Elon University is out with their latest statewide poll, and I hate giving it any more attention that it deserves (which is none). But since the media are fascinated by it for some unknown reason, I will make a couple of comments and you can judge for yourself its legitimacy.
1. I will quote directly from their press release:
The poll, conducted Sept. 15-18, 2008, by the Elon University Institute
for Politics and Public Affairs, surveyed 411 North Carolina residents.
The poll has a margin of error of 4.9 percent. The sample is of the
population in general and does not restrict respondents by their voter
eligibility or likelihood of voting in an election. (emphasis mine).
Kudos to Beckwith at Dome for at least pointing this out. Unlike this WRAL article (that may be lifted from the AP, but it has no byline, so I can’t tell) which incorrectly calls this a sample of voters.
2. Take a gander at their data. Notice anything missing, like say, demographics?
Now if a poll won’t share with you the demographics of the people who make up the sample, what good is it? They do admit that they weigh the results by race, but they don’t tell you what the percentage of black/white in the poll is!