It seems to me that Yglesias is [...] kind of jumping to conclusions. The only poll I've seen to date was conducted on Thursday and Friday (that is, half of it before Kerry had even given his speech). It showed a 4-pt. bounce for the Democratic candidate ... disappointing, but rather more than the "no bounce" conclusion of Yglesias.

Maybe other polls have started appearing, but seems to me that you generally have to wait at least a week -- sometimes more -- to get a real handle on the actual direction the opinion polls may be trending. In the end, Yglesias may turn out quite correct ... but crowing like this, so quickly, with so little real information, just doesn't look right to me.

I think that Roger is right, and I also think that the GOP has done an excellent job of spinning the Kerry bounce, both before and after the convention -- with the usual assistance from the media, of course.

First, Matthew Dowd predicted a large bounce, saying that Kerry should come out of the election with a 15 point lead. This was ridiculous, but then it's Dowd's job to be Bush's chief strategist, which raises the obvious question of whether his saying that was just part of the strategy, which, of course, it was. So, everyone knew he was deliberately raising expectations, but they nevertheless reported it anyway (Update: and some, like the NY Times even leant it some extra credibility), and even if they reported it as being spin it still got the prediction out and into circulation, where it became enough a part of the conventional wisdom to help make any bounce which didn't come near to Dowd's prediction look very much like a failure. ("I know that he probably wouldn't go as high as Bush's guy said he was, but look, he didn't get anywhere near it, so I guess he didn't do very well.")

Second, they started the public hunt for a bounce practically the moment Kerry's acceptance speech was over -- I mean, the convention technical guys were still on CNN's audio cursing the balloons for not dropping, and the GOP was already out walking around the neighborhood in their bathrobes and a flashlight calling out "Here bouncey, bouncey, bouncey! Where are you bouncey?" and whistling to attract its attention.

The point is, as Roger said, the bounce, if there is any, isn't going to show up right away, it takes time for things to settle in on people, especially when it involves either changing your mind (in the case of former Bush supporters who decide to go to Kerry) or (in the case of people heretofore undecided) making up your freaking mind in the first place (which those people seem to have a wee bit of difficulty doing).

So, yes, I think a week after the close of the convention, or maybe even 10 days (to get past the weekend to the next workweek) is a reasonable time to see a bounce starting to show up. If there's no indication of a bounce by that time, then it would then be acceptable for young Matt to take a chance and declare that there's no bounce.

Which there may not be in any case.

Before the convention began, on Swing State Project, Chris Bowers calculated that the largest bounce Kerry could get would be 4%, and that getting that much would be "remarkable." This is because of Kerry's already high poll numbers as a challenger, the relatively large percentage of strong supporters for each candidate and the relatively small percentage of undecideds, and also because Kerry was already coming off a bounce from his selection of Edwards as a running mate.

Surely the running mate announcement is usually one of the factors which goes into creating a convention bounce, but this cycle, for this convention, that factor was removed.

In fact, it should not be overlooked that a candidate is going to have more leeway for a convention bounce when he or she is not doing well overall, and there's therefore some place to go! For a challenger who's essentially beating the incumbent (if only by keeping pace with him), one would expect a modest bump rather than a dramatic bounce.

(Some may well consider that after-spin -- so be it. It still seems to me to be a reasonable interpretation of the situation, certainly much more reasonable than Matthew Dowd's objectively silly -- but strategically brilliant -- "prediction".)

P.S. I should also mention that the obsession with bounces in the national horse race/beauty contest (take your pick) numbers is a little bit of a red herring. The important "bounce" I'll be looking for will be in the poll numbers in the 22 swing states in which the election will be determined. If Kerry's electoral college numbers (as measured by the 28 sites survey below) start to move upwards, instead of down (as they've been drifting lately -- although he hasn't stopped being ahead), then we'll see that will be a real and concrete "bounce".

Update: Chris Bowers points out that there are other ways to detect a bounce:

Since Kerry's speech, his lead over Bush in unvaforables has risen from 7.4 to 10.2. Also, Bush's job rating has once again dipped into negative territory.

This is in line with my comments above, that in a tight somewhat locked race the beauty contest numbers aren't going to move much, which means you have to look for movement below the surface in the polls' innards.

Also, Billmon:

I'm just glad Kerry didn't get a bigger bounce in the polls from the convention, or the White House would have taken us to Defcon 1 by now.

Update (8/4): Kos:

While all of us would've liked that 15-point bounce Mellman warned about in his "playing the expectations game" email, reality is that we have the most polarized, most hardened electorate probably in the nation's history. Some polls have shown a bounce, others have not. We'll probably see just as little movevement after the GOP convention. Not many people are changing their minds this time around.

Also, as I've noted before (we saw it with Edwards and Reagan), bounces take a week to really manifest themselves as people digest the latest info and talk to other people about their impressions. So we may still some movement in Kerry's direction. See Rasmussen's tracking poll. The same phenomenon happened during Edwards and Reagan -- no immediate bounce as people digested the news, and then a slow, gradual rise the week after.

Will Kerry rise further in subsequent polling? Who knows. But I know I'd rather be in Kerry's shoes than Bush's.