Skip to main content

View Diary: Punish CNN for Bogus Poll: Remove Candy Crowley as a Moderator (281 comments)

Comment Preferences

  •  DKElections tweet (6+ / 0-)

    link:

    This "CNN only polled Southerners" thing is ridic. MoE on Southern sample was 8.5%, meaning 133 respondents, or 30% of poll (1/2)

    Proportion of Southerners in most recent Daily Kos/SEIU poll from PPP? 31% (2/2)

    •  That should be obvious from the fact that ... (2+ / 0-)
      Recommended by:
      angry marmot, Bob Duck

      the MOE for Southerners isn't identical to the MOE for the entire sample, but some things that SHOULD be obvious apparently aren't -- especially if you want your fantasy to be true.  I thought we were supposed to be the reality-based community, but things like this diary being on the Rec. list really make me wonder about that.

      Bin Laden is dead. GM and Chrysler are alive.

      by leevank on Thu Oct 04, 2012 at 09:37:50 AM PDT

      [ Parent ]

    •  I conceded the Southerners thing (0+ / 0-)

      Now try it with 50+ vs. <50.  I dare ya!

      Pro-Occupy Democratic Candidate for California State Senate, District 29 & Occupy OC Civic Liaison.

      "I love this goddamn country, and we're going to take it back." -- Saul Alinsky

      by Seneca Doane on Thu Oct 04, 2012 at 11:21:00 AM PDT

      [ Parent ]

      •  Demographics look OK to me (1+ / 0-)
        Recommended by:
        Seneca Doane

        I think you may have made a mistake with this one...

        I back-calculate 67% >50 in the CNN poll.

        In 2008, 64% of debate watchers for the first debate were >45 according to Nielson. That's really not too far off.

        •  It's certainly possible (0+ / 0-)

          The age of the median voter in 2008 was about 47.  Maybe it's just the difference between voters and debate watchers.

          67% in 50+ seems too small to me.  (Among other things, it would mean about 140 people in the "<50" group, which I'd think would have been enough for a SE that of no more than 8.5.)  May I ask (nicely, not confrontationally) if you can explain your calculation in some greater detail?

          Pro-Occupy Democratic Candidate for California State Senate, District 29 & Occupy OC Civic Liaison.

          "I love this goddamn country, and we're going to take it back." -- Saul Alinsky

          by Seneca Doane on Thu Oct 04, 2012 at 12:11:37 PM PDT

          [ Parent ]

          •  Estimate (2+ / 0-)
            Recommended by:
            raincrow, Bob Duck

            Margin of error they give us is not a simple margin of error - not sure what they're doing exactly, probably calculating a margin of error based on weighting.  BUT.  If you just back calculate the raw number based on simple margin of error you get 318 >50 out of 475 total sample. Now, we already know that their sample is 430, not 475, so this shows they are using weighted margin of error calculations or something similar, but the simple back calculations should be close enough.

            So 67% >50 is only an estimate of their demographics. But for the sake of argument let's say their sample was actually 75% >50. Would this be unreasonable given how they constructed it?

            1) We know in 2008 it was 64% >45 watching debates, but there was unusually high excitement about the election with young people which there is NOT this year. EDIT: I screwed up. That includes under 18. In 2008 it was 67% of those over 18.

            2) The poll was not of registered voters, and not of registered voters who watched the debate, but of registered voters who planned to watch the debate, AND said before the debate they'd be willing to answer a follow-up survey. (This is probably the population they are weighting to, as not everybody who said they'd complete a survey would have actually done so.) I'd bet that demographic skews older and more white than debate viewers as a whole this year.

            So I think all the numbers seem reasonable, given the construction of the survey. And I really don't think CNN could have done anything else - at least this way they were able to weight to a measured population.  Calling respondents after the debate that they hadn't called before would 1) really really piss off people on the East Coast, calling so late, and 2) they would have no way to weight their sample because they would have no idea who actually watched the debate. So I can't fault the design of the poll much either.

            Now, if they are flogging this poll on TV as representative of all voters, that's a different issue entirely, of course, but the numbers look OK to me.

    •  Oh, weird, thanks for this AdamB (0+ / 0-)

      I've removed my tip and rec (given before any updates at all this morning).

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site