I don't think he is doing that but we could use an update from anyone who actually knows what Raz is currently using for his Turnout Model.
IIRC Ras is using a D+3 model other are higher.
the ARG poll yesterday showed Romney +3 with a +4D sample.
If the numbers are like 2010 (R&D 35% I30% and apply this to the polls it comes out Romney +6. Which is the same as the CU economic model predicts.