One of the more poignant responses to the 2016 presidential election came from Paul Krugman, the CUNY economics professor and New York Times columnist. On election night, as it was becoming clear that Donald Trump would win, Krugman contributed some thoughts to a Times forum on what the election meant, under the headline “Our Unknown Country.” That pretty much summed up his argument. As he put it, “What we do know is that people like me, and probably like most readers of The New York Times, truly didn’t understand the country we live in.”
Of course Krugman meant mainly to evoke the objects of his misunderstanding, those heartland voters who’d given the election to Trump. He was saying that he and his educated, big-city readers had underestimated the ignorance, irrationality, and bigotry lying out there in the heart of American darkness; Trump’s election showed mid-America to be even more benighted than he’d imagined– more benighted than anyone could reasonably have imagined.
But there’s another way to read Krugman’s comments, and it carries a much sadder message: namely, as an unwitting confession of intellectual failure. After all, Krugman’s one of America’s best-known social scientists. He’s devoted his professional life to understanding how societies work, America’s in particular, and he’s brought brain-power, big data, and sophisticated mathematics to the job. In his Times columns and blogs, he’s sought to make all that high-end expertise available to the non-specialist reading public. It wasn’t his intent, but his election night reflections were actually telling us that after thirty years on the job, he doesn’t know what he’s talking about. His version of social science doesn’t work, and it hasn’t helped “most readers of the New York Times” either.
If you’ve read some of the posts here, you’ll know that I don’t find that failure exactly surprising. I’ve complained from the outset about the weaknesses of contemporary social science knowledge, and especially about the versions of it you get in America’s economics departments. Mostly the complaints have centered on the radical over-simplifications that the economists have foisted off on us. My line has been, you get a better understanding of the real world from novels and history than from econ department modeling, that is, from forms of knowledge that take human complexity seriously. You may even get a better understanding from the movies, though I go back and forth on that one.
But the 2016 election has foregrounded another, more surprising problem with contemporary social science–because what we’ve witnessed hasn’t been just the failure of an over-abstract vision of how people operate, the kind of abstraction built into so many economics department models. There’s also been an empirical failure, a failure to get the right kinds of data about what’s going on around us. That shouldn’t happen these days, because (as you know unless you’ve been seriously off the grid) Big Data’s become a major force in our world. We’ve now got mountains of the stuff, opinion surveys, endless statistics about income, race, homeownership, and the like, and more and more behavioral indicators, as various agencies track our web-surfing, phoning, shopping, driving, Facebook-friending, and God knows what else.
A lot of contemporary social science involves slicing and dicing these data, calculating averages and correlations, then serving them up as guides for real-life decisions– for ad campaigns, interest rate policies, terrorist monitoring programs, and any number of other purposes. You’ve heard the various metaphors that get used, like “data mining” and “connecting the dots,” and they capture the basic idea: the real story lies hidden below the human surfaces, the social scientist digs out the deep truths, or finds meaningful pictures lurking in the apparent mess. That’s why it’s “social science”– the meanings aren’t visible to just anyone, they only emerge when the scientists apply their techniques.
Usually we don’t get to test out those Big Data claims. Either there’s lots of secrecy involved, as governments and businesses insist on keeping their methods to themselves), or the circumstances are too complicated to know whether Big Data worked or not; you may not think much of the Federal Reserve’s economic policies, but no one thinks it’s easy to get the right take on our multi-trillion dollar economy, no matter how much data you assemble.
But with the election, we got to see the intellectual failure in real time, on an issue that had a simple yes/no answer. We got to see the polling misread the situation, despite endless, carefully analyzed data, and we got to see analysts like Krugman throw up their hands in bafflement. We should consider the possibility that comparable failures are happening all around us, it’s just that we don’t see the screw-ups as clearly.
Maybe 2016 can teach us this: if we want to “understand the country we live in,” we can start by seeing our fellow citizens as actual people, not as bundles of discreet data points.