The Polls Were Wrong. I Saw it Coming

I sing in praise of the liberal arts.

I am not a scientist.   I am a political tourist.   I watched performers.  I watched the audience.   I wasn't trying to measure things closely.  What I was doing was acting a little like a movie critic or an anthropologist visiting people in a strange country where I happened to know the language.   

Before the election Michael Moore predicted a Trump win.  Michael Moore (and I) saw something the polls did not see. 

How could this guy be right and the scientist be wrong?

Oh, the election polls were not wrong by much, but they were off some 5% in the areas that mattered—the upper midwest swing states.  And they were systematically wrong in all the same direction.  That was where it mattered in the electoral vote.

They weren't wrong because the election for fore-ordained, that it was the Republicans' turn in the pendulum nor that Hillary was unelectable from day one, nor that the election was rigged.   She lost the electoral votes because Trump had a message people wanted to hear in the crucial battleground states.   She was talking about identity and what was wrong with Trump while Trump was talking about the economy and what was wrong with Hillary.   The opponent bashing more or less cancelled each other out, leaving Hillary talking "fair-to-women" and "minority oppression" while Trump talked about jobs.  The jobs message was better than the identity message.   A majority of white women voted for Trump, pussy grabbing talk and all.

The scientists and the statisticians among the punditry missed something that Michael Moore saw.   How it is possible a grungy Michael Moore saw something missed by real scientists and statisticians who were measuring data to a tenth one percent?  

The story teller, the movie maker Michael Moore was looking at Trump’s message and he saw that identity politics was not as strong a bond as thought in the consensus view by the pollsters and data miners. He saw that “women”" were voting like “Republicans" or like “shoppers” not like "women."   It wasn’t the gender, it was the economy, stupid.  Same with Hispanics, except there was the added complication that voters—i.e. people who were here fully legally and now voting citizens—had some resentment about line-jumpers, even if these included extended family.  So some 30% defected.

Parents, don't be alarmed if your college student children study the liberal arts.  They are useful.  Watching people, hearing their stories, and evaluating the persuasiveness of message and story narrative is the stuff of literature and history and anthropology.   Students of the humanities can collect stories and pretend to measure it to 3 or 4 significant digits, but we are measuring hopes and aspirations and fears.  It is more honest to present what we learn as narrative and explanation, not closely measurable data, so we lack the credibility borne of precision.  The most important things in life—or at least in politics—are real but they are not precise.  They are the basis for ideas that persuade:  Hope and Change, or Make America Great Again.  It is hard to weigh or measure them, so the pundits focus on the pollster and not the narrative that may show the polls wrong.  We speak of “Political Science”, but there is another reality simultaneously: Political Humanities.  The humanists saw something the scientists did not.

Polling did not fail badly, just a little.  The failure seems large because so many had been so confident in it.  Its success in the last couple of elections went to the public’s heads.   People thought they were doing science and science can be precise and therefore persuasive, even when it is inaccurate.

So now the pendulum may move back on the credibility of polls, and the commentariat will better appreciate the soft and fuzzy and immeasurable elements of a campaign: who is interesting, who tells a story people want to hear, who connects.  And they will listen to voters better, in a long form rather than a quantifiable answer to a poll question.   In that case they might tease out that a black voter or a Hispanic voter or any other voter is complex and multi-dimensional, and that it reduces their full humanity to be seen as a demographic category.  A voter might respond to an identity message or a religious value or an economic message or a fear message or a combination of them.   The voters won't be voting something closely quantifiable.  They will be voting their hopes and dreams and aspirations.  

Hopes and dreams and aspirations are the work of the humanities, the liberal arts. The humanities have their place in the world.