Doubts About Columbia Law School Affirmative Action Study

When Columbia Law School accuses the law school admissions community of dissing minorities, I sit up and take notice. A recent press release from Columbia trumpets a new study it conducted that makes the following argument:

  • African-American and Mexican-American applications have been steady over the last 15 years.
  • During the same time, their LSAT scores and GPAs have been improving.
  • During the same time, the number of law schools seats has grown.
  • And yet, enrollment from among these two groups is down.

Their conclusion:

"The net result is that for African Americans and Mexican Americans, law schools are not progressing towards more inclusive admissions.  This affects everyone who is concerned about better education and a more representative legal profession."

The press release raises the alarm more generally about a "disturbing decline in minority enrollment in U.S. law schools," and the study calls for a "diversity focused admissions plan," or what the rest of us call affirmative action.

The charges being leveled are indeed troubling, if true, so I took a look at the underlying LSAC data to see if I could identify the same problems. Not even close.

Here's what I see in the data*:

Black/African American enrollment went up by 8.9% in the last ten years. Not earth-shattering, but not too shabby either.

During the same period, there was a sizable drop in the category LSAC calls "Chicano/Mexican American":  -18.9%. That does look pretty bad.

But wait... what's that row right beneath Chicano/Mexican American in the LSAC tables? That would be the row called "Hispanic/Latino," which saw a 62.4% increase in enrolled students in the same time period. That spike dwarfs the percentage increase in Asian American (43.3%), "Other" (57.3%), and white (5.8%) enrolled students. And for some reason, Columbia chose to focus on the drop in Chicano enrollment and ignore that much bigger jump in Latino enrollment.

So what's the scandal here? I don't think it's any "disturbing decline in minority enrollment," because I'm not seeing one. What's disturbing is that Columbia looks at this data, cherry-picks its numbers, and pretends there is a disturbing decline.

It's pretty cheeky to cherry-pick the data to make a point and then issue a breathless press release about how "the statistics help people focus on the numbers, not on ideology."

Another interesting omission: while the Columbia study basically accuses the admissions community of failing to admit enough minorities (hence the call for more affirmative action or "diversity focused admissions plans"), here's what the LSAC data show about the growth in the number of minority applicants who were admitted (note that admission is different from enrollment, hence the different sets of numbers):

Hispanic/Latino:                  +62.8%
Other:                                 +51.7%
Asian/Pacific Islander:          +36.5%
Black/African American:        + 7.8%
Caucasian/White:                 +  4.0%
Chicano/Mexican American:  -23.0%

Those numbers also paint a rather different picture than Columbia's study or press release. In those numbers, I see admissions officers working very hard over the last ten years to create an ethnically diverse class, and doing a pretty good job at it, too.

Incidentally, I haven't even touched on whether Asian Americans count as minorities. If they do, then Asian American admissions and enrollment data undermine Columbia's argument even further. In any event, it would be helpful for Columbia to use more precision when it throws around terms like "diversity."

But back to the study: Columbia needs to be less selective with the data, and I would encourage the media who respond to the press release (some of whom have been calling me for comment, hence this posting) to dig into the underlying data. And if I'm reading the data wrong, I'm happy to be corrected.


*Here's a caveat: the Columbia study looked at LSAC data going back to 1992. The LSAC website currently gives data going back only to the class that started in fall 1997, so I have five fewer years to work with; I do have the latest ten, however. For those of you who want to take a look, go here, then roll over the Data button, then click on LSAC Ethnic/Gender Volume Summary. In that section, I looked at the enrollment and admissions tables in particular. I compared the numbers for "Fall 2007" to the numbers for "1996-1997" and calculated the percentage changes between those two points in time.