The movement has been driven by assuming more Dems are going to show up to vote. Count me skeptical.
Uhhh...
I don't think it's just an assumption. Obviously they are recording what the people are saying to them.
Rasmussen weights their poll according to partisan affiliation IIRC.
More specifically, Rasmussen weights their base sample according to partisan ID, and then actual responses from the people doing the survey decide if they get deemed likely.
For example, if Rasmussen's data suggest that there is (to pull numbers out of a hat) that the totality of the population is split 40/30/30 D/R/I the total sample will be weighted to 40/30/30, but if certain groups indicate greater enthusiasm and likely hood to vote, the actual likely voter sample may differ from that 40/30/30 breakout.
Rasmussen runs an entirely seperate survey of 500 people a day that tracks party ID, and based upon year long natiopnal aggregations, he uses that (along with other data) to set targetsd for each state by partisan ID, and then "likely is a variation off that base sample based upon actual responses.
