Skip to main content

View Diary: Nate Silver Calls Out Those Whose "Goal" Is To "Entertain" Rather Than "Inform" (176 comments)

Comment Preferences

  •  What was the actual margin of error on that poll? (3+ / 0-)
    Recommended by:
    TomP, walkshills, Ian Reifowitz

    I don't have a link, but I could have sworn the MOE was 3.5%

    Which means that a lead of 4 points would actually NOT be within the MOE.  

    Thus, one would need to round that 3.5 up to 4 in order to get the 4 point lead to be within the MOE.

    I am pretty sure this is what CNN did, which makes it even worse than calling a 50 - 46 lead a tie.

    "Why do we see the same old Republicans all over the news all the time when they were kicked out for screwing everything up?" - socratic's grandma

    by Michael James on Sat Nov 03, 2012 at 11:21:24 AM PDT

    [ Parent ]

    •  Found the link and MOE from that poll is 3.5% (3+ / 0-)
      Recommended by:
      walkshills, Matt Z, Ian Reifowitz

      http://politicalticker.blogs.cnn.com/...

      This important detail is buried in the second to last paragraph of the article:

      The CNN poll was conducted by ORC International from October 23-25, with 1,009 Ohio adults, including 896 registered voters and 741 likely voters, questioned by telephone. The survey's overall sampling error is plus or minus 3.5 percentage points.

      "Why do we see the same old Republicans all over the news all the time when they were kicked out for screwing everything up?" - socratic's grandma

      by Michael James on Sat Nov 03, 2012 at 11:38:44 AM PDT

      [ Parent ]

      •  I believe I read (2+ / 0-)
        Recommended by:
        BobBlueMass, Ian Reifowitz

        that margin of error of 3.5% means plus or minus 3.5% for each -- If that's true, then O could go down 3.5% and R can go up 3.5%, putting him ahead.

        Please correct me if I'm wrong.

        The GOP: "You can always go to the Emergency Room."

        by Upper West on Sat Nov 03, 2012 at 01:41:05 PM PDT

        [ Parent ]

        •  You're right (3+ / 0-)
          Recommended by:
          bontemps2012, WinSmith, Ian Reifowitz

          Obama could have as little as 46.5 and Rmoney as much as 49.5.

          But that's one poll.

          When you aggregate polls and Obama is ahead. When nearly every poll says Obama's ahead. Obama's ahead.

          Or you could use a lower confidence margin of error, say 80 or 90%, still very good odds, and Obama would clear even the Double MoE mark.

          That's what Nate and Princeton do, look at the whole data set. and assessing probabilities, not just Stats 101 confidence levels  from one poll. And that's why they see Obama ahead. Way ahead in the EC. Still gives Rmoney a 15-20% chance of winning.

          But... Wolf Blitzer should never be allowed to explain numbers, any numbers. Or CNN needs to hire someone with some actual statistical knowledge to clearly and succinctly - using big letters - write for him. The man is innumerate.

          A society is judged by how well it cares for those in the dawn of life, the children. By how well it cares for those in the twilight of life, the elderly. And, by how well it cares for those on the edge of life; the poor, the sick, and the disabled.

          by BobBlueMass on Sat Nov 03, 2012 at 02:47:56 PM PDT

          [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site