Mismeasuring the mismatch: a response to Ho.

• Author: Sander, Richard H.
• Position: Response to article by Daniel E. Ho in this issue, p. 1997

Daniel Ho claims that if one tugs at a single strand of my analysis of affirmative action, A Systemic Analysis of Affirmative Action in American Law Schools, (1) the entire structure collapses. (2) As I explain briefly in this Response, Ho is wrong. Ho seems to miss the central analytical framework of my article, is vague in his claims of bias, and offers an alternative approach that violates the very methodological precepts he lays out.

I

Systemic Analysis documents that black law students are nearly two-and-one-half times as likely as white law students to not graduate from law school, (3) four times as likely to fail the bar on their first attempt, (4) and six times as likely to never pass the bar. (5) Around half of this disturbing black/white gap can be explained by differences in pre-law-school credentials, but no more than that. (6) None of my critics, including Ho, deny these basic facts, and none have proposed an alternative to my explanation, which I call the mismatch hypothesis.

I argue that large racial preferences in law school admissions elevate blacks to law schools where they labor under a significant academic disadvantage. This disadvantage leads to low grades (roughly half of black law students are in the bottom tenth of their law school classes), and very low law school grades lead more often to academic dismissal, dropping out, and trouble on the bar. (7) I reach these conclusions by using white students as a control (i.e., a group of students who generally do not receive admissions preferences) and comparing the fortunes of blacks and whites in law school and beyond.

Ho's critique of my paper focuses on Tables 5.6 and 6.1, which present the results of two linear regressions that compare the role of law school GPA, law school eliteness (measured roughly by a variable I call "law school tier"), race, and a few other variables in predicting who successfully completes law school and passes the bar. (8) I offer the regressions to establish two points: First, law school grades shape these outcomes much more powerfully than does law school eliteness. Second, blacks and whites with similar law school grades (when controlling for school and entering credentials) have virtually identical graduation and bar outcomes. Blacks and whites, then, would have the same outcomes if persons of both races with the same credentials went to the same schools, but admissions preferences induce blacks to swap good grades for more prestige. Blacks suffer from that tradeoff and have worse outcomes--lower chances of graduating and passing the bar--than do similarly credentialed whites.

II

The first part of Ho's Comment purports to critique Systemic Analysis on methodological grounds. But each criticism is oddly detached from the article itself.

First, Ho suggests that my article is flawed because there is no "control" group--a group that has not received racial preferences to whom blacks can be compared. (9) Not so: The...