Wouldn't that mean that 6.6% more Democrats were sampled by Rasmussen than the new reality supplied by Gallup requires? Would not one expect 80% plus of these 6.6% to support McCain? I do not understand why the differences in methodology between Rasmussen and Gallup have to do with establishing the correct ratio of Republicans and Democrats.
Could you kindly explain?
Rasmussen should use his own methodology and numbers for his own polls. Which is what he is doing. But he admittedly states that his 90-day rolling average, which tends to smooth out the peaks and valleys, may well have a serious disadvantage in that it misses out on any quick surges in party identification.
I personally think Gallup's 1% differential is closer to reality than Rasmussen's 5.7% differential from a couple of weeks ago. I base that on spiked fundraising, previous election totals, surges in state polls, reports from the campaign HQs, some voter registration numbers, etc. But I think it's too early to say, without more concrete data.
Do I think Rasmussen poll is underestimating McCain/Palin's performance? Yes, I do.