Sunday, November 13, 2016

Election analysis

This isn't exactly a kids-learning post, so apologies to anyone disappointed. This is my attempt to organize some thoughts and analysis around the results of the 2016 US election. The questions I'm posing are "Why did Trump win/Clinton lose the presidential election?" and "What are the lessons for 2018 and 2020?"

This analysis is not complete, so apologies to anyone who wants a nicely packaged story.

I welcome data, analysis, and different perspectives supported by evidence.

Range of theories

CNN starts us off with a nice collection in their article: 24 theories why Trump won.
Before I summarize their list, I'll note that I intentionally use the combination "Trump won/Clinton lost" rather than focus on a single side. What I'm thinking is that there are factors specific to both sides as well as relative factors at play. It isn't appropriate to look at one group in isolation. That idea helps me with a simple taxonomy of their theories:

Trump-focused theories
  1. Media. There are several sub-theories: (a) social media and fake new echo chamber: Trump supporters were ill-informed because of systematic problems with modern media. Facebook is particularly cited as a key culprit. (b) celebrity plays stronger than substance, 
  2. Trump appealed to white males, so they supported him. Subtheories are (a) xenophobia/racism/sexism, (b) backlash to political correctness.
  3. Trump appealed to voters who have been struggling economically
Clinton-focused theories
  1. Voter suppression directed toward minority voters who would typically support the Democratic candidate. 
  2. Leaked information related to scandals. Some sub-theories: (a) Russia provided hacked information, (b) FBI.
  3. Third party candidates drained support from Clinton.
  4. Clinton was a weak candidate who did not appeal sufficiently to those who formerly voted for Obama (a) saddled with too much negative baggage (b) not properly tested or vetted through the primary process
Relative theories
  1. anti-establishment fervor

The most popular story

My impression is that the most popular theory is Trump 3: Trump won the votes of those who are struggling economically. One strong example from before the election is the Guardian's My Journeys in Trumpland.

Now, empathy is good and I applaud people trying to understand each other. That said, is the economic plight of the non-elite white voter the reason for this election outcome? It is a popular story with a nice human/humanizing face, but doesn’t ring true to me. Here are the reasons I’m skeptical.
First, from my own experience, this group has been struggling since the 1980s (maybe earlier). The story is not new.

Second, it looks like Trump’s overall support was about what we should expect from a “generic republican,” maybe underperforming a little. I’m basing this on the comparison against the past two elections (see here for example).

Third, the evidence that Trump’s supporters (on average, whatever that means) are fairly well-off (see 538 primaries and the income table from CNN exit polls.)

Confusion about voter turn-out

I am currently confused about voter turn-out. These articles from CNN and 538 seem contradictory, but I haven't had a chance to work through their numbers and reconcile.


Perhaps one useful test here is the degree to which incumbents were re-elected compared with past elections. I haven't yet gathered the statistics for congressional elections, but my impression is that incumbents were overwhelmingly re-elected, basically in line with past experience, if not more. See how few districts and states are cross-hatched in these maps, (marking that shows a flip in the party controlling that seat):

2016 House of Representatives results from NY Times as of 14 Nov 2016

2016 Senate results from NY Times as of 14 Nov 2016


  1. This tweetstorm, written by Brendan Nyhan during the primary period, makes the same claim I did about racial/economic pressures: they are not new or heightened in 2016.

  2. EducationRealist, a Trump supporter since the primaries, illustrates my claim that Trump support in the general was just Republicans supporting their candidate:

    “Hey, I’m a Republican. It’s kind of what we do—you know, vote for Republicans.”

    See his full post here. My two quibbles would be the claim (a) that this election was about ideas and (b) whether the support for Trump was enthusiastic (in aggregate).

  3. Perhaps this is just confirmation bias, but I didn't seek these out. Two more very good discussions of factors around the US election:
    (1) Vi Hart looks at an important demographic divide (old vs young voters) and also flags factors behind party identification: The Great Divide

    (2) Brendan Nyhan on Julia Galef's podcast, Rationally Speaking. Brendan starts out with the baseline macro political expectation that it would be a tight race with generic republican slightly favored, people generically voted along with their party affiliation. They also explore issues in the primary process that contributed to the outcome: bonuses for state winners, earlier/rushed process, crowded field of candidates splitting "establishment" vote, skewed candidate incentives.