Bounce Comes Into Focus

Over the course of Saturday we’ve started to get a clearer picture of the effect the Wednesday night debate had on public opinion nationwide. There’s little doubt the immediate effect was to significantly boost Mitt Romney’s standing. (I’ll assume that you’ve read my post from last night going over the polls as of then. If not, you can read it here.) What we’ve learned since last night is that a significant post-debate boost for Romney has shown up in virtually every poll. Ipsos, Rasmussen and Gallup all show marked tightening — with the difference mainly corresponding to the number of days their tracking poll covers.

A poll by Gravis Marketing gave Romney roughly a 4 point lead in Colorado. That’s been a very Republican poll this cycle. But for Colorado that is almost a 10 point move from the poll they did earlier in September.

Also this evening PPP released a poll of Wisconsin which showed Obama with a narrow 2 point lead over Romney. That’s a 5 point move in Romney’s favor since their last poll there in mid-September.

Put all this together and you start to see a pretty clear picture. In the immediate aftermath of the debate, Romney made a rapid move in all the tracking polls. And that move — in the low-mid single digits — appears to have been replicated in a number of swing states to various degrees.

A bounce like this can be ephemeral of course. And the silver lining in today’s numbers for Obama is that there’s at least a bit of evidence that Romney’s momentum has plateaued or even fallen back. The Reuters-Ipsos online tracking poll moved to a 2 point margin yesterday from a 5 point margin the day before. That was with 2 days of 4 of post-debate data. Today though it held steady at a 2 point margin for Obama with 3 of 4 days of post debate data.

PPP polls also gave some hints about the polling its done over the last 3 days. PPP’s twitter feed said Friday’s polling was actually worse for Obama than Thursday. But it then noted that “Saturday interviews we’ve done for polls across the country look a lot more like our pre-debate than Friday numbers.” In other words, PPP’s data seemed to go from bad for Obama on Thursday to really bad for Obama on Friday and then back to something more like the pre-debate numbers on Saturday.

As far as I can tell, PPP and Ipsos are the only two outlets from which we have late Friday and Saturday data. And both seem to suggest either a plateau or fall off of Romney’s surge of support from the debate.

It’s worth noting that this kind of volatility is one reason many pollsters often prefer not to do quick polls right after major news events. The results can be accurate in a purely statistical sense but also quite misleading about the true state of the race. It’s not simply that opinions can shift quickly right after a major event. But surges of enthusiasm on one side or another can look like shifts in support, especially with likely voter models.

All of this remains uncertain. And the evidence for a fallback for Romney is quite fragmentary at this point. But from incomplete information, what looks clear is that Romney got an immediate boost which pushed him into something like a tie. But there’s now some evidence suggesting the surge is falling back.