Party ID is not irrelevant. In fact it is an interesting measure of voter attitudes. It should be looked at like approval rating or right track/wrong track. It is not however a tool to analyze the accuracy of a poll.
the problem is there is a very strong correlation between the obama score and the D advantage in the sample. Generally: a sample with D+10 and you will have Obama +9 and so on...
COnsidering I'm pretty sure that the sample will not be D+10, I disagree with Obama +9.
For now, some (bad) pollsters show a very big democratic turnout (greater than in 2008 !). I strongly oppose that!
I remember than in 2004 (the golden age of this forum), all nationals polls were corrected by vorlon with a rational party id.
And in 2004 and 2008, Rasmussen gave the final result.
