>>
For September, the targets are 39.7% Democrat, 32.1% Republican, and 28.2% unaffiliated
>>
Remember, Ras gets these targets with a 3 *month* survey. The recent Palin-derived GOP surge in affiliation won’t show in his targets at all pre-election because there’s only one month left. Next month’s targets will see a GOP increase and that will boost McCain’s daily results, but it will be averaged with August and July.
Ras results don’t move much because he forces that party mixture.
Also, McCain’s surge in polling is the result of big McCain nights in the daily surveys. They will roll off.
All bounces fade and this one will, too. But the fade may not erode much for the same party mixture reasons . . . volatility is held down by that mix.
As for some other talk floating around, some debunking:
1) There is no weekend bias in polls. If there was, the pollsters would apply their own counter bias to achieve accuracy.
2) There is no agenda in polls other than accuracy. Accuracy gets more paid contracts.
3) Bounces erode.
4) There is no failure to poll the young. Cellphones are now included by most pollsters. Many people dump the calls via caller ID, but there is no evidence GOP vs Dem voters do so more often than the other.
5) The same thing is true of lying in poll results. There is no evidence one side does it more than the other for whatever imagined agenda.