In my debut feature at the Guardian, I present a series of objections that the Obama campaign might raise about these dynamics. Can the Clinton numbers be taken at face value? I think there is at least one reason why they probably cannot:
Please see the article for the other objections, which I think are significantly more debatable.
Increasingly, while the Obama-McCain polls are measuring an actual election matchup, the Clinton-McCain polls are measuring a hypothetical one. The polls presume an instance in which Clinton spontaneously replaces Obama as the nominee - the "Obama is struck by lightning" scenario. But in order for Clinton to actually get from here to there, a lot of blood would be shed in the process. Her nomination (1) would require her to take her case to the convention in August, and (2) would be actuated by an overwhelming number of superdelegates siding with Clinton and contradicting Obama's advantages in pledged delegates, the current Obama-Clinton national polling and some or most versions of the popular vote count.
Were this to occur, what fraction of Obama's supporters would feel as though the nomination had been stolen from him? And how many of them would turn out for Clinton in November? There is no way to know for certain. But at the very least, Clinton would need to tie down a lot Democratic votes that aren't usually in play, and would have only three months between the convention and the election to do so.