Skip to main content
ABC News
Dueling Ohio Polls

Rasmussen has a new poll out in Ohio where they show John McCain with what they headline as a “modest lead”. Actually, McCain’s lead is 6 points, and 10 including leaners, which I would probably describe as a little bit more than modest. This comes on the heels of a PPP poll released just yesterday that had shown Obama with a not-so-modest lead of 8 points.

So, what the hell is going on here? Is this the party identification issue again — PPP tending to identify more Democrats in its sample than Rasmussen?

Only up to a point. Neither pollster lists their party ID figures explicitly, but from what best I can tell, Rasmussen has the numbers at about 36/43/21 (Republican/Democrat/Independent) and PPP at 32/44/24. The party ID advantage accounts for about 4 points’ worth of difference. For instance, if you took PPP’s internals and weighted them as Rasmussen does by Rasmussen’s party ID numbers, Obama would hold a 4-point lead rather than 8. That’s still pretty significantly different from McCain leading by 10 points.

What else accounts for the differences between the two polls? Rasmussen initially permits one to select a third-party candidate — and 7 percent of voters do — whereas PPP does not. But then they push voters who have picked a third-party candidate toward one of the major-party candidate with a standard “who are you leaning toward?” question — and most of those leaners wind up with McCain. So it’s possible that you have a number of fairly conservative voters who are dissatisfied with John McCain and are flirting with the idea of voting for Bob Barr — but will gravitate back toward McCain in the end.

The polls were also conducted at different times; all of Rasmussen’s interviews were conducted Monday whereas PPP’s were conducted Thursday through Sunday. The conventional wisdom has been — and frankly, my assumption has been — that Obama would get a little bit of a bounce out of his Iraq trip. This would directly contradict that, although I think we’ll need to see quite a bit more evidence before we can reach a firm conclusion.

Nate Silver is the founder and editor in chief of FiveThirtyEight.