Most of the controversy over affirmative action in higher education concerns undergraduate admissions, but the American Educational Research Association has just published what it calls "important findings on the impact of banning affirmative action" in six fields of graduate study -- natural sciences, engineering, social sciences, business, education, and humanities -  in four states that have enacted affirmative action bans: California, Florida, Washington, and Texas (during the period when the Hopwood-imposed ban was in place.

The author, Liliana Garces, an assistant professor in the Graduate School of Education at George Washington University, revealingly notes in her first paragraph that these bans were enacted "despite" the Supreme Court's Grutter decision, "which allows institutions to practice affirmative action...." Since she also endorses the familiar if false equation of "racial and ethnic diversity" with a "diversity of perspectives" that fosters "the innovation necessary to tackle complex research problems" and finds the "low representation" of minority graduate students "relative to their representation" in the population "troubling," she predictably laments the effects of banning racial preferences.

These effects are most pronounced in the sciences, "where the average standardized test scores of enrolled students is higher than other fields" and the gap between "underrepresented" minorities and others is correspondingly greater. Garces argues that lower GRE scores "are not necessarily indicative" of potential for success in graduate school, citing some critics of the test but not other studies that find a "substantial" correlation.

I noted here recently that defenders of racial preferences often praise their alleged benefits while ignoring their cost or -- the other side of the same devalued analytical coin -- they bemoan the supposed costs of eliminating affirmative action while ignoring any possible benefits. Prof. Garces's study, alas, is no exception. Thus she ignores all the recent "mismatch" evidence indicating that eliminating racial preferences for applicants interested in pursuing careers in science would actually increase the number of minority scientists, not to mention the substantial social benefit of ending discrimination based on race.

Finally, in Prof. Garces's study, as in so many others, the claimed benefits of producing more minority STEM graduates are often no more than unsupported shibboleths that, say, more minority mathematicians are essential to allow us to compete in "a globalized market." Where, as I asked here, is the evidence that the return on the large investment required to produce more minority scientists will generate a higher return than the increased number of Asians and Jews that we could get without discriminating against anyone?