Skip to main content

View Diary: Nate Silver Doesn't Think This Election Is Important. He's Very Wrong. (188 comments)

Comment Preferences

  •  I don't think he's gone off his rocker (8+ / 0-)

    He's good with numbers and making predictive models.

    Where he's not so good is drawing conclusions about the numbers in fields he's not an expert in.

    There's no numerical model you can create to decide that "The 2014 elections won't be very important" so claiming that based on a numerical model is somewhat foolish.

    Krugman has some great analysis in that you can't just throw numbers out there - you have to provide context and understand why the numbers are important.

    Otherwise you get into ice cream sales spike in summer, crime spikes in summer therefore ice cream sales lead to crime.

    When we stop putting leaders from the past up on pedestals and ignoring their flaws, we can start seeing our present leaders for what they really are.

    by PhillyJeff on Wed May 14, 2014 at 08:30:41 AM PDT

    [ Parent ]

    •  In fact, he seems to know next to nothing about (0+ / 0-)

      anything except the numbers.

      Thus his tendency to insist that the numbers establish what must be, rather than what will be, unless someone does something to change it.

      Long ago, the tried to predict the weather this way.

      Then computers got powerful enough to model the dynamical systems that give rise to the weather -- and the statistics-based approach died. (I'm told, however, that it's experiencing a bit of a resurgence, now that we have enormous, high-resolution databases and computers powerful enough to deal with that data.)

      Silver's approach to political prediction is the same. He is indifferent to any changes in the characteristics of the dynamical system that is electoral politics.

      To put the torture behind us is, inevitably, to put it in front of us.

      by UntimelyRippd on Wed May 14, 2014 at 08:50:43 AM PDT

      [ Parent ]

    •  Krugman right on as to why I dont like the new 538 (4+ / 0-)
      Recommended by:
      ferg, pollwatcher, Deep Texan, orestes1963
      Krugman has some great analysis in that you can't just throw numbers out there - you have to provide context and understand why the numbers are important.
      It really feels like a total data dump, with half-asses analysis at best.

      As a scientist, we see this all the time, especially with genetics.  Oh, let's sequence the entire genome of population with a disease and look for all the mutations in there.  Well, we'll find so many that we don't know what to do with them.  Data alone is just numbers.

      There needs to be context, as Krugman says, and I find it lacking on 538.  Despite the fact that Nate Silver thinks we can apply data analysis and statistics to everything, there are certain things were physical experiences and emotions mean more.

      "Give me a lever long enough... and I shall move the world." - Archimedes

      by mconvente on Wed May 14, 2014 at 10:16:47 AM PDT

      [ Parent ]

      •  some things you just can't quantify (0+ / 0-)

        And deciding the "importance" of elections may be one of them, at least for now.

        But it's kinda worse than that.  When a scientist, or a number cruncher goes out of their way to put a number on something you really shouldn't, how many in the general public are going to buy into it because it has someone with respect behind it?  Now if 97% of the professional political number crunchers tell me this isn't an important election, and they show me the data, I'll gladly change my mind and admit I'm wrong.

    •  i think you could make the argument (2+ / 0-)
      Recommended by:
      pollwatcher, bryduck

      that since it's almost certain to return divided control of the Congress that it's the election least likely to upset the status quo since forever, as even if Dems retain the Senate there's no chance of getting to overcome filibusters on legislation.  but that doesn't make it unimportant for the reasons the diarist laid out.  It does need, however, an extra argument that the consequence of 2 more years of Republican legislative veto points means Obama would be more aggressive at that point in pursuing executive actions, which is the only sense in which i can see it playing out.   Not a perfect substitute in any case.  

      I think Silver's usually pretty good at going beyond unimportant correlations and avoids inferring causation -- it was after all the subject of his book.

      Difficult, difficult, lemon difficult.

      by Loge on Wed May 14, 2014 at 10:42:00 AM PDT

      [ Parent ]

      •  If they control the Senate (2+ / 0-)
        Recommended by:
        Mokurai, Garrett

        then the efforts of executive action will be MUCH harder.

        Every time the President issues and executive order, the house and Senate will attach an amendment to a very popular piece of legislation, and he'll be forced to either reverse his action, or piss off a lot of people.

        I'll guarantee this is how it will play out.  The President is going to be forced to make a ton of decisions like, sign into law something that cracks down on NSA spying, and stops all Solar/wind incentives offered by the government.  For 2 years the Republicans will reverse many of the gains of the last 6 by attaching a gazillion amendments to extremely popular pieces of legislation, and many here will blame Obama and the Dems for reversing course.

        You simply can't ignore the powerful, powerful politics that comes from controlling both the house and the Senate.  And you can't deny the willingness of the Republicans to do ANYTHING to regain power.

        (I don't mean "you" personally)

        •  true, but in the scheme of things, (0+ / 0-)

          obnoxious amendments are small relative to the fact any further progress is effectively stymied since the Scott Brown election.  Relatively unimportant isn't the same as totally irrelevant.  

          My point is simply that while the differences between the parties have never been greater, Nate's prediction is sensible if understood as a prediction of likely outcome.

          Difficult, difficult, lemon difficult.

          by Loge on Wed May 14, 2014 at 01:19:30 PM PDT

          [ Parent ]

    •  Rachel Maddow just highlighted a Web site (2+ / 0-)
      Recommended by:
      pollwatcher, Garrett

      about that.

      Rachel Maddow 05/12/14
      Secret link between bed sheets and skiing?

      Rachel Maddow shares charts of questionable data correlations from the Spurious Correlations blog to make the point that while some correlations are coincidences, others are not.

      Back off, man. I'm a logician.—GOPBusters™

      by Mokurai on Wed May 14, 2014 at 11:48:40 AM PDT

      [ Parent ]

Subscribe or Donate to support Daily Kos.

Click here for the mobile view of the site