Whew

Polls coming out immediately after the Republican Convention seemed to make this prediction (fourth paragraph) look foolishly naive. Specifically, this Time poll had Bush ahead 52-41, and this Newsweek poll had Bush ahead 54-43. There were plenty of reasons to be skeptical of both results; most of all, the methodologies used in the polls skewed heavily towards Republicans. In the Newsweek poll alone, the Republicans had a 7 point advantage in voter ID, whereas all the previous available data suggest a Democratic advantage in voter ID that is trending more Democratic. I can offer a bit of observational evidence as well, namely that nothing about the Republican presentation seemed terribly persuasive in terms of bringing independent, undecided, and wobbly Kerry voters into Bush's camp. McCain and Arnold did their things without really making any case for President Bush. Cheney and Bush themselves gave ho-hum performances that weren't likely to produce sea-change swings in the electorate. Then there was the Dark Lord Miller, whose primetime Georgia-twanged fascist screechings, and subsequent sociopathic musings about killing Chris Matthews, could only have done more harm than good for the Republicans. (Put it this way: If the Bushies didn't see Miller as a liability after Wednesday night, then why did they exclude him and his wife from the Presidential box?)

I realize that I am at an Ivy League university between New York and Boston, and thus, in the middle of Blue America's most vital artery if not its heart, but I don't think I'm so out of touch that what I perceived as an above-all boring occasion would have been the thing to tip the election to Bush. Even so, when two polls were released showing Bush ahead by 11, my heart fell into my stomach.

Ruy Teixeira, back from vacation and coming to the rescue just in the nick of time, goes in depth on the polling results here. He happens to think that the reason Time and Newsweek conducted their polls so poorly has to do with magazine publishing deadlines. That's probably the most benign explanation; yet any plausible theory puts paid to the Insta-whiners' complaints about the lib-ruhl media.

At electoral-vote.com (sorry, no permalinks available; current score is Bush 275, Kerry 247), we find what has to be the most important corrective to the notion of a major Bush bounce:

Rasmussen has started publishing a 3-day rolling average every day. For Friday, Saturday, and Sunday, (all post-speech), Bush's lead nationally has shrunk to 1.2%. Rasmussen looked at the Time and Newsweek polls we had yesterday and said the samples had too many Republicans in them. When he corrected for this effect, he concluded that the Time and Newsweek data might support the conclusion of a 3% Bush lead, not more. This observation is noteworthy because it is relatively rare when one pollster says that his colleagues blew it.

Let's get some attention on this right away, because punditizing about "bounce" and "momentum" is self-fulfilling, and the worst thing that could happen right now is for our side to give up on an election that is still very winnable.

Nonetheless, there is no getting away from the fact that Bush is ahead right now, and there is a lot of work left to do. I stand by my prediction of a swing back to Kerry, but now, unlike mid-August, I would not be comfortable putting money behind a bet on him.