Weekend results generally favor Democrats. I do believe Rasmussen is an excellent pollster. However, I too think there is something flawed with the input, perhaps the Bradley effect comming into play. But, if this poll is accurate, McCain/Palin would win maybe 2 states. I just don’t see it.
He’s using a rolling average of party ID - that he gets from the state by state polling, IIRC - to weight the daily tracking poll...so, he’s weighting a poll with a poll as I understand it. That’s a hot mess waiting to happen. I don’t believe that this is the same methodology he used in the past.