Nate Silver sifts through the before-and-after polling:
If we simply take the three polls at face value and average them together ([an improvement of] 6.7 points), they in fact point toward a statistically strong likelihood of a bounce. Concluding that there is no bounce on the basis of the ABC poll, as some smart commentators appear to have done, while ignoring the other polling, is not objective, plainly put. There should, however, be plenty more data out before the end of the week to help settle any arguments.
The real question, of course, is not whether there's been a bounce, but how long-lasting its effects might be. Bounces usually dissipate. That's why we call them bounces; they go up and they go down. If Obama's health care polling is back in the low 40s by early next week, well then, who cares if there had been a bounce -- it's entirely an academic question. I'm not yet prepared to render a prediction on this subject, although for a variety of reasons -- basically, the GOP having used up a lot of its firepower coupled with Obama having underachieved his overall approval ratings on health care reform -- I think the bounce (if there is one) is more likely to have "oomph" than it usually does.
As Jon has pointed out, of course, the crucial audience for the speech was a handful of centrist Democrats and Olympia Snowe. We should have a better idea where they all stand pretty soon as well.