of this is real movement and how much of can be attributed to the effects of a substantially more Republican sample.


I felt the same way yesterday when I read the Newsweek polling, but today, after looking at the new Gallup and Washington Post polls, am wondering if there isn't some real movement back to Bush. Not sure optimism is realistic at this point, just wait and see.

You can make lemonade out of these lemons of polls, or take it tart--the fact is no one really know exactly what is going on, there is a dearth of reporting from the candidates rallies (I think the press prefers to feed on itself rather than go out and actually report), and no one has a clue to how it is going to turn out.


Zogby also has men going for Kerry and women for Bush. My only explanation is that Bush did manage to scare the security Moms, so to speak, that if they want to protect their kids, they have to swallow hard and vote for him. That's where the draft issue may be effective in balancing the fears, since nowadays it would apply to children of both gender and would affect even more Moms.

An update: CNN went the Newsweek's route and their headline is not the screaming one of the USA Today -- as in Bush Surges!!! -- but simply -- Poll: Presidential race still tight:

The latest poll, taken after the third and final debate last Wednesday in Tempe, Arizona, indicated an edge of 49 percent to 46 percent for Bush...

That is still practically even -- given the margin of error of plus or minus 4 percentage points. The October 9-10 poll had a margin of error of 3.5 percentage points.

"1. Independents favor Kerry/Edwards by 11 points, 51-40."

"2. First-time voters favor Kerry/Edwards by 21, 57-36."

Those are some outstanding numbers. If those numbers happen on election day, we're obviously in great shape.

Kudos to DR on Newsweek leading with their RV numbers. I'm not saying he's due 100% of the credit, but there can't be much doubt that his yeoman's work is paying off here.

Still gotta work on Gallup...grrrr.

I just don't get Newsweek. As a subscriber they definitely tend towards more progressive content yet their polls go the other way.

I know from doing the lawn sign poll system that in Southern MN the towns are definitely leaning Kerry over Bush Cheney and in the Country its the opposite. I think its pretty clear that this illustrates a Retro Vs Metro split.

Look at the cnn release also. They are pushing the same headline as Newsweek. Ignore our likely voter sample and look at our registered voter sample. They must be realizing there is a ridiculous GOP bias in their likely voter model. We really did some good. Take a look here: http://www.cnn.com/2004/ALLPOLITICS/10/17/poll.sunday/index.html.


I would really like to see some data on a demographic that isn't being quantified; Republicans who are planning to vote for Kerry, or are disgusted enough with Bush to just stay home on Nov. 2. I wonder if the polling firms could design a poll that would sample that group? My gut sense is that it might be the Mother of All Internal Numbers.

We really need to start mentioning the polling results for the Libertarian party. They apparently are polling almost as strong as the Nader candidacy but yet only Nader is getting any mention and Badnarik not a peep. You would think that if Nader is pulling from the Dems then Badnarik could pull some from Bush. So, Ruy, please start mentioning the Lib party polling results too.

And we should start getting on these polling companies to ask about the Libertarian party candidate if they are asking about Nader.

If I remember correctly, Nader was polling about 1.8% and Badnarik about 1.2%.

As a Kerry supporter, I'd very much like to believe that the polling data that have been reported this weekend in the media constitute good news for the Kerry campaign. But somehow, I'm having a (very) hard time getting past the unpleasant fact that Kerry still trails Bush anywhere from 2 to 8 points. So, I can't help but view this article with more than a trace of skepticism.

I'm an engineer by training, and I had a fairly rigorous coarse in probability and statistics in college --- I don't mean to brag, but the probability and statistics coarses offered to science majors in most colleges are MUCH more rigorous and exacting than those offered to non-science majors.
Thus, I can easily appreciate the alice-in-wonderland aspects of statistical studies in general, and I certainly understand the need to really dig into polling data to grasp their true significance. Still, I can't help but wonder if the "Kerry's doing OK!" conclusions stated here are more self-delusion than astute analysis.

Much of the optimism expressed here has its root in the "undecideds break for the challenger" rule, coupled with Bush's sub-50percent poll standing.
But none of the elections that that rule is based on was quite like this one, which was preceded by
an attack where more Americans were killed than in Pearl Harbor. Given that background, and the fact that Bush continues to poll better than Kerry on national security concerns, is it realistic to blithely assume that undecided will continue to abide by the usual rules?

Since I'm in Japan I have no way of knowing but how much play is Kery's lesbian comment getting in the media? The Daily Howler seems to think it's getting enough to affect the race.

My feeling is he shouldn't have said it, not because it was derogatory or mean-spirited but that it's given the Rebuplicans something to scream all self-righteuous about.

Analyzing the Newsweek data in line with the 2000 exit poll ID I get Kerry leading 48-47 and on Democracy Corps I note that their sample voted for Bush over Gore in 2000 by 51-43 while Bush gets only 47 in their poll which means he has lost support among this sample of voters since 2000 and may also mean that Kerry may actually be doing even better than this poll shows since it appears to have some oversample of Bush 2000 voters.

I wish this were entirely truly, but meanwhile, over at Slate, suellentrop is starting the Kerry is falling behind bangwagon. Luckily the major media outlets are going with your narrative which I actually think is accurate as well (not that I am in any position to change the national narrative :) I emailed him calling him on his obvious bias toward Gallup by asking why other than instituional and societal inertia should we trust Gallup over other competing and viable polls?

"Men favor K/E by 50-46 and women favor Bush-Cheney by 49-43." Right!

Every once in a while I think I'm grasping at straws when I discount poll results I don't want to see. But when you see something like this, it confirms the wisdom of not taking any of these polls at face value.

At some point, I wonder if there doesn't need to be a serious national discussion about how these polls shape public perceptions -- and, indeed, about the wisdom of permitting private, profit-making firms to drive events as citizens exercise their franchise.

Good for Newsweek, and for once, good for CNN, who down-played their flawed LV model. Unfortunately, several other news sources have not followed suit; even the reputable Reuters is pimping the "8 point lead".

maybe they switched BOTH the genders AND the overall numbers
* I heard that there had been a terrible shortage of lawn signs in Wisconsin and Iowa a while back. Surely, given that this election, after 100s of millions spent, is coming down to less than a dozen states, including those two, the problem has been addressed
*The Boston Globe reports that the organizers on the ground have insisted in Arizona of going WITHOUT TV advertisement, arguing, "without the spin, we win". It would be interesting to see if Kerry can even come within a couple of points there
*polls in Nevada are said to be now a dead heat -- and that is one state that really could swing; NH trending towards Kerry
*Repubs set to steal Ohio and Fla, and that way the election
*I don't understand why more attention isn't paid to W Virginia -- it's a state that Clinton and Dukakis carried three elections in a row, and where Edwards might be a big plus campaigning
*BUSH is giving a major address on Terrorism, Kerry must be ready to follow suit with a MAJOR address that takes Bush's approach apart limb from limb. Frankly, I thought the NYU address could have been longer and meatier, with a more detailed critique of Bush on Iraq; that is what is needed on the terror issue, including rejecting the Bai spin. Oddly, if Kerry denies the Bai spin (WHY DIDN'T HE DO SO RIGHT AWAY?) and Bush denies the Social Security issue (which sounds more plausible), Kerry would be more believable. But Kerry has been Dukakissing like mad -- on the flipflop issue and now on the Bai spin on terror. Can't win that way. Maybe it's not the program to be gotten with.
* If Kerry carries all the Gore states plus Nevada and New Hampshire, and there's no court-accepted division of Colorado, the election goes into the House of Reps, where Bush gets re-installed
It seems clearer by the day that the Bush campaign ALWAYS defends itself vigorously and attacks VIGOROUSLY where it can on key issues, while Kerry (other than in the debates, where he did well enough to win the debates but didn't really knock Bush out of the ring) doesn't. As the election approaches, that could counter the 'natural' tendency of the economic issues to push the election more towards the Democrats. That would be three out of the last 5 elections where that has happened (Dukakis, Gore, Kerry).

The strange results among women may be a reflection of the "Mary Cheney attack" meme that's circulating. If that's going to take hold anywhere, it'll be among Republican, homophobic women. Republican men don't care about "cruelty", and Democrats don't see the mention as cruel. All of this being conjecture, of course, but the numbers aren't as unbelivable as they immediately seem, and they should also trend back to what's expected as people forget about the lesbian "attack".

At this point, I think w/ almost every poll showing either a tie or a 2 pt Bush lead, I don't know how anyone can come on here saying that here has been a trend other than toward stagnancy after the first debate. In fact, several sites have looked at all the polls, and they show little if no movement other than momentary blips. For example, the Zogby/Reuters poll today have them again at 45/45. I believe of the negative people who come on here several things are going on: 1) You are really incapable of differentiating between what you feel, and the reality that there is a tie or close enough situation that we shouldnt be hand wringing (notice I don't say our guy is secretly ahead or some other such nonsense), but I am also saying that it is winnable; 2) You are a Republican or conservative troll (I am deeply suspious of posts that rely on one poll, but not the general picture formed by all polls, and CW about what the polls mean. Every pollster, including the Gallup poll CW is saying that this is extremely close based on history, present anecdotal evidence, and polling data. 3) You are by nature pessimists? 4) and to me this is unlikely- you re seeing what no one else is seeing on both the left or right who all agree that this is close.

To the guy who says he is engineer- don't confuse what y ou can say definatively with engineering w/ what can be said using the same modelling systems in a social science. Polling of people and this is where we get into trouble is a very inexact science that relies on a huge number of subjective assumptions including the issue as people have mention here of voter turn out- in fact no on at this stage knows for example if the 250 percent increase in new voters in Ohio will change the polling data (Gallup for example is of Likely voters not RVs) and there are also I would argue demographic problems (ie. higher than normal African American participation that can deeply influence the outcomes in Florida (1 million black voters) and Ohio and PA but polling models assume according to weighting what these numbers will be each year based on past data. The exact thing you are saying we can not then also use for voter id of party.

Since I've not yet seen it posted, Zogby has it back at a 45-45 tie.

Now I'm really confused. Bush leads among women and Kerry among men, a reversal of all previous polls? I must have missed something last week that led to this inversion of previous voting patterns.

And exactly how did they get a 6-point GOP swing in party ID? An unfortunate possibilty is that voters responded to the "liberal, liberal" labeling--one bit of 80's/90's retro that we don't need to revive. I also wonder if the Mary Cheney brouhaha had an effect; if so, it'll likely be modest and short-lived.

Here is a thought...besides LV, and RV we really need a third category: LCV, likely counted votes. Looking back on the number of disenfranchised votes in 2000 in Florida alone, perhaps a 1-2% Kerry lead in RV, equals a dead heat RV, and a 1-2% lead for Bush in LCV.
Am I crazy and paranoid? Or part of the reality-based Community?

Our engineering friend is confusing probabilities in relatively closed systems in the physical world with those used in social sciences. The latter is no less "scientific" just because it tries to measure social type data and predict behavior. Intervening events can always intrude on anything. Murphy's Law, for ex, came out of an engineering project. Right now, the history of these elections is that the incumbent polls about 1% more than his number going in and that usually undecideds break for the challenger.

the poll is tilted toward Republican women because the stay-at-home suburban mom (who is now a demographic minority) who picks up the phone is a Republican Bush voter and her working mom or working single woman sister (the demographic majority) who is not there to pick up the phone is a democratic or an independent Kerry voter. I really think it is that simple. Whoever wins this race, what the results are going to show is that the pollsters underestimated the revolution in telecommunications since the last presidential election. In our nuclear family, we are four Kerry Voters. Two of us, age 52 and 49, are never home to get the phone. Two of us, age 18 and 20, will never have anything but a cell phone. More people were born in 1984 than in any year since the 1960s. I think it would be overly optimistic to extrapolate from our personal scenario into a Kerry win but I feel confident extrapolating that the polling methods have become obsolete


The dictum that undecideds break overwhelmingly for the challenger -- a simple historical fact, not hopeful Democrat propaganda -- has been explained this way: an incumbent officeholder, especially a sitting president, enjoys overwhelming name/ character recognition, and as much policy/achievement recognition as any politician in the country. You already know him awfully well, or as well as you could know anybody in office, and you have had all sorts of information with which to form your opinion. As a consequence, if you've come this far you're STILL undecided, that means you must have such deep reservations about the incumbent that, in the end, you are highly likely to give your vote to the "devil you don't know."

Now as to your question about whether this election is different than others b/c of the terrorist threats we face -- I would argue that, on the contrary, the thing that really makes this election different is the remarkable degree of polarization in the country. GWB is the most ideological president in memory -- do you think that will make people LESS knowledgeable of what he's done?! I believe it will fuel MORE frustration among people who are not sold by his faith-based presidency, include the undecideds, the vast majority of whom are quite unhappy with Bush (Zogby polls say only 13% of them believe Bush should be re-elected!!!) and are just waiting to figure out if they can convince themselves that Kerry is the man to replace him.

By the way, read Alan Abramowitz' "inside scoop" post of this weekend -- apparently Karl Rove does not agree with your assessment.

In this election the polling of registered voters seems to be the most accurate view. LV numbers are based on past elections or maybe "wishful thinking" on the part of the polling company as to who will vote. Without the dem/rebup sampling ratio its hard to even know how accurate the RV results are.
Its all about GOTV at this point in time. As to independents breaking for the challenger, this is an election like none other in recent history, so it is hard to really know if that will occur.
If the dems and aligned organizations can get their voters to the polls Kerry will win, its that simple.

Rasmussen reports that just 42% say their personal finances are good/excellent. (Their front page says 36%, which is what I am pretty sure was posted yesterday. HMMM...) 56% say fair or poor. I believe swing voters are looking primarily at their own pocketbook issues (rather than the greater good). If they apply the "are you better off than 4 years ago" test, I believe the election will go Kerry's way. Record high oil prices, loss of jobs, flat or down 401K's (and those third quarter statements will all arrive before election day), huge college expenses, more people slipping into poverty and without benefits, etc. This is morning in America? I don't think so.....

I am repeatedly seeing the argument that an incumbent's final percentage is in all likelihood going to be the same as his percentage in the polls just before the election. It is the challenger whose percentage number will rise. My question is how has this been demonstrated in the past? The polls showed widely varying numbers for Gore and Bush immediately before the 2000 election and we are seeing that again this year.

MYDD posts this observation:
According to Newsweek and Gallup from the weekend, Kerry leads among independents, 51-40 and 47-42 respectively. Kerry is also gaining because he is shoring up his base, with Democratic support for Kerry now close to Republican support for Bush in nearly every poll. When both polls are re-weighted according to the Party ID model that either Rasmussen or Zogby are using, Kerry holds a small lead.

What does this tell us? In all of these polls, Kerry currently holds a lead in a scenario where 2004 turnout by Party self-identification is roughly the same as 2000 turnout Party self-identification. In other words, we only lose if we get beat on turnout.

Words to live by. Sign up at Johnkerry.com today.

And another poll suggests that things are not trending Bush's way...Rasmussen has moved to a virtual tie, at 47.5% for Bush, 47.3% for Kerry. Remember that three days agao, Bush was up 3.5% in that poll.

Rasmussen and Zogby have trends going one way, the Washington Post the other. Who knows? It's close. 15 days to go. Keep at it all.

I hadn't visited Larry Sabato's "Crystal Ball" Web blog recently, so went to it this morning. And his most recent entry (October 12th): contained the following paragraph:

"We continue to believe that President Bush absolutely, positively MUST have at least a 51 percent approval rating in the nation as a whole to be reelected. Second, given our surmise that the undecideds/leaners will break somewhat more heavily for Kerry, we think Bush needs to have built a lead of at least several points overall to win narrowly. In other words, a polling tie probably results in Kerry's election."

Now, given that Bush's average approval rating for the past couple of weeks has been about 47%, and his "lead" -- in those polls where he's not tied or behind is well under 7 percent (except for a couple of outlier-polls), this seems to be a strong prognosis of victory for Kerry.

BTW, I got the group mailing today from the Kerry campaign (as did, I suspect, many of you) reporting that over the weekend, the Kerry campaign decided to spend money trying to win Colorado. Good news, that. I'm wondering what other states they've targeted. Ohio, Florida, New Hampshire, Colorado and Nevada are all Bush 2000 targets. Are there others? West Virginia?

Today there are two polls which have shown to be more accurate than gallup's etc. for past elections which show that it is tied. Rasmussen is 47 to 47, and zogby is 45 to 45. Those two polls were the most accurate in determining the past election.

I am not watching the same CNN that you are watching. They act as if the election is over. They include every poll that Bush in the lead, and do not metion Rass, Zogby and Dem. Corp.
Futhermore, I would expect a little more from Ron Brownstien. He did not have a clue about the over sampling of the repubs. Or at least he did not want to mention it.

Apologies if this has already been posted.

Under the category of "For what its worth", in an article in the Straits Times of Singapore posted on Zogby's website at


John Zogby states that:
"Polling can be uncertain only if you rely on statistics. Thats why an effective pollster has to rely on culture, history and sociology. I repeat, polling is the study of human behaviour, not simply of people's preferences."

Then he predicts a Kerry victory on Nov.2.

As I say, for what its worth.

I am slow to give out good news but BOTH the Zogby AND the Rasmussen tracking polls today show the same dead heat (rasmussen a 0.2% edge for Bush) In BOTH cases, this represents a steady gain over the past four days. In Rasmussen's case, it's the best showing since the post SECOND prez debate polls came in (Oct 10). I had expected that the whole 'soft on terrorism' issue would hurt Kerry more but so far it doesn't seem to have. Bush giving major address on terror, while MSM trolls continue a phalanx of protestation pursuant to the Bai piece, like Sebastian Mallaby in today's Washington Post and others across America. Kerry failed to counter the flipflop spin and it stuck -- he'd otherwise be in a clear lead now I'm sure. THIS BLUNDER IS BEING REPEATED, THE DUKAKIS PLAYBOOK. Kerry MUST give a major LENGTHY address, not short and catchy, completely rejecting the Bai spin, and detailing what the media have been saying. This is infinitely more important than on-the-road campaign speeches. One major address followed by a press conference and an interview on a major TV station like CNN would stop any bleeding that is or might occur. But this is crucial.
I think the Cheney flap has helped buy time on this but that has run out. Also, check out the Abigail Thernstrom protestation about the Civil Rights Commission report. (Wall St Journal opinion)
She says it is made to look like it has Commission approval but on the USCRC website it clearly states it is a "draft staff report" and it is "for commissioner review". I just mention it because it's so characteristically bogus, with the underlying theme (like the REpub convention in 2000) of playing the (white supremacist appeal) race card. I think the fuss over Cheney is an indirect way to put the whole gay issue in the eye for the fundamentalist base. They didn't break over it in 2000 and certainly won't care now -- but the reminder helps fire them up. Then there's the usual suppression of the vote efforts -- from Thernstrom NOT A PEEP about that, in all her outrage!

I think the national polls are probably oversampling and biased towards older more conservative voters. The younger voters who tend to be cellular only users will not be reached by Gallup et.al. simply because the numbers of young/new voters are not reachable for opinion sampling with the existing sample selection models.

Kerry is going to do very well. His reputation as a tough closer on campaigns is a fair predicter of the next 2 weeks.

GOTV is the key!

Please forgive the slightly off-topic post:

For those interested, I've posted my latest (10/18) survey of 62 Electoral College tracking / prediction / projection / forecast sites at http://unfutz.blogspot.com/2004/10/electoral-college-survey-1018.html.

Executive summary: The gap between Bush and Kerry continued to close, but with less speed than last week, with Bush's 50 vote lead of 2 weeks ago now down to about 10 votes. Kerry stayed where he was, at 247 to 249 votes, but Bush lost 5 to 7 votes, landing at 255 to 260 votes. About equal numbers of sites (36 and 37) showed Bush and Kerry leading or ahead, but more sites showed Bush with over 270 votes (24) than did Kerry (17). Neither candidate averaged enough votes to be the clear winner at this point.

Every day there are polls showing the race tied, and every day there is some recent poll which shows Bush with an unexpected 8 point lead.