What Goes Up…
So in the last month Obama has actually lost ground in Colorado, which is one of the states that he needs to flip to be able to win without Ohio. Third party support is impressively high at 8%, and 7% are undecided. 50% believe Obama to be too inexperienced, and his favs have dropped to just 50%. The good news for him is that he has a strong lead among independents (14% of independents prefer a third party candidate) and he does get 77% among Democrats, but the inexperience question reveals some dangers for him: 25% of Democrats think he is too inexperienced, as do 41% of independents. If those who take this view and those who aren’t sure either way swing to McCain, Colorado slips out of reach.
Rasmussen also shows Obama still trailing in Florida and basically tied in Ohio. Someone’s model of likely voters–Rasmussen or Quinnipiac’s–is badly off.
I wouldn’t put much stock in any single poll. They have been so far off most of this season it’s a little absurd. Now we have McCain up in Georgia by only +!. WTF? In what universe is Obama losing ground in Colorado and gaining hugely in Georgia?
Well, “likely votes” is certainly a huge part of the problem, though not the only one. Voter turnout in the Dem primaries was unprecedented, and if the general election follows that pattern, what with Obama’s huge voter turnout organization, this won’t even be close. The good thing about all these conflicting polls is that they require pundits to actually THINK and reason about this election, rather than just slavishly quote certain polls while ignoring others and then fit them to their pet theory.
btw, the Rasmussen Colorado polls both have margins of error of +/-4.5, making it well within reason that NOTHING has changed in Colorado, not even looking at likely voter patterns.
According to Rasmussen, McCain is up in Georgia by 10, but supposing for a moment that the Insider Advantage poll is reliable (I have little reason to think so, but let’s go with it) we can understand the difference between them quite easily. InsAdv has Obama at 43 in Georgia, and Rasmussen had him at 41 a little over a week ago, which is pretty close. McCain is at 51 in Rasmussen, but 44 in InsAdv, but the latter poll also includes a named third party candidate in Bob Barr, who pulls 6% in his home state. The generic “some other candidate” vote in Rasmussen’s Georgia poll is also 6 percent, but it seems likely that the undecided vote will increase when there are three named candidates that the respondents have heard of before than when there are just two.
That probably means that the undecided vote is higher in the InsAdv poll, and in the Rasmussen two-way poll that undecided vote more readily goes to McCain in a traditional Republican state such as Georgia. Having Barr as a named candidate in a Georgia poll scrambles the calculations because some Georgian conservatives are probably more attracted to Barr on policy or reputation or because of local connections but are conflicted about backing a third party, and so they become undecided rather than default to McCain in a two-way match-up.
“In what universe is Obama losing ground in Colorado and gaining hugely in Georgia?”
A universe in which there is higher black turnout? It’s not so crazy to imagine that Obama could improve some in Georgia while relatively losing ground in Colorado, but still end up winning CO and losing GA. Of course, he isn’t “gaining hugely” in Georgia. McCain is getting kneecapped by Barr (three cheers for Barr!). That’s what’s happening down there.
That’s very interesting about Barr. With heavy black turnout, Barr could possibly swing things to Obama, though third party candidates tend to poll heavy early and then fadeout. Thanks for helping clarify what’s going on there.
And then there’s the new Newsweek poll that has Obama up nationally by +15. That’s way out of the recent range. And then up by +12 I think in PA. Seems crazily optimistic for Obama, but you never know I guess.
One possible factor: McCain’s campaign has been running television ads in Colorado with noticeable frequency.