Do Yoon and Rothstein Rebut Sander? Gail Heriot

Over at the Volokh Conspiracy, a comment to a post by David Bernstein suggests that a recent study by Jesse Rothstein and Albert Yoon may go a long way toward disproving Sander's mismatch thesis (that racial preferences have actually reduced the number of African American lawyers).  I don't think so--not yet anyway.  Here is what I've written on the subject:

Recently, an article by Jesse Rothstein and Albert Yoon, although couched as a criticism of Sander, actually provides partial (though by no means conclusive) support for Sander’s mismatch hypothesis in the law school context. Rothstein & Yoon studied graduation, bar examination and employment outcomes. Not surprisingly, due to widespread hiring preferences, Rothstein & Yoon find that African-American law graduates have better employment outcomes than whites with similar credentials. But with regard to graduation and bar passage rates, they did indeed find what Sander’s research predicted they would find: that there are very large, unexplained disparities in graduation and bar passage rates between African-American and white law students with similar entering credentials. Rothstein & Yoon assure the reader that any mismatch effect is confined to the African-American students in the lowest quintile of entering credentials, but later in the article, they admit that 75% of all African-American law students are in the lowest quintile. (Sander also found that graduation and bar passage outcome disparities were greater in the lower ranges. See Sander at 441.)

 Rothstein & Yoon are not ready to attribute this unexplained gap to mismatch (defined narrowly to mean that the affected students would have more likely graduated and passed the bar if they had attended a less-competitive school). See Rothstein & Yoon at 29 (“For students in the bottom quintile of the entering credentials distribution, the data are consistent with sizable mismatch effects on black bar passage rates but also with differential selection into law school.”). 

 Specifically, Rothstein & Yoon, argue that comparing white and African Americans with the same academic index towards the bottom of the law student credential range can be misleading. Particularly among the disfavored whites, many law school applicants in the lower academic index range were never admitted to any law school and hence never became law students. Only those with something extra--like a particularly good undergraduate institution or a particularly tough major (attributes that cannot be detected with the LSAC-BPS data)–make it into the pool. Comparing these specially-selected whites with African-Americans with the same academic index (but who received a preference and are therefore not necessarily as academically qualified as the average member of the white group) may be deceptive. If Rothstein & Yoon are correct on this, one would expect a gap in graduation and bar passage rates between the two groups.

But there's a problem:  One issue that Rothstein & Yoon do not deal with is how African American and white students in the bottom academic index quintile differ on law school GPA. A key finding of Sander’s was that African-American and white law students with similar academic indices perform about the same when they compete against each other in law school. But when it comes time to take the bar examination, the whites substantially outperform the African Americans. Sander attributes this to the mismatch effect. Rothstein & Yoon do not weigh in about whether Sander is correct in that finding.  They don't talk about whehter “bottom quintile” whites earn higher GPAs than “bottom quintile” African Americans when they compete against each other (as their alternative model would suggest) or about the same (as the mismatch model would suggest). If Sander is correct about this, then Rothstein and Yoon's theory doesn't fit the data.

      It's also worth pointing out that Rothstein and Yoon's tentative conclsuions in no way are at odds with the Commission on Civil Rights' recommendations. Nothing in Rothstein & Yoon suggests that more research into the mismatch issue is unnecessary. To the contrary, their uncertainty as to whether the 75% of African Americans in the bottom quintile of law students are victims of mismatch or simply less qualified than the white law students with similar academic indices is an argument in favor of further research. Moreover, even if the Rothstein & Yoon alternative (i.e. that white law students in the bottom quintile have academic qualifications that are not detectible from the LSAC-BPS data set, but which nevertheless cause them to have higher graduation and bar passage rates than African Americans in the bottom quintile) turn out to be correct, that does not detract from the Commission’s recommendation that law schools disclose to all students the likelihood that a student with particular qualifications will graduate and pass the bar examination.