My point is that the other polls show some movement in response to events. Rasmussen never does. The recent polls have shown the President with a lead beyond the MOE and I believe that Kerry had a lead above the MOE in some earlier polls before the SBVT ads hit. Even the ultra-partisan Chrissy Mathews agrees that Bush is ahead by around seven points. The actions of the Democrats since the Republican convention have spoken loudly that they know Kerry is behind.
Merging several polls and keeping a moving average is not the best way to determine if individual polls have leads above the MOE. I guessing that if you drop Rasmussen out, this last round of polls would show the President with an above-the-MOE lead.
But by the same token, relying on individual polls is not the best way to determine the state of the race. An individual poll could either be at the upper or lower end of the MOE, or it could be an outlier. No way to know which.
My point is that the other polls show some movement in response to events. Rasmussen never does.
I'd disagree with that assessment. You're probably looking only at the 3-day averages, which aren't real good at showing what's happening, because one day of bad data throws them off. But if you look at the week-long averages, you see movement. Not huge. But a couple points. We got it after Kerry's convention (he went up 4 in Rasmussen's poll). We're getting it now after Bush's convention (throw out Saturday's data point, and Bush is up 4-5 according to Rasmussen).
It's also important to note that Rasmussen controls for partisan breakdown in his polls. He's assuming a turnout of 39% Democrat & 35% Republican, which mimics the turnout from 2000. These other polls that have shown double-digit Bush leads do not control for partisan breakdown. They're showing a about a +7% turnout advantage for the GOP which is completely unrealistic. When you suddenly oversample Republicans, of course you're going to see big movement in Bush's direction. That doesn't make it real (at least to that magnitude).