Still Hazy after All These Years : The Data and Theory Behind 'Mismatch'

In the context of reviewing the book "Mismatch" by Sander and Taylor, the authors provide a comprehensive review and synthesis of dozens of social science research studies regarding affirmative action, mismatch, graduation rates and labor market earnings. In addition, the authors look at the recent graduation rates of nearly two hundred thousand black and Latino students at one hundred U.S. research intensive universities (Table 1). The authors conclude that the social science research overall, and particularly the best peer-reviewed studies, do not support the mismatch hypothesis with respect to affirmative action and African American and Latino college graduation rates and earnings (Table 2). The authors find that Sander and Taylor's contrary claims are the result of cherry-picked data. The authors next critically analyze Sander and Taylor's claim that California's ban on affirmative action brought about a "warming effect," i.e., a rise in the rates at which African Americans and Latinos applied to and accepted offers of admission from University of California (UC) campuses. The topic is relevant to the issue of "stigma," including Supreme Court Justice Clarence Thomas' judicial opinions discussing stigma and affirmative action. The authors conclude that the direct evidence on UC freshmen yield rates do not support the warming effect hypothesis, and Sander and Taylor's contrary claims are not consistent with other research and do not properly account for factors including (a) admitted candidates likelihood of choosing to enroll at private selective institutions with affirmative action instead of UC; (b) UC's anomalous price advantage in the years when affirmative action was first banned because its tuition dropped at the same time tuition at competitor institutions rose; (c) missing data; (d) inability to separately analyze African American and Latino candidates; and (e) disproportionate reliance on admittees in the bottom third of the pool, where there are pre/post affirmative action confounders such as the changed ratio of African American student-athletes. The authors show that Sander and Taylor (and Justice Thomas) rely on a theoretically impoverished and shallow account of "stigma" that is also not supported empirically. The authors provide a better, alternative test of the warming effect hypothesis by looking at 2008-12 campus racial climate survey data for twenty thousand African American and Latino undergraduates at UC campuses, University of Texas at Austin and other leading research universities (Figures 2A-2C). These survey results show that at the UC campuses where African Americans are only 2-4% of the student body, merely 34.5% of African Americans "agree" or "strongly agree" that they are respected on campus, compared to 81.8% of white students at these same schools. Results are more encouraging at research universities like UT Austin and UC Riverside that have higher "critical mass" and/or affirmative action, all of which calls into question the warming effect hypothesis. The authors conclude that the flawed and one-sided nature of the arguments in the "Mismatch" book are also evident in the disproportionate focus on underrepresented minority students when Abigail Fisher -- the white plaintiff in the high-profile Fisher v. UT Austin case recently remanded by the Supreme Court -- had a relatively lower academic index score similar to many of those black and Latino students Sander and Taylor claim are being harmed by race-conscious affirmative action programs. Affirmative action programs at U.S. selective universities, the authors conclude, provide access to important leadership, educational and career opportunities that are important to America's future