Personally, I think Rasmussen is oversampling Democrats. But I don’t mind, because we can’t get complacent with the favorable polls. This is just a reminder that we have a long way to go.
There's no thinking about it. Rasmussen had an article last week stating that the gap between Dems and Pubbies had closed to 5.7% (and this was immediately after Palin's selection and before the GOP convention). Yet he's been using a model that has about 9% points more Dems than Pubbies. That's due to a rolling 90-day average he uses. No problem. He's completely up front about his methodology.
So if you think that Dems will outvote Pubbies by about 9 points in November, the race is indeed about 50/50. If you believe the 5.7% number is accurate, then McCain/Palin are still enjoying their bump.
If, like me, you think (based on new enthusiasm, fundraising numbers and registration numbers) that the gap will continue to close through November, then McCain/Palin are in pretty good shape in what is still a close race.
Polls historically have given the dhimmirats a few points more than the actual election results.
0bama will have to be 5 points up to make it an actual elective close race.
He is, but not because of bias. Rassmussen uses a 90-day rolling average of party affiliations to balance his sample. In the current sample that meant an extra 7.9% of the sample were D’s compared to R’s. His own current data says that the registration advantage is now 5.7%.
If any FReepers are Premium Members of the Rassumssen site, they can probably get the breakdown of the polls by party affilitation and recompute the result weighting the parties by current affiliation numbers instead of the rolling average. I suspect the calculation would show McCain/Palin up a few points.