One thing that these kind of studies usually haven't counted is comparison of graduate income with their parents' income, which usually are highly correlated. This is why there are some claim that philosophy or art major have some of the highest incomes, people who take that major usually are high class already. If this study is somehow able to eliminate that bias, I bet even more majors will have negative ROI.
> One thing that these kind of studies usually haven't counted is comparison of graduate income with their parents' income
This particular study actually does try to take that into account:
> I estimate counterfactual earnings for each program’s typical students, based on their demographics, socioeconomic status, academic ability, family background, and geographic location. [https://freopp.org/does-college-pay-off-a-comprehensive-retu...]
I couldn't find an explanation either. They mention a paper somewhere, but all the links just point to that Medium article which doesn't explain the methodology in enough detail to replicate.
