Score contribution per author:
α: calibrated so average coauthorship-adjusted count equals average raw count
Using data on 19,000 whole siblings, it is shown that earnings vary significantly among students who have graduated from different colleges. The cross‐section estimates are up to twice the within‐family estimates, indicating that a regression estimator of college effects that does not adjust properly for family characteristics will overestimate the earnings premium of college. This study also shows that the effects of college choice vary between sisters and brothers and that there is a relationship between teacher quality and the college effects. These findings suggest that there is no straightforward interpretation of college in individual earnings equations.