Saturday, November 12, 2016

Political Science is Dead, Long Live Political Science

There is a lot to think about in the aftermath of Trump's win.  Lots of early hot talks will be wrong.  One of the first reactions has been to wonder about the value of political science:
To be clear, folks in my discipline will be thinking about this election for some time to come, trying to figure out why most political scientists (and everyone else) got it wrong.  But the idea that this is an indictment of the discipline is foolish.  It presents the entire discipline as predictors of elections.  Because my thoughts are still scrambled by anger, fear, sleep deprivation and shock, I will just listicle why this claim about political science is so very bogus:
  1. Hillary Clinton won the popular vote as predicted.  Oops.  What we got wrong was the distribution of the vote--that the rust belt states went red.  But only barely so.  It does not take too many people to show up in a few states to flip the narratives.
  2. Ours is a probabilistic social science--we can and should indicate how uncertain we are.  I pooh-poohed Nate Silver's discussions of uncertainty because I hated the metaphors, but he was right to indicate that there was significant uncertainty.  The race was too close to be so confident (I will be issuing mea culpas for years).  And there were models that predicted Trump winning.
  3. The claim that we couldn't predict the election and what that says about the discipline reminds me of what people said about the collapse of the Soviet Union.  That is, political science may be bad at predicting singular events.   This election had a lot of stuff going on that was far from normal, and we are less good at explaining outliers.  Never before had an outside power (Russia) and a non-state actor (wikileaks) conspired to favor one candidate.  That does not mean that political science is bad at explaining.  More on that below.
  4. Most people are forgetting (thanks to folks like me who wrote out of their lane) #notallpolitical scientists are elections experts. We are a big discipline addressing all kinds of political phenomenon.  Real election experts are a small proportion of those who study the US--lots of Americanists study Congress, the courts, Con law, public policy, and on and on.  Oh, and Americanists are not all political scientists, despite how it sometimes seems in department politics and when picking up certain journals.  Comparatives, IR scholars, and theorists have much to say that is not related to electoral behavior.  
  5. Oh, and there is plenty of political science that can explain what has happened and what the effects are LIKELY (but not guaranteed) to be.  
    1. There is much work on populism that can explain the rise of Trump and why he played so well.
    2. There is much work on ethnic conflict, especially on ethnic outbidding, which explains why homogeneous parties nominate ethnic nationalists and why heterogeneous parties have a hard time balancing appeals to the largest group in the country with appeals to the smaller groups.  In short, Sri Lanka and India may be better models for US politics now than ever before. 
    3. There is much work on how democracies creep into authoritarianism.  This was a growing business for poli sci, thanks to Turkey, Hungary, Poland and others, before, but will get more energy (although probably not more NSF money).
    4. There is much work on how non-violent protest is better than violence for achieving outcomes, and that is more relevant in the US today than perhaps even during the civil rights movement.
    5. There is much work on the impact of uncertainty on alliances.
    6. There is much work that I am not mentioning here--because my reading of my discipline is as incomplete as Luke's training before Bespin.  But the key point is that the field is fast and election prediction is a very, very small part of it.
 Some might say that it is too soon to defend political science against the criticisms, that doing so appears to be defensive.  And as usual, my answer (which explains why I am easily trolled) is that if someone attacks you, you either fight back or flee.  I tend to prefer the former, at least when it comes to arguments on the internet and elsewhere.

We all have much work to do after this election.  Some will organize protests, some will figure out ways to ameliorate the worst impacts of Trumpism.  The primary job of political science, as always, will be to explain what is happening, to make sense of it, and, contrary to the hot takes, it is something we do very, very well.

1 comment:

Anonymous said...

I've spent my career as a quantitative political scientist, but I've also been at it long enough to notice a bit of arrogance creeping into the conversation. Unfortunately, that arrogance has corresponded with the increasing national prominence of political scientists as "informed" pundits. Punditry and science don't coexist comfortably. Science is about skepticism, doubt, and modesty in one's conclusions. Punditry is about hot takes and bold, attention-grabbing predictions. The Monkey Cage wouldn't keep their valuable real estate in the Washington Post very long if they said, "hey, guys, this ain't physics: a lot of what we think we know is pretty tenuous, based on a limited number of studies, and entirely dependent on an unshifting set of assumptions. That may sound unfair, and perhaps it is, but after whiffing on so many predictions and explanations it may be time for our profession to take an introspective step backward, even if it costs a few of us fame and income.