Thursday, March 17, 2016

Innumeracy and the Narrative

With the 2016 election in full swing -- in all its glory and chaos -- I've noticed certain arguments being made on the Democratic side that shows a marked lack of understanding of how primary elections work in the USA.  So, this blog is going to attempt to educate the reader on a few of the more specious ones, so that they can be pushed back against effectively.

  1.  "Primary turnout is low!  We're going to get crushed in the general!"

  2. I can't say the Democratic candidate won't get crushed in the general, but I can say that primary turnout is not in any way correlated with general election turnout.

    YearPrimary %Primary deltaGeneral %General delta
    196819.2(0)61.9(0)
    197230.9(11.7)56.6(-5.3)
    197629.6(-1.3)55.1(-1.5)
    198026.0(-3.6)54.7(-0.4)
    198423.9(-2.1)56.0(1.3)
    198825.5(1.6)53.1(-2.9)
    199221.7(-3.8)58.1(5.0)
    199617.5(4.2)51.5(-6.6)
    200019.0(2.5)54.3(2.8)
    200417.2(-2.8)60.9(6.6)
    200830.2(13)61.6(0.7)
    201215.9(-14.3)58(-3.6)

    This data set is tiny, but still, the correlation is -0.196. That's noise.

  3. "Hillary Clinton's unfavorables are sky-high!  She can't win the general election!"

  4. Way, way, way too early to make this claim. Favorability numbers are extremely volatile and the track record of them predicting election results is very poor this far out from the election. In 1988, as an example, Dukakis was +18 at this point, which GHW Bush was -1. That obviously did not coast President Dukakis to victory.

    Also, the odds-on favorites for the Republican nomination, Donald Trump and Ted Cruz, both have worse favorability numbers. So, I'm not sure how this argument is supposed to actually make sense. Being viewed favorably doesn't hurt, certainly, but it doesn't indicate who is going to win.

  5. "Bernie Sanders can't win!  He's too far behind!"

  6. That is simply untrue. He can win -- he has not been mathematically eliminated.

    Is he likely to win? No. And the reason is proportional awarding of the delegates. I'll use the current numbers from http://www.thegreenpapers.com/P16/D to illustrate:

    A candidate needs 2383 delegates to secure the Democratic nomination. Currently, Clinton has 1121 to Sanders' 799 (not all of Tuesday's primaries' delegates have been finalized yet). That means there are 2131 pledged delegates yet to be awarded. Ignoring superdelegates for the moment, that means that Clinton needs 1262 pledged delegates to secure the nomination outright, or 59.22% of the remaining. Sanders needs 1584, or 74.33% of the remaining.

    Seventy. Four. Percent.

    Those are ridiculous numbers. Now, both sides can cut down the number of pledged delegates required by gaining superdelegate endorsements. Currently Clinton has 465 to Sanders' 24. Should neither candidate gain or lose further endorsements (not realistic, but an assumption required to run the math), Clinton now needs a further 797 delegates (37.4% of the remaining), and Sanders needs 1560 (73.21%). Should Sanders gain every single superdelegate not currently endorsing, he would drop all the way to needing 1336 further pledged delegates (62.69%).

    Clinton without a single superdelegate is in better shape than Sanders with every unallocated superdelegate. That is a bad position for Sanders to be in.

    So, can he win? Yes. Will he win? Very unlikely.  

No comments:

Post a Comment