Did They Learn Correlation and Causation in College?
It looks like Peter Thiel won’t be unopposed advising kids to stay out of college
Thanks to a new report from Georgetown University economist Anthony Carnevale, and a David Leonhardt column based on Carnevale’s study, over the last few days the college‐for‐all crowd has been striking back. But they seem to have missed something in their own college training: correlation does not equal causation.
Carnevale, Leonhardt, and others’ argument is basically that there are big, positive returns on a college degree. It’s something, frankly, that’s not generally in dispute. I say “generally,” because while on average college grads make a lot more than people without a degree, there’s a lot more to the story than averages. Indeed, there are at least three major problems with making averages the basis for a universal‐college offensive, problems that Andrew Gillen recently laid out in a terrific blog post. I won’t reinvent the wheel by going into them all (read Andrew’s post) but I’ll summarize them: (1) There are huge throngs of people who attempt college and never finish, a giant population ignored when you just look at completers; (2) at least part of the college wage premium is simply a function of a degree signaling something about the intelligence, work habits, etc. that graduates already possessed; and (3) there are some majors and degrees that confer no great wage premium and are in about as much demand as Betamax or gangrene.
What is most concerning about the Carnevale report, however, is how the report and its fans make the very basic mistake of conflating correlation with causation in implying that the roughly one‐third of bachelor’s holders in jobs not requiring degrees are much better workers thanks to their BAs. They base that conclusion on degree‐holders in non‐degree jobs earning appreciably more than workers with only high‐school diplomas. Heck, a graphic to go with Leanohardt’s column trumpets that dishwashers with college degrees make a lot more than dishwashers without them, a data point seized on by the Fordham Institute’s Peter Meyer to attack anyone who dares say college isn’t the best option for everyone.
Once the dishwasher example comes up, is there any way to escape the causation/correlation problem? Any way to not at least seriously contemplate that it isn’t what someone learned in college that makes him or her a better dishwasher, but that someone able to graduate college will tend to be more punctual and reliable? Heck, even if you believed that the proverbial underwater basket weaving major existed, it would be very hard to conclude that the skills one would need to make the finest submerged wickerwork would be useful for getting dinner plates spotless, even though that often occurs underwater.
And many of the public service jobs cited in the graphic, such as firefighters? At least from what we know about teachers, government employee pay scales often give salary bumps for degrees, but degrees don’t necessarily have any bearing on job effectiveness.
People like Carnevale and Leonhardt are right to guard against efforts, especially by public‐school employees, to actively push kids away from college, in particular if that’s driven by students’ class or race. But shoving everyone into ivy walls? Based on what we know, that’s equally unjustifiable.
