There is now no question about the immediate effect of last week’s debate: It produced an immediate, sizable shift in the polls toward Mitt Romney. That’s true, despite the just-announced CNN poll which shows a four-point lead for Barack Obama in Ohio. Obama partisans will take this as a sign that even Romney’s surge won’t be enough, but as always, the polling averages tell the story.

At this point, despite conservative triumphalism and liberal despair, we don’t even know whether the debate bounce will ultimately matter at all.

For example: Gallup’s polling showed an immediate five-point swing to Mitt Romney in the three days after the debate…but that swing entirely disappeared by Sunday and Monday. So far, every national poll except Rassmussen released since the debate includes at least two days from that immediate post-debate period. The best poll for Romney, the Pew survey which found a four point lead for him, covered Thursday through Sunday.

At a minimum we need to wait until October 4-6 washes out of the polls; even after that, we’ll have to see if any remaining effect dissipates over time or remains as a permanent shift.

The previous history of presidential debates suggests that debate changes are more likely to be “bounces” which go away rather than long-term bumps in support for the winner. The theory explaining those results is based on findings that most people have made up their minds by October, and that those who have already decided are likely to take in new information through partisan biases. That is, even if Obama’s supporters believe he lost the debate, they’re likely to minimize the scope of the loss…and to dismiss debates as a proper method for evaluating candidates anyway. The opposite is true as well; Romney’s supporters can be expected to exaggerate his victory, and to suddenly believe that debates should drive vote choices.

Moreover, the mechanism of polling allow for changes in head-to-head results even if all that’s changed is enthusiasm for the candidates, particularly in the way that voters answer likely voter screens.

However, none of that is to say that the current swing to Romney is a mirage. We simply don’t know. There’s good theory to suggest that at least most of any debate bounce shouldn’t last, but we also appear to have an unusually large debate bounce. Even if past data and theory agree, we’re only talking about a handful of previous electoral cycles featuring debates, and there’s always the possibility that the analysis was wrong or that something has changed.

The bottom line in the polling is that we just have to wait — and to remember to keep watching the polling averages, not individual polls.