Undecideds Are Breaking Bush

The latest round of Zogby polling shows something amazing. Nationally, Bush is up by three points. This isn’t earthshattering news, but A:) it’s a Zogby poll that measures far more Democrats than Republicans and B:) it’s a weekend sample which traditionally lowers Republican numbers. Why is Bush running so well?

About 5 percent of likely voters are still undecided heading into the final full week of the campaign, but Bush has opened a 12-point lead on Kerry among independents.

This is very bad news for the Kerry campaign. And while Zogby argues in the Reuters article that he doesn’t know how undecideds are breaking, an astute reader of the Kerry Spot notes him saying the following:

Pollster John Zogby: “Bush had a stronger single day of polling, leading Kerry 49% to 46%. For the first time, in the one-day sample Bush had a positive re-elect, 49% to the 48% who feel it’s time for someone new. Also in the one-day sample, Undecideds were only 4%. Could Undecideds be breaking for Bush?

Yesterday, Zogby noted:

Pollster John Zogby: “Another good single day of polling for President Bush. In today’s sample alone, he leads 50% to 43%—the first time we’ve seen either candidate hit 50%. Each candidate continues to tighten his own constituency, and Undecided voters are now at only 4%

Consistant with that, Zogby’s polling finds Bush up by 5 in Ohio and 3 in Florida.

If Zogby is showing these kinds of results, I’m wondering what is going on here. The ABC/WaPo tracking poll is showing a tighter race than Zogby – which is a major reversal from the norm. If it’s true that undecideds are breaking for Bush, it’s good news for the President. Nor would it be particularly surprising. The CW that undecideds break for the challenger isn’t always true on the national level, and given that this is a wartime election there are a lot of people who aren’t fond of Bush’s domestic policies but simply can’t trust Kerry to defend this country.

I have a feeling that this last week of polling will show a break for the President, and if that’s true the President’s margin of victory could get bigger in the popular vote – and if Zogby is correct about Ohio and Florida Bush may be cementing a lock on the Electoral College as well. Then again, given how reliable Zogby’s state polling is, the Electoral College race is anyone’s guess at this point…

7 thoughts on “Undecideds Are Breaking Bush

  1. Cherry-picking polls to declare your candidate is cruising to victory is a good way to get egg on your face, particularly when other polls contrast with it (Rasmussen and ABC/Washington Post). The fact is that Kerry has the same poll standing Al Gore did at this time in 2000, and Bush is an incumbent this time. It’s certainly a possibility that undecideds could break for the incumbent, but history has shown that it’s not a common scenario. Furthermore, Bush’s efforts to ramp up his already unprecedented fear-mongering to fever pitch with images of wolves and mushroom clouds could backfire as Kerry is primarily projecting optimism and could hit Bush right between the eyes with a forceful foreign policy statement in the closing weekend.

    And by the way, I expect Clinton will do far more for Kerry in Pennsylvania than socially liberal Arnold will do for Bush in Ohio.

    And another by the way, if Bush were really leading Kerry by 12 points among independents, he’d be ahead by far more than three points nationwide…unless you believe the Dems have a double-digit party affiliation advantage even though you insisted Gallup’s seven-point GOP party advantage was correct for the last two months. When you cherry pick polls, you should probably use a little common sense to determine if the methodology is the least bit credible to justify the result. With today’s Zogby numbers, the answer is clearly no…unless of course you believe Kerry and Salazar have commanding leads in Colorado.

  2. Given that Zogby oversamples Democrats, it’s possible than a 12 point lead with independents could only show a Bush lead of 3 overall. As for Zogby’s state polling, they’re hit and miss. Remember that there are different samples for each state, which means each poll can have different biases, MOEs, etc. For example, it’s clear that the CO polls dramatically oversampled Democrats, while I’d guess that the OH poll oversampled Republicans unless OH is really much redder than the other polls have indicated. (And as much as I’d love to believe that, I rather doubt it at this point.)

    The ABC/WaPo and Rassmussen numbers are showing the typical weekend depression of Republican numbers. By midweek they’ll be back to showing a 5 point Bush lead.

    If the undecideds begin decisively breaking for Bush, the margin of victory for the President will only increae. Given that his RCP average is around +3% over Kerry, the only way that Kerry can win the popular vote at this time is to hope that there’s a dramatic shift in this race. Barring an October Surprise for Kerry, I wouldn’t count on it.

    This isn’t 2000. We weren’t at war in 2000. We are at war in 2004, and Kerry’s weakness on national security will be his downfall. In no American wartime election has the more dovish candidate ever won – and I highly doubt that 2004 will be the exception to that rule.

  3. Previously, you’ve only used Sunday numbers as an excuse to discredit upticks for Kerry. It’s good to see the days of the week where Bush can’t compete with Kerry is increasing. Hopefully, Tuesday will soon be included in that growing lineup here in the next week. 🙂

  4. Again, if Zogby is showing Bush up by 3 on a Sunday, that could just as well mean that Bush is up by more as it could that Sunday polling doesn’t descriminate against Republicans.

    There’s more than one variable involved in polling…

  5. And now that I think of it, we know that Zogby applies an artificial partisan balance to his results, which may explain these figures. If we know he’s normalizing the data to be something like 37% Democrat 32% Republican and the rest independents (I don’t remember what his exact figures are, so this is just a hypothetical) when his sample has fewer Republicans due to the weekend poll he may be normalizing the Democrats to match. In that case, his results would actually be more favorable to Republicans than other weekend polls that don’t normalize their samples in any way, especially if his data about independent voters is correct.

    I’d have to delve into the raw numbers to figure that out, but if there are any enterprising statisticians willing to take a crack at it they’re welcome to try.

  6. I believe Zogby’s Democratic partisan advantage is only two points…at the most three.

    That may well be, although he’d still be normalizing the results and the consequences would be the same. If you figure (using hypothetical numbers again) that because of the weekend a pollster only gets 25% of Republicans in their sample, that would logically mean Bush would not have high numbers. If Zogby either discards Democrat respondants or extrapolates the numbers to equal 35% Republicans his numbers would still be off, but they’d be closer than the other polls.

    At least that’s the theory – without delving into the raw numbers I’ve no way to be sure.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.