Skip to main content

View Diary: Senate Snapshot, October 18th: GOP takeover becoming extremely unlikely (69 comments)

Comment Preferences

  •  author mischaracterizing Nate's model (1+ / 0-)
    Recommended by:
    Code Monkey

    It's correct that Nate doesn't treat all of the races as independent, but this is not only to account for systematic model error.  It is also based on the empirical observation that movements in a collection of political races between time X and election day are not independent:  Frequently the situation gets better for one party or the other in response to national events, etc.

    So put another way, Nate's 18% doesn't just reflect that the polls might systematically be biased against Republicans, but that the political climate may get better for Republicans overall.  While I think it's correct that polling is likely to be biased against our side, there's no macroscopic reason to think the national picture will turn in our favor in the next two weeks.    

    •  Precisely. It's not error per se that he's (0+ / 0-)

      correcting for, it's the chance of a wave election. If half the teabaggers do substantially better or worse than predicted, it's just more likely that the other half will, too.

      Formerly known as Jyrinx.

      “If I can't dance to it, it's not my revolution.” ― Emma Goldman

      by Code Monkey on Mon Oct 18, 2010 at 06:30:39 PM PDT

      [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site