This is big.
If pollsters are not taking this into consideration when making their assumptions as to what the actual electorate will be in November, they are missing what would be a landslide in Romney/Ryan’s favor.
It seems that the polls reflecting a D+4 or greater should be highly suspect.
“It seems that the polls reflecting a D+4 or greater should be highly suspect”
Especially because the 2010 exit polls the, if anything over-sampled the Ds, came up with a 35-35 D vs. R split. Looks to me like an accurate poll would use that split.