In fact, median is very very explicitly what you want to use, because outliers are just that and the vast majority of people aren't.

One study on UNC-Chapel Hill graduates in the 1980s said that the best-paying major was geography. Why? Michael Jordan majored in geography.

Yeah, that's my sense of it. Median as opposed to mean. I'm just not sure what OP's "opposed to" / preferred measure is.

It's wrong because average people are more likely to do average things in their life, like pursuing a college degree because that's the common thing to do.

Outliers on the other hand sometimes choose their own path. Sometimes they drop out of college because of financial problems, but still attend courses that interest them. Sometimes they choose to learn on the job, by making mistakes and fixing them, by screwing around disassembling stuff, by learning from mentors they admire, by trying out new things and so on and so forth.

Saying that Gates, Jobs, Zuckerberg are college dropouts really ignores that these people are/were more educated than the hordes of people coming out with master degrees. And this I think is the biggest problem with our educational system, the fact that these alternate paths aren't properly recognized.

Disclaimer: I'm also a college dropout, I regret it and intend to finish college, on the other hand I have a monthly revenue at least 3 times bigger than the software industry's medium salary in my city for senior developers.

Um, you're not OP.

What are you suggesting as a preferred measure of central tendency if not median? The usual suspects include arithmetic mean (strong outlier bias, especially for incomes), or mode (which has other problems).

Or are you suggesting that measures of central tendency are inappropriate for this sort of analysis altogether? In which case, what population/sampling moments would you use instead?

I'd suggest that 2 is false. For a policy analysis, you're interested in either total (mean) or typical (median) outcomes. Probably median. You might toss in some floor/ceiling limits as well (if a particular course of action has a highly bi-modal tendency, with one mode being particularly low, it's not so hot).

As to my own feelings on the matter: a good college education is really valuable. I've certainly met people who are brilliant and/or successful who haven't finished, or even in some cases attended college. One of the biggest problems (and it ultimately does become a problem) with the latter is what I call the "chip on the shoulder" issue: it's an awareness that the person has that they haven't been to college and a mix of compensation and confidence issues resulting.

Much of what you learn in college comes from the context: who you're exposed to, what you're exposed to, library access (much less significant today than 20 years ago, given the Internet and its resources, but there are still huge amounts of information not readily available online). You absolutely can experience much of this outside the academic environment, but you've got to work at it.

At the same time, there are many schools (and student bodies) which are uninspired, and programs which do a poor job of preparing students for the outside world. There were holes in my education. There was much that I learned in college which was self-directed, or incidental.

One of the bigger lessons was the degree to which you might want to discount some random professor/researcher's claim of something ;-)

I'm on the same boat. I didn't drop, but just not attending courses. I'm curious, however, to know why you did regret it.