These poll numbers do not add up. In the latest Marist poll 44%, say they’ll vote against Bush. Bush's strong fans, 38% in number, are ready to re-elect him now. “The remaining 18% are not committed either way.”
It follows that if the election were today, in order to win Bush would need to get two votes out of the undecided column for every one his opponent got: a big gap, albeit not unbridgable. (And it's early days anyway, and the Republicans are going to have 200 million dollars to make up the gap.)
But, despite the negative re-elect numbers, among those polled Bush's “approval rating” is 53%. These numbers are not mathematically inconsistent, but they don't make political sense.
It could be that of the 18% who are not ready to vote for Bush yet, 83% (15/18ths) are leaning to Bush. But that seems politically unlikely. What's going on? Is it that lots of people “approve” of him but don't approve of him enough to vote for him?
More coherent, although also slightly odd, are these numbers from Newsweek's new poll . It too finds the “approval rating” at 52%. But, 50% of those polled (including persumably at least 2% who 'approve' of Bush's on-the-job performance) don't favor his re-election, while 44% do. As for independents, usually swing voters, “A majority of independents, 53 percent, said they oppose Bush's re-election, while 40 percent favor it.”
I know that serious politicos treat the “re-elect” number as the one that matters, and that despite this the press usually focuses on the “approval” number. These polls suggest that the approval number means even less than I thought.