“Personally, I don’t think that you can just adjust a poll according to self-reported political affiliations.”
__________________________________
That might be singularly the stupidest thing I’ve ever read on FR regarding polls. Voter turnout is NOT random!! One more time invade you we’re unable to let it sink in the first time. Voter turnout is NOT random.
Usually, I ignore personal attacks like this. You should know better. And if you don't dial it back a notch, this will be my last response to you.
No, voter turnout is not random. But, you can't expect a telephone poll to predict it reliably, especially when the poll results are well within the margin of error.
Instead of taking my comment out of context (like the Obama Media Group does for Romney), maybe you should read the rest of my response.
Poll responders are not randomly selected, either. The "opportunity" may be random, but the choice to respond is not random. There's too many factors that essentially make a poll respondent self-selecting. So, you can't adjust the results of one self-selecting poll with the results of another self-selecting poll and say that it is "better".
However, as I noted in the part that you deleted, I think that it is ethically and intellectually dishonest to claim that a poll is "accurate" when it is clear that your samples have a significant number of outliers.