No empirical study is ever dispositive. However, this one is qualitatively different from previous work in that it employs fixed effects to neutralize unobserved differences in student ability. You are pointing to the older version of the study, BTW. A far more comprehensive and influential paper is available in the Quarterly Journal of Economics in 2002, and is available on JSTOR. It responds to comments made about the earlier paper and demonstrates that the findings are robust to the NLS database as well as College and Beyond.

I’ll be happy to chat with you about this sometime! My paper has moved slowly through the publication pipeline, but should finally come out in the next few weeks.

If anyone with an econ background wants to tell me whether the study is any good, they can find it here:

http://www.irs.princeton.edu/pubs/pdfs/409.pdf

I wonder what would happen if the same kind of study was done with grad school.

I am by no means being merely snarky here. And it occurs to me that the “limits of cronyism” theory may dovetail with the current “over-promotion” meme. To wit: if Harvard admits black students who would do well at UMass but struggle at Harvard, then Marty Peretz’s universe comprises a bunch of white students, some of whom are obvious hotshots, and a smaller number of black ones, most of whom are struggling. Ditto for whoever TAP’s talent-spotters are at our better institutions.