Monday, March 23, 2015

FLG Learned Something About Statistics In College

A few days ago, Frank Bruni wrote a piece, to hype his book, arguing that where you go to college doesn't matter.  FLG is somewhat sympathetic.  If a smart kid, who could otherwise gain entrance to an elite school, decides, for financial reasons, to attend their flagship state school, they'll be fine.  In fact, if they then go on to an elite graduate school, they probably saved a bunch of money and end up in the same spot.

Regardless, FLG saw a major flaw with some of the data Bruni used to support his position:
among the American-born chief executives of the top 100 companies in the Fortune 500, just about 30 went to an Ivy League school or equally selective college.

Okay, so only around 30% of the Fortune 100 CEOs went to a selective school, but here's the issue:
elite college and universities, however, they enroll fewer than 6 percent of U.S. college students 
So, elite school graduates are roughly 5x over-represented from what one would expect from a random draw, which sorta makes it harder to argue that there isn't that much value in attending those schools.   FLG made a note to look at more of the data, but he knew he'd never get around to it.

Luckily, somebody else did.  


Andrew Stevens said...

Wow! Jonathan Wai made such a glaring, egregious error in the write-up that I have to say it casts doubt on the entire quality of his work.

He says, "Note the purple bars, which show that nearly everyone in the US elite graduated from college. This flatly contradicts the stories glamorizing college dropouts—such as Bill Gates and Mark Zuckerberg—who both were accepted by and attended Harvard." However, it is very clear, both from his introduction to the graph he's referring to and the actual paper it came from that college graduation doesn't come into it. All of those bars are only measuring attendance at college. And this is the guy who did the study making that mistake!

I haven't had time to check into his work (and will never have time to actually double check his data), but I would be very cautious before concluding anything from his study. (Never mind that he studied it in a way which is quite different from what I would recommend.)

FLG said...


I totally missed that. I can't believe it.

It's not even really confirmation bias on my part, because I am more sympathetic to Bruni's case.

Andrew Stevens said...

There are a lot of problems with virtually any way you study this. There are just tons of confounding factors. Most studies just show something like "people who go to Harvard make more money than people who go to other schools" and stop there. Well, people who own Gucci handbags make more money than people who don't. I wouldn't recommend buying a Gucci handbag in order to try to boost your income.

This study's a little better, but I'm very wary about a lot of the methodology. In particular, lumping "elite school undergrad" and "elite school graduate school" into one big pool of "attendees of elite colleges." It's fairly easy to know what percentage of undergrads go to elite colleges each year, but it's much more difficult to figure out what percentage of people have attended an elite college at some time in their lives. For example, I have a friend who is very smart and successful, but he's not really "elite." He does have an MBA from Yale, so would probably qualify on that list and he might be able to use the degree to buffalo rubes. But an MBA from Yale is not an MBA from Harvard. Yale is not "elite" at the business school level.

The very major problem with any study like this is that a lot of things which are very helpful in getting into Harvard are also very helpful in making money or getting elite jobs: ambition, intelligence, family wealth, etc. Any of those things will make both goals more likely as well as either/or. Untangling all of these is extraordinarily difficult and just about nobody has really attempted it. The schools themselves aren't always the most helpful at providing data, for one thing.

As near as I can tell, by the way, choice of major appears to have a larger correlation with future income than the college one attends. You're probably better off majoring in chemical engineering at a state school than majoring in folklore and mythology at Harvard.

Anecdotally, I have no doubt that credentials from an elite school make entree to certain types of jobs much easier. Before the crash, Wall Street was snapping up Ivy Leaguers (regardless of major) like there was no tomorrow. It may or may not be a coincidence that Wall Street basically doesn't exist any more.

Andrew Stevens said...

On that last point, it is my own personal thesis, anecdotally observed, but with no solid data to back it up, that virtually all great financial minds grew up poor. Nobody thinks about water more than someone in the desert.

Creative Commons License
This work is licensed under a Creative Commons Attribution-No Derivative Works 3.0 United States License.