« Kerry Gains in New FL Poll | Main | Kerry Holds Slight Lead in New PRC Poll »

Kerry and Bush Neck and Neck in New Gallup Poll

According to a nation-wide Gallup Poll reported on August 13th, "Among all registered voters, the race is just a one-point difference, with Bush at 48% and Kerry at 47%. The late July/early August poll had each candidate at 48%.

If independent Ralph Nader is included in the ballot, Bush receives 48% among likely voters, Kerry 46%, and Nader 3%. Among registered voters, Bush still has a one-point lead over Kerry, 46% to 45%, and Nader gets 5%."


What's the matter with Gallup? Are they registering something the others don't? Or are they just polling Texans?

They sure make me nervous. And even if they are off they're constantly offering solace to the Bushies and preventing a panic which might be helpful to Kerry.

It is a sign that either a) Kerry hasn't peaked yet (which might be a blessing) or b) something isn't connecting. Please spin me, folks.

In the text of the Gallup report, there is this comment:

"The likely voter model assumes a turnout of 55% of national adults. The likely voter sample is weighted down to match this assumption."

Is this a hint that they're weighting by party affiliation to reflect previous election turnout (and ignoring just how p*ssed off we Democrats are about Bush)?

That's the only way I can see them coming up with numbers so out-of-sync with all the other national polls of the past couple of weeks.

Michael, we are p*ssed, alright. But I'd sure like to know just how much.

No offense, really, and I sure hope you're right. But I keep hearing this everytime the polls aren't matching o u r outrage. The indication for the optimism being that unusual numbers of people are turning out to Kerry's meetings, or that people know Bush voters who've changed their minds. It's all very encouraging. And it seems to reflect in most polls, actually.

But a turnout of 55% would already be a significant increase. In 2000 it was 51. Last time it was 55 was in 1992 (okay, we won). That included a popular third party candidate. You seem to expect an ever higher turnout this time. But last time it was over 55 was actually 34 years ago (61 in 1968). (Again with a popular third candidate and no, I don't think Nader qualifies for that category.)

So, asking the other way around: Are all other pollsters brewing an overconfident "sauce"?

The Gallup organization -- like the Washington Post editorial page -- benefits from being remembered as one thing (Gallup: gold standard of polling; WP: a liberal voice) so many don't notice it has shifted into being something else (Gallup: consistent Bush-favorer; WP: war hawks).

Gallup gave Bush a 13 point lead over Gore ten days before Election 2000. Throughout the Bush presidency, it has shown a consistently higher level of support for Bush (in approvals and match-ups) than just about any other polling organization. Now, its "likely voter" model pushes that standard further, pruning it to make it MORE GOP-friendly. In the process, it sets itself apart from not only every other national poll, but also the aggregate results of all state polls -- i.e., if Bush were ahead by 1 (RV) or 3 (LV) overall, there's no way Kerry'd be taking such wide leads in FL, PA and MI, or leading NH or OH.

Now, there's always the possibility Gallup has discovered a special sauce that gives it unique insight into this year's electorate (though a LV model overstating GOP participation seems wildly at odds with anything perceptible out there this year). But the more likely conclusion is, Gallup has, for whatever reason, gone off a deep end.

It reminds me of a snarky thing my father used to say when I was a kid and persisting in some stubborn, unpopular position. He'd say "It must be the rest of the world that's wrong". On a rare occasion, I WOULD be vindicated. But most of the time, the rest of the world had it right.

When Fox has K/E ahead 3 points but Gallup has them behind, somthing is seriously wrong somewhere.

It's true, if you look at the Gallup chart for 2000, that Gallup showed Bush far ahead for most of the fall:


What Up???

The Gallup poll is getting trumpeted far above all other polls, especially when it come out showing Bush's approval rating is over 50%. How convenient that just before their convention the public is now more favorable to Bush! We'll have to wait and see if this is for real or an outlier. One indicator is whether or not other polls show approval trending up over the month. Anyway Gallup is getting way too much publicity.

It's not the likely voter screen, folks. Check out the actual release (and for that matter the original posting). Even among registered voters, Kerry is only up one, which is 5-6 points off from the other recent polls. They have some sort of methodology that is setting them apart. There are many possibilities, some of which are:

1) They're cutting corners on re-connecting with voters they can't find on the first try ... this is a very reliably pro-GOP shortcut.
2) They're calling at a time or on a day when younger voters are unlikely to be home.
3) They're not looking closely at their "random" sample to see if it makes any sense, i.e. if the R versus D ration is close to accurate.

Or, another possibility is that we'll start to see close to dead even numbers from the other polls soon. Remember, it's the battleground states that matter, not the national numbers. What seems to be happening is that Bush is rolling up huge numbers in the non-battleground states, but continuing to lose ground elsewhere.

In addition to switching from "registered votesr" to "likely voters," another thing poll readers should do is switch from national polls to state polls about now. The reason is NOT, actually, because of the electoral college. In fact, I'll bet the house that the winner of the popular vote this year will win the electoral vote too. Rather, what happens is that, inevitably, non-battleground states see lower turnout than battlegrounds (check the historic numbers ... it's not a perfectly true statement, but close). As a result, a candidate who is running up big leads in his base states but losing a lot of close ones will see his final numbers go down substantially from the final polls. That's what happened to Bush in '00, and it looks like it's happening again. Most of the "swing" states right now aren't even in the margin of error, Kerry's doing so well. But he's getting creamed in Texas and Alabama. Bush's national numbers will almost certainly be inflated as a result.

With respect to Frenchfries' point -- I'm not sure where you're getting your turnout numbers, but last I checked turnout in 2000 was exactly 55% for all over-18's (60% for all eligible voters, meaning over 18 citizens, which I strongly suspect was the universe for the Gallup poll). So in fact Gallup is assuming (at best) an identical turnout, perhaps a lower one, depending on exactly how their screen works.

Oh and one more thing ...

Here's the census link for turnout in 2000:


Also remember that turnout of REGISTERED voters in 2000 was over 86%. One of the least understood truths of American politics is that, in presidential elections, almost everyone who is registered does in fact vote. Or at least 6 out of 7. The problem is 30% of eligibles aren't registered.

I'd have to disagree with those assailing Gallup's methodology - they might very well be right.

It is going to be very hard to read this year's electorate, response rates are low, old assumptions about turnout might be way off and people are seriously polarised.

Another possibility to account for the pro-Bush results of the Gallup and the Pew polls is that Bush's relentless negative assult is starting to work. I'd love to live in a world where it didn't. But the simple fact is that tearing down your opponent is a good way to win.

Bush has got his negative strategy, but Kerry doesn't have his. Flip-flopper resonates. Kerry's got a few options: incompetent, misleading, buffoon, he just needs to pick a couple and slam away.

Kerry will "slam away" when he and his astute team decide it's proper to do so. He's run an excellent race so far, something which could never be said of Gore.

The problem, Samuel Knight, is you'd have to assume that Gallup was picking up the results of this "successful" negative barrage at a time when no one else was. There've been all sorts of polls, national and state, issued contemporaneously with this Gallup, and none of them show the same result. The poll Gallup most resembles, in fact, is ITS OWN, from two weeks ago, right after the convention -- which suggests it's their methods/sampling, not their time-frame.

Gallup is, unquestionably, different from the rest right now. If it turns out they had it right all along, and everyone else was wrong, my hat will (sadly) be off to them. But I find it difficult to lean that way right now.

If democrats are truely outraged and totally p$$sed with GWB, then kerry has no real problem no matter what the polls say.

I would tend to think that outrage would rally the party and get them mobilised to vote, as well as mobilise others to vote and to convert those repubs. who are also peeved at GWB.

I would think that this outrage would cause the democrats to support the 527s and other groups that are labouring to produce ads and cover the floor with blanket canvassing..

I would think that the outraged democrats would shower the stations like FOX with the real versions of the stories and challenge the Hannities and O'Reileys with something positive and truthful to talk about.

I am sure that outraged democrats would spread the word and sow the seeds day and night, just to show how outraged they are.

Outraged democrats will team up with other grass root organisations and literally flood the market with the Kerry message.

Outraged democrats will cover for Kerry when he mis-speaks and defend his positions to the death..

If this is what outraged and p$$sed democrats are doing, then Bush is merely a sounding brass and a tinkling cymbal.

Lets face it, if Bush is still running tied with Kerry, then the word is not getting out.. and the effect of the campaing aint biting yet.

While its true that I believe and am pretty sure that Kerry will win, I am not convinced one ounce that its because of the work of the democrats. I am happy with the work that the candidates are doing but they are a lonely pair.

I expect kerry and edwards to win simply because the electorate are simply fed up with Bush. Look.. the republicans cannot point to a single issue that Bush has done that benefits the nation. They cannot point to a single policy that is not in dispute as being useful or useless. Bush and his crew does not have a record or a platform on which to run but yet they are running tied with Kerry who has truck loads of ammuniition. Doesnt this say that something is desparately wrong?

Yes it does and hence I insist that if democrats are outraged and all these excellent descriptive terms, then they need to go out and change the numbers in the polls and change the president in November.

Being outraged and p$$sed dont means one scrap if democrats are not interested in palying a meaningful part. Being angry or any such unique term dont say squat until democrats quit chatting and start working...

Actually I am p$ssed that democrats are allowing the candidates to run this show all by them lonesome selves... what a pity.

I agree with you, Bel. Stop whining, guys. Sit and write a letter to a newspaper or TV talk show.

I put in about two hours a day for the locl Kerry campaign. I have given about three hundred a month to either Kerry, Move-on, or some Democrta's campaign. I have a yard sign, bumpersticker and pin.
I think the reason Kerry's campaign is stalling is that he isn't running that great a campaign. The Bush negativity is working. Kerr's ads (expect his war ones) are too vague and platitudinous. He should be slapping that 500billion dollar debt up in front of everyone's eyeballs. He should have a ad which says point blank that Bush took taxes off the rich and put them on the middle. he needs to be more concise, more pointed and more informative and back off on the Democratic buzz words and Democratic emotional buttons.
The move on testimonial ads are also good and provide models for what the Kerry campaign should be doing.
He doesn't know how to speak to the independents.

Can some one tell me why all these polling firms include Ralph Nader when he has made it on to a grand total of TWO STATES? Maybe there are more since I last checked but it can't be more than ten. And most of those signatures to get him on the ballot are made up by over 90% conservatives/republicans. Why can't these firms get that A) national numbers mean squat, it is the state-by-state numbers in swing states that matter and B) Nader won't be on the ballot in almost all of those swing states.

In reality, Nader will be lucky to get more than 1% of the vote. Nice try in handing the election to Bush a second time, old man. Give it a rest.


I didn't do much research. I just googled and found this:


and this:


What are your numbers?

If everyone feels and knows that they are playing their part to get kerry elected and the polls are still showing these close numbers, then I think that someone needs to get the message over to the Kerry handlers that as soon as the GOP convention is over that they have to attack Bush on all his failed policies, his almost invisible military record, his constant failures while toying around with the economy... and the list goes on forever.

The democrats has sooooooooooooooo much fodder and fuel that thy can literally blow the GOP outta the water but somehow they seem afraid to use it... or maybe they are waiting to dump just before November.

I cant quite fathom how Bush can be running so close to Kerry and yet have such dismal records. This is more fascinating than anything. Personally, I think its an indictment on the democrats and the people as a whole. How can you throw your support behind such failed leadership? CEOs who function in the manner that Bush functions gets their desk cleaned out.. stuff placed in a box and they are given a formal package at the door the next morning.

How is it that nearly 50% of the electorate think that this CEO aint doing too bad and that he deserves the right to make things worse? This is interesting beyond measure and I doubt that anyone can come up with solid explanation for this. Its highly unusual.


Frenchfries: Check out that census link I posted up higher. I think it's pretty much the gold standard for turnout numbers. Maybe you're confusing 2002 with 2000? If I remember correctly, turnout in 2002 was indeed around 50%.

I'm not confusing 2000 and 2002. (You did look at these links I posted, didn't you? One is called "americanpresidency".)

And frankly, I checked out your census link but I didn't find the exact numbers. It says somewhere "55% of the voting population". Is that it?

Here are the numbers of the Federal Election Commission for 2000:


I'm not sure what Gallup is doing but it is telling that they are the lone poll out so it's likely they're wrong. I did read a comment of Charlie Cook's where he notes that Gallup is doing "likely voter" polls right now and those are wildly off the mark this far out from election day. I'd also note that Gallup was one of many pollsters in 2000 who didn't catch the swing back to Gore in the last 5 days of the election and who didn't predict the Dem GOTV effort.

The fact that all other polls put K/E ahead, plus a lot of anecdotal evidence (huge crowds for K/E, lots of stories of former Bush voters now swearing off voting for him again) all point to K/E doing well. Plus there are undercurrents that the Gallup organization (and other pollsters, actually) are not catching. One, the dem vote is *highly* motivated to vote. The rep vote is apparently subdued. I have yet to see any evidence of any kind of "fire" to get out and vote for Bush. I'm guessing that Gallup's "likely voter" is totally missing the fact that this year's likely voter is probably going to the polls to Get Bush Out. Second, the undecided vote historically goes against the incumbent. Add to this the fact that the undecideds this year are even more anti-incumbent than usual. Incumbents get 1/3 of the undecided vote *at best*, and Bush looks likely to get not even 1/5 of it. If he's not going into election day at least 3% points ahead of Kerry, he's probably going to be toast.

Frenchfries: Well, I'll grant you that the inconsistency is strange. And also, the link I sent you was wrong. Here's a better one:


I'm going to guess that since the census report came out 2 years after the election, it may be the more accurate one. But who knows. I am totally confident that Gallup is NOT assuming a higher turnout than they think happened in 2000. That's just not their style ... they're very very conservative (small "c").

I also expect a significantly higher turnout, if only for the fact that such high percentages of people have already made up their minds. It's stunning.

I don't trust the anecdotal evidence, however: Dean drew the biggest crowds in the primaries, as well and we all knew what happened afterwards. And the Bush voters who changed their minds and talk about it? They're a good story, granted, and I hope they provoke many articles.

But many people change their minds from election to election. In a climate, though, that is slightly more hostile to one candidate (in our case: Bush since April) you only hear from one sort of switchers. The other ones keep silent. Maybe the Kerry-Republicans outnumber the Bush-Democrats (I sure hope so), but hearing from them and not much from the latter ones doesn't prove it. If you know what I mean.

Anyway, let's just hope Gallup comes to other results, soon. Before their numbers change the climate. Because I don't think the Democrats have the discipline (or stubborness) to keep the spirit up through more than, say, three weeks of bad press. There'd immediately be Mickey-Kaus-ian infighting...

I am bothered by the fact that this seems to be the poll the SCLM cites when discussing the horse race.

Bob Schieffer (sp?) on Face the Nation asked Tad Devine yesterday if Kerry had had a "reverse bounce" since Bush's number have continuously risen since August 1. Devine pointed out that most polls, including CBS's show a Kerry lead, but the loaded question probably made a bigger point than the reply.