<p>The study FCCDAD describes is similar to the Krueger and Dale studies, but with additional constraints. Hersch challenges K&D because they did not study students of similar ability whose choices varied enough in selectivity. I think it would be very hard to find enough students who were (a) admitted to top schools, and (b) chose to attend much less selective schools, while satisfying all other constraints AND tracking them over their entire lives. </p>