Skip to main content

View Diary: CNN Snap Poll - Admits Oversampling Republicans (136 comments)

Comment Preferences

  •  best comment so far, maybe a bit off (0+ / 0-)

    I think your bottom line is excellent:

    Given this problem, it seems like the CBS poll may be the more valuable measure of how the debate played to average voters, but I don't think CNN's is some kind of foul play.
    I would quibble over "average," but I think you're definitely on the right track here.

    Rewinding:

    CNN's poll was measuring the opinion of registered voters who watched the debate.
    That's true, but maybe unintentionally misleading. ORC didn't draw a fresh sample of RVs who watched the debate; it called back previous respondents.

    I wouldn't expect people who agreed to be reinterviewed to be entirely representative of viewers: even if willingness were randomly distributed (surely it isn't), those people would be watching the debate differently. I wouldn't even expect the reinterview sample to be entirely representative of viewers in the original sample. So, it's quite possible that the reinterview sample has a Republican lean vis-a-vis the viewership -- and that the overall viewership had a Republican lean, which wouldn't really matter.

    ORC knows who in the previous poll said they intended to watch the debate -- whether or not they were reinterviewed. It's possible that they would obtain more "representative" results if they weighted to characteristics of self-reported intended viewers. But given the nature of the sample, I'm not sure that would be worth doing. It's less crazy than trying to weight to the entire electorate (with the problems you discussed), but it's still a thought experiment at best.

    Election protection: there's an app for that!
    Better Know Your Voting System with the Verifier!

    by HudsonValleyMark on Fri Oct 12, 2012 at 04:10:51 AM PDT

    [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site