For what it's worth, I think Gak's right that there were strong headwinds blowing against the GOP this year, and McCain was definitely not a first-string candidate (the first-stringers probably saw how bad the environment was in which they'd have to run, and stayed out).
However, I find the implication (both Gak's and Jerome's) that Obama's campaign basically had the White House handed to them, and deserve little of the credit for winning, is rather silly.
I think his take on this is correct. The electoral picture is so bad right now for McCain/Palin that they really have nothing to lose campaigning in odd places. Only a true x-factor event can save them now, and by definition, no one can predict where they should campaign as a result.
I think you're right on regarding the effect this would have for GOP morale, especially since Palin is just about the only morale booster they have right now.
But it gets even worse: by focusing on downticket races, it leaves Obama/Biden almost free-and-clear to also campaign on behalf of those downticket races. In essence, the GOP would be ceding the highest battlefield to the Dems, and the Dems could use it to mercilessly shell from above.
Of course, if the RNC is desperate, they may make foolish choices.
He's got a damned good point, Jerome. Add in the fact that only the CBS poll shows this kind of movement over that time period, and I have to say you have an uphill battle supporting your theory.
The GOP's going to get a bounce, and I fully expect the McCain campaign to be leading the race post-convention. But let's not get carried away with apocalyptic interpretations of single suspect polls...
The CBS poll was run from Monday to Wednesday. I'll admit that I had a hard time deciding on how her speech would come across to the average person last night (and this is speaking as a "middle America" native). But I saw literally nothing prior to last night that would have moved people toward McCain or her.
So I think the theory that the party ID swing is actual is wholly unproven and unevidenced, at best.