I’m no math major but if Romney is up 48-47 in a dem +3 model, if the turnout is actually R+2 the the result is 50-45 (not accounting for undecided that will likely break for Romney).
You nailed it. The polls are all over the place because nobody knows which turnout model to use.
If you assume that Obama will have the same turnout that he had in 2008, then he’s ahead (CBS, Reuters, NY Times, etc.. assume that). If you assume that we are back to the historical turnout model for presidential elections, then Obama is behind by 4-5 points.
And if we assume the the turnout model will be the same as in 2010 then we’ll have a Romney landslide.