Such data also can't be used to make inferences on causality. It is quite possible that people with higher earning potentials tend also to be accepted by elite schools, instead of elite schools adding "value" to these people.
They did try to control for this. From the article:
(It’s important to note that we controlled for numerous other factors that might influence postgraduation earnings, such as family income, race/ethnicity, gender, marital status, SAT score, postgraduate degree and age at graduation and more.)
