Must read: Ruy Teixeira has a great breakdown of that Gallup poll

on his blog, http://www.emergingdemocraticmajority.com

Several points are worth noting here. First, there are three national polls proximate in time to the Gallup poll, all of which show a slight Kerry lead: Rasumssen (1 point among LVs on 3/29, rising to 2 points today); Pew (1 point among RVs); and Newsweek (1 points among RVs).
Second, it is possible that the Gallup results are a bit of a pro-Bush outlier because of the composition of the sample. The Gallup report on this poll points out:
There has been no change over the past several weeks in the percentage of Republicans and Democrats supporting their own party's candidate, or in the candidate preferences of independents. Rather, the increase in support for Bush over the past few weeks comes mostly from an increase in the percentage of Americans identifying themselves as Republicans, and a comparable decline in self-identified Democrats. This could be the natural result of a shift in the political environment now that the Democratic primary season has concluded.
Maybe. Or it could be that, for whatever reason, Gallup happened to draw a disproportionately Republican sample. That would explain some of the anomalies in the poll that have just been outlined. It would also imply, of course, that the bad news for Bush is a bit worse than it appears in this poll.

more...

From the article --

"It's hard to find this movement toward Bush over the last three weeks in state polls that have been released in various battleground states. Given the magnitude of the shift implied by the Gallup data, you'd think these polls would be chock full of good news for Bush. But, by and large, they have not been."