Originally Posted by SoCalBronco
Mercyhurst is apparently showing a D+10 turnout model. I think we can accept as a given that whatever shows up on election day will be somewhere below D+7 (2008 turnout- Democrat recent high in last several cycles), and prolly around the norm of D+3.
I wish we could get state by state polling using a D+3 model. Most of what we're seeing now is at or even above 2008 levels...not realistic.
SoCal, where are you getting the D +10 number? I'm not seeing it in the methodology of the Mercyhurst study. As a general rule, I've always understood polling to count a higher turnout for Republicans (given the usual class reasons).
Alternative, Nate Silver, looks at the disparity in polls using cellphones and ones not (and how the former favor Obama). http://fivethirtyeight.blogs.nytimes...de-cellphones/