Calif. Bar Exam Results and Affirmative Action Critics

From Cheryl Harris at UCLA Law:


I am writing seeking your help and counsel in preventing the disclosure of private data regarding our students that would have little research value but could produce significant harm. Rick Sander, in collaboration with two other law professors, Bill Henderson of Indiana University School of Law and Vik Amar of UC Davis, is seeking to get the California Bar Examiners to release the bar exam scores, as distinct from the the passage rates, for Black and Latino law school graduates. He wants the LSAT scores, race, gender, law school attended, repeater status, and bar exam scores for all those taking the bar exam for the first time between 1997 and 2003—the classes admitted from 1994 to 1999. He furthers wants similar data on Black and Latino graduates from the classes of 2004 and 2005. His argument is that this will help evaluate his prior claims attributing poorer bar passage rates and lower law school performance to affirmative action ( or as he prefers to call it “racial preferences” ) which admit Black and Latino students with lower entering academic credentials into institutions with significantly higher median scores.

I am attaching a National Law Journal op-ed authored by myself and Walter Allen, Professor of Education and Sociology at UCLA, explaining why the Bar Examiners should stick by their original decision to deny him access to this material. The reason they point to is that the test takers provide the background information to the bar examiners for the purpose of determining testing validity—that is whether the test is fair. There is no specific request or consent given to provide access to a group of researchers to test a hypothesis. (I should point out that this disclosure is different from that g iven to LSAC projects like the BPS study or the more recent, After the JD study, in that institutional actors like LSAC who are governing bodies for the administration of evaluations have a distinct responsibility to engage in ongoing evaluation to determine best practices—a different inquiry than verifying a hypothesis.)This privacy concern is compounded by the fact that while his team promises to take precautions in structuring how the data will be reported, given the extremely small numbers of Black students in some of the cohorts, it would be possible for someone to extrapolate from the reported data back to a particular set of people.

There are serious problems with the research model that Sander et. al. propose. While this time the research team includes people who, unlike Sander, are not committed to the mismatch thesis, the reason that the research has twice failed to get National Science Foundation funding is that as the peer review letters disclose (all of this is on Sander’s website), the project is grounded on a set of assumptions—among them that bar scores reflect what is learned in law school—and encumbered by a set of problems that skew the pool to be tested—so-called selection biases.

Rather than addressing these issues, and figure out how to redesign the proposal so that it will meet peer review, Sander has now engaged in a campaign to publicly pressure the California Bar into giving him this data. He first went to the US Civil Rights Commission which is now populated by people like Abigail Thernstrom and Gail Heriot, of the Proposition 209 campaign, who unsurprisingly support his request since his research supports their political opposition to affirmative action. Heriot wrote an article in the Wall Street Journal chastising the committee for giving into political correctness and then Sander and Amar wrote the LA Times op-ed Sept 26 to which Walter and I responded.

Sander has succe eded in getting the Board of Governors of the California Bar to review the initial decision to deny and set the matter for a public hearing on November 8, 2007 before the Board’s Committee on Regulation, Admissions and Discipline Oversight at 2:30 here in Los Angeles. Thus far, there are letters on record supporting the general idea of Sander’s project and urging the release of this data. If the board is to be fully apprised of the issues and take account of the concerns regarding potential harm, it needs to hear from as many as possible. I know that colleagues at Stanford are planning to appear and that students and alum from Stanford are wanting to be heard as well. I will be there also.

I am writing to ask if you would be willing to weigh in. Regardless of whether one thinks that the mismatch hypothesis has been empirically demonstrated or not, the problem here is that the method proposed to test it is deeply flawed and risks putting our students in harm’s wa y, without their even having given consent to such examination.

If you think you might be interested, I would ask that you contact me via email and then perhaps an appropriate response can be coordinated. Excuse the length of the email but I wanted to be as thorough as possible.

The op-ed is here.