Yesterday, I criticized Newsweek for passing off its poll as a measure of Kerry's bounce from the convention, when it didn't provide the right data for doing so--the "before" poll was too early (July 8-9) and, most important, the "after" poll wasn't really after since half of it was conducted Thursday, before the end of the convention and Kerry' speech. Since Newsweek's own poll indicated that Kerry did much better in the second half of their poll, after his speech, than before and since other data indicated that Kerry's pre-convention support was probably less than that indicated by their July 8-9 poll, it seemed to logical to me that their measure was probably an underestimate of Kerry's true bounce.

But now we have Gallup data that were collected entirely after the end of the convention (Friday and Saturday) and compared to a poll reasonably close to the beginning of the convention. And, quite oddly, they show, if anything, less of a bounce than that measured by Newsweek. Newsweek, in the RV Kerry-Bush matchup, had Kerry's support level going up a point and his margin increasing by 2 points. In Gallup's data, Kerry's support level also goes up a point in this matchup, but his margin actually decreases by a point.

Huh? How did that happen, when Gallup's polling on two days that should have been very good for Kerry and Newsweek only caught one of them? And beyond the bounce, how does Gallup wind up with a 50-47 Kerry lead, while Newsweek has the race at 52-44 for Kerry? After all, in early July, when these two polls were conducted at about the same time, they wound up with virtually identical results (51-45 Kerry in Newsweek; 51-44 Kerry in Gallup) in this matchup.

Apparently, there was something about that Saturday when Gallup polled. Gallup reports, consistent with the Newsweek data, that Friday was quite a good night for Kerry. But Saturday came in very differently, with quite good results for Bush--hence, the horse race results they reported.

Maybe this was just the shortest bounce on record--it only lasted one night! But I don't know; it's hard to think of a good reason why this would be so. What could have happened on Saturday to turn things around so quickly?

Also, just to deepen the mystery, other results from the Gallup poll suggest a good bounce for Kerry in almost all other respects. Compared to their pre-convention poll, Bush's approval rating went down and his disapproval rating went up. By 57-39, the public now agrees that Kerry "has the personality and leadership qualities a president should have" (up from 53/41) and by 52-43, the public now says they agree with Kerry on the issues that matter most to them (up from 49-42).

On who can best handle specific issues, Kerry's lead on the economy has gone from 8 to 11 points; on Iraq, he has gone from -5 to +2; on terrorism, from -18 to -12; on health care from +17 to +21; and on taxes from +2 to +3. Kerry has also regained the lead over Bush on who can manage the government effectively (+1); increased his lead from +8 to +15 on "cares about people like you"; reduced his deficit from -19 to -9 on who is a strong and decisive leader; took the lead over Bush on "is a person you admire" (+2); increased his lead on having an optimistic vision for the future from +3 to +11; took the lead over Bush on being honest and trustworthy (+5); and registered a large lead on "will unite the country, not divide it" (+13).

And how about this one: Kerry is now preferred over Bush, 51-46, as the candidate the public trusts more to handle the responsibilities of commander-in-chief of the military. And he has a higher net rating than Bush on being capable (or not) of handling the commander-in-chief resonsibilities (+30 vs. +21).

Heck, Kerry even went up on having a clear plan for handling the situation in Iraq, so that his net rating on having a plan (-14) is now actually a little bit better than Bush's (-16). Before the convention, Bush's net rating on this issue (-9) was a great deal better than Kerry's (-23).

But, somehow, all this didn't affect the horse race much. Hard to figure out. It must have been a very strange Saturday.

Well, we'll see what the other polls have to say. And Gallup apparently is going to stay in the field for a third night (Sunday night), so we'll see what happens with that. In the meantime, bounce or not, Kerry still seems to have helped his general political position by the convention which, as I've argued here a number of times, is the really important thing.