I can't seem to find it, but there was a report from a few years back that compared an income prediction based solely on incoming student demographics (using a

I can't seem to find it, but there was a report from a few years back that compared an income prediction based solely on incoming student demographics (using a regression model on things like parental income and education levels) with graduates' actual income a few years out. It still doesn't prove causality, but if there's a big difference between the two in the positive direction, it at least suggests the university is providing some kind of actual value-add, vs just passing through students who were already from a high socioeconomic background.