Statistics for Lawyers and, Also, Your Test is Still Racist

Ah, Aaron Taylor.  He pops up every now and then in the sorts of high-class periodicals I peruse.  For example, in this article, exploring the relationship between declining enrollments and greater diversity at less-endowed law schools, which contains this outstanding footnote:

Hoards of so-called “scamblogs,” on which the authors, often anonymous, offer searing critiques of legal education, have gained popularity over the last few years. These blogs have tapped into the larger climate of frustration, generating attention and, of course, page views. Some of these blogs have contributed useful insights to discussions of legal education; others have contributed little more than snark. See, e.g., Profiles in Deliberate Misinformation and Dishonesty: Aaron Nathaniel Taylor, “Law Professor” at St. Louis University,THIRD TIER REALITY (Oct. 12, 2011, 6:25 AM),http://thirdtierreality.blogspot.com/2011/10/profiles-in-delib erate-misinformation.html (providing a particularly snarky assessment of my views on legal education).

Feel the burn, you gnarly whoreson!  If that isn't like lemonade on a sunny Saturday afternoon...

His latest triumphantly castigates the LSAT and law school admissions folks for being racist. I feel like I've heard this record before, but it feels fresh every. single. time.  So academics, don't worry - you can repeat yourself and still get tenure. Look at how often the Beatles get played on the FM radio,or the teeming millions who read this blog. People want to hear good shit over and over again.

The actual article is about "marginalization," a word omitted by the Orwellian press cited above entirely.  The piece deep-dives into this concept of marginalization, not just Du Bois-for-idiots but getting out the dictionary Scalia-style:

The terms marginalization and its lexical cousin, marginality, are both derivatives  of  the  word marginal.Thus,  the  meanings  and  conceptions  of marginalization and marginality descend from how marginal is defined.  The Oxford dictionary defines marginal as “[r]elating to or situated at the edge or margin of something.” Therefore, marginal is a relative construct denoting deviation or distance from some notion of normality, center, or power.

Smooth as the related word margarine, and not at all filling space like a high school salutatorian speech.

The article's real gem, however, is brash acts of grammatical/statistical dissimulation worthy of an immediate letter e-mail to prospective students.  For example, this glorious page:

 

Note that the upper chart doesn't support his thesis in the least; though it uses averages, a black student with a 152 appears to have a greater chance at landing a merit scholarship than a white or Asian student with a 152.  Confronted with this hostile microaggression, Taylor ignores it and pulls a denominator switcheroo, rephrasing the issue as the chances of a random student from said race independent of LSAT score landing a scholarship - which is simply restating the LSAT disparity coupled with the idea that merit scholarships are yoked to LSAT scores.  This indispensable variable removed, the racism of the imaginary 148 LSAT black student not getting a scholarship compared to the imaginary 155 LSAT white peer becomes patently obvious.

Elsewhere, the piece lumps together people with 135 and 149 LSATs to make some claim that low-LSAT whites have a higher admission rate than low-LSAT black applicants without even looking at GPAs or other factors, and at other places it makes logical presumptions without appropriate citation or premises (e.g., "[a]pplicants  who  apply later in the cycle are often less likely to gain admission, due to fewer available seats  remaining  in  the  entering  class").

In statistics terms, it sucks, but in rhetorical terms, it's amazing, perfect not only for law school scholarship but also indicative of the sorta thing that would thrive on the Plaintiff's bar or in a BigLaw managing partners' meeting.

Being blatantly pro-LSAT, Inside Higher Ed, of course, gave the institutional test-makers a chance to defend their hot garbage exclusionary tool:

"The trends we have observed in applications from African American candidates that affect outcomes include the factors that Mr. Taylor identifies such as submitting applications late in the application cycle and LSAT scores that fall in lower bands. We see other factors as well such as [grades] and age at time of application."
...
"We believe that the overreliance that Aaron criticizes is driven by too much focus on U.S. News & World Report rankings and, in some cases, a failure on the part of schools to understand that even if they choose to focus on rankings, they have room to admit a wider range of LSAT scores."

Emphasis added.  What a snippy, spiteful bitch.  It wasn't enough to successfully rebuke the adequacy of Taylor's work in single paragraph; she just had to knock the entire industry's ability to understand how averages and medians work.

As a leader of the industry's preferred white-power testing mechanism, she should know that the willful mishandling of statistics to do whatever the fuck we want is de rigueur for this industry.  

Without statistical obfuscation, the Million Dollar Express just doesn't run.