I hate to call out an organization’s poll as being bad, but, the latest numbers put out by InsiderAdvantage are just plain wrong.
InsiderAdvantage conducts their poll and then weights the respondents according to race, age and party affiliation (a common occurrence in the polling community with IVR polls — PPP and SurveyUSA both do it).
Unfortunately, when you attempt to mold your data into something, the final results can get a little screwy.
First of all, their weighing is off on some demographics (equal numbers of R’s and D’s? 16% of blacks are Republicans?). Also, their weighting skews too young.
Second, the numbers they come out with make you scratch your head.
They have McCain garnering 17% of the black vote — which if you believe that 16% of blacks are Republicans sounds plausible. (And if you believe that, I’ve got a bridge you may be interested in purchasing). The real number is around 3%.
But then they only have McCain getting 6.8% of the Democratic vote. Every other poll has McCain around 20% of Democratic vote.
Basically, the crosstabs of this poll are so far off from every other poll taken recently (Civitas, PPP, SurveyUSA) that the poll cannot be considered an accurate sample of North Carolina.