Learning from our elders/locating ourselves in history

This semester, I’m taking an amazing course, Fundamentals of Population Thought, from a social demographer I greatly admire, Jenna Johnson-Hanks. The first day of class was devoted to going over the syllabus, and even this was an intellectually engaging exercise. (Yay!) I was particularly intrigued by two comments she made about the last week of class, which will focus on big data and new frontiers in social demography:

First, she compared Adolphe Quetelet – an astronomer who got into social demography because he happened to live at a time when administrative data on populations was exploding – to our contemporary computer engineers-turned-data scientists. I enjoy the joke that a data scientist is just a statistician with a Mac as much as the next person, but I really liked this comparison, and the implication that a data scientist is actually (often) someone with formal training in a totally different field who is invigorated into doing social science by the times we live in (at least in the optimistic view).

Second, she made a distinction between her generation of social demographers and mine. From about 1985 to 2005, she says, population thinkers were pushing against a prevailing trend in the social sciences toward studying fine-grained connections between covariates at the level of the individual. It hadn’t occurred to me that this is a particular methodological orientation – but of course it is, and this comment immediately illuminated my experience at ICPSR this summer, wherein much of the material assumed this orientation. Although this may still be an assumption at ICPSR (which is dominated by political scientists), the new vogue for big data means that population-level analysis doesn’t have to elbow for a place at the table.

I also thought of this comment this morning when a friend Tweeted, “mixed feelings about anthropology’s absence from the recent rash of “social science is a dumpster fire” pieces.” Maybe part of the reason anthropology (and demography) haven’t been skewered is that run-99-regressions-and-finally-connect-hair-color-and-voting-behavior is what people are laughing at (rightly or not), and that’s not what we do.

Or maybe it’s that no one cares. But if no one cared, we wouldn’t have all these data scientists.