Its the sample stupid!
&&&&&
As soon as a pollster starts talking about "weighted" sample, I think of someone adjusting numbers to suit any client or desired result.
It makes my head hurt that I am supposed to believe that
a) questions are developed to be neutral - NOT!
b) respondents are solicited randomly
c) results are then shifted around based on some kind of scientific number of how many dems/repubs men/women etc there actually are in the US.
And we are not supposed to notice that the people writing the questions, soliciting the answers, and "weighting" the results are as biased as anyone else in the country.
I would love for one time to see poll results that are not weighted - just 1006 phone calls - period - wouldn't that be a hoot.
Ya know, I had a whole class on this when I was a psych major...about random samples and stuff like that. And I don't ever remember learning that you should change the results of your sample based on some preconceived notion of the types of groups involved. The whole POINT of a random sample is that you get a reflective sample of the whole group.