I looked at the data and was struck by the fact that the top schools by ten-year earnings were medical schools. This led me to believe that the data doesn't distinguish between undergraduate and graduate programs. If that's the case, then an undergrad-focused school like Bennington is going to be penalized versus a school that has mostly grad students. If your pool of students is, say, 70% grad students, then they're going to be further along in their careers at the ten year mark than an institution where 90% of students are undergrads.
