Do elite high school produce elite results…over and above what would be expected from the fact that students admitted to elite high schools are academically gifted? Will Dobbie and Roland Fryer looked at three New York City high schools where entrance is by exam only. As a first step, the authors show that you really do end up with much better peers at these high schools. Here’s their picture showing peer 8th grade scores against high school entrance exams. Note the discrete jump in which students who get into an elite school end up with higher performing peers than those who don’t. (The exam cutoffs are given by the dashed lines.)
Nope. Here’re the authors’ pictures. Note that there is not an interesting discontinuity on outcomes.
Having shown that peers are different in the elite school, and noting that teacher pay and other characteristics aren’t terribly different in the elite schools, Dobbie and Fryer argue that there aren’t any peer effects. It would have been better if the conclusion had been limited to that the total effect of the elite schools is roughly zero. That is, that the effect of academically gifted peers plus whatever else makes Stuyvesant et. al. different is roughly zero. Maybe the elite schools have very positive programs that offset negative effects of smart peers, or something.
Whatever is going on, the finding that three of America’s best-reputed high schools don’t seem to add much to what their smart, incoming students bring is pretty surprising.