Opinions, Context & Ideas from the TPM Editors TPM Editor's Blog

What Do the Post-Debate Polls Say?

Pcodhzqukjeohnu7m6rz
Newscom

The clearest evidence so far comes in the Reuters-Ipsos online tracking poll. Yesterday I noted that Ipsos's first day of post-debate polling showed an ambiguous picture -- some move for Romney but possibly not that much or even any. The before and after was from Obama by 6 to Obama by 5. (Remember these are individual day samples in a tracking poll rather than the full rolling sample of several days. So you've got big margins of error.)

Today though the four day rolling sample moved to only a two point margin for Obama. Two points is close. But to push a move that big in a four day rolling sample, Thursday/Friday's numbers for Romney must have been really strong, most likely in the lead for that one day. So if we're going by just this one poll it shows a big kick for Romney.

On the other hand, Rasmussen, which has a three day rolling average has been stuck at a 2 point margin for Obama for the last three days. So it doesn't seem like any move has shown up there at all.

Similarly, Gallup actually ticked one point up for Obama today, from a 4 point to a 5 point margin. Gallup is on a 7 day rolling average. So it's hard to read too much into that. But there's at least no evidence yet of a strong move for Romney.

So for national polls it's a mixed picture, though it seems clear that something pretty dramatic happened in Romney's favor in the Ipsos poll. The current PollTracker Average is Obama +3.5. But PollTracker reports online polls but does not include them in its average. So it doesn't include Ipsos.

Then there's the state numbers. Two pollsters -- Rasmussen and We Ask America -- released polls today of Ohio, Florida and Virginia. Each showed Romney either ahead or roughly tied with Obama. If Romney had sustained leads in those three states we'd be in a very different election.

But again, caveats. We Ask America is a robo-poll which has had spotty reliability this cycle. I talked to PollTracker Editor Kyle Leighton today and his impression is that they've been bouncy and reactive through the cycle, generally magnifying trends of the moment.

So let's look closer at Rasmussen's numbers. Rasmussen had Virginia +1 Romney, Florida +2 Romney, and Ohio +1 Obama. Against the context of the last month those are strong numbers for Romney.

Yet these aren't very different from Rasmussen's most recent polls of those states. In Florida, it's a 3 point move in Romney's favor since Rasmussen's previous poll in mid-September. In Virginia it's a 2 point move for Romney. In Ohio it's no move at all, though the latest poll is 50-49 Obama compared to the mid-September poll of 47-46 Obama. So you've possibly got some move in Romney's favor. But a lot of it seems simply to be Rasmussen's standard 'House Effect'.

What does it all mean? Clearly, we need more data to have any real idea. But it seems like there is a real and clear move in Romney's favor. What's not clear is the magnitude of the change. And of course its persistence we really have no sense at all on, especially since Friday's job news seems to have changed the news conversation pretty dramatically.