Alexia Brunet Marks and Scott Moss (my former Marquette colleague) have an interesting paper on SSRN this week that was also profiled on the WSJ Law Blog. "What Makes a Law Student Succeed or Fail? A Longitudinal Study Correlating Law Student Applicant Data and Law School Outcomes" should be of interest to anyone either applying to law school or admitting students to law school. The two authors collected seven or eight years of applicant data from the University of Colorado Law School and Case Western Reserve Law School for 1400 enrolled students and matched the applicant data with resulting law school grades to see which application factors had the most predictive power for law school grades. I was on an admissions committee with Scott Moss for two years, so I was very interested to see whether this study confirmed or disproved some of our heuristics.
Table 3 sums up the findings. "Positive Predictors" are LSAT, UGPA, LCM (LSAT College Mean), STEM or EAF (economics, accounting or finance) major, post-college career last 4-9 years, being a teacher, and a rising UGPA if the UGPA is not old. Though the abstract states that the LSAT underperforms compared to conventional wisdom, I think the actual findings resemble what most of us thought about the LSAT: it predicts first-year grades, but not necessarily cumulative grades. UGPA does correlate with long-term grades, but it seems like only somewhat better than the LSAT. This is surprising to me only because I tend to discount UGPA in the era of grade inflation. The combination of high LSAT/low UGPA has a negative correlation with grades, confirming a gut feeling I have been spouting off for years. The variable of have a teaching career being positively correlated with law school grades is intriguing, though it seems to match my experiences with the very small number of ex-teachers I have taught.
I will let others pore over the statistical findings. The authors do a good job of describing the limitations of their data -- grades, not job placement or satisfaction, are used as a proxy for law school "success." Only matriculants are in the pool, so these are students who may have been chosen despite low UGPAs or LSATs because of other qualities that may not show up in the data -- in other words the pool is selected to succeed. And of course, the data cannot code for personal qualities such as ambition and drive.
What is an interesting thought experiment is whether law schools would have changed any admissions practices if Marks and Moss had proven zero correlation or even a negative correlation between LSAT and law school grades. Given the oppression of the USNWR rankings, which have worked to put undue emphasis on LSAT scores, then other things would have to change before law schools could throw out the LSAT (including changes at the ABA).
As an aside, having been in admissions meetings with Scott, the most interesting finding is that a disciplinary or criminal record has a negative predictive value equivalent to over a 7 point drop in LSAT. This is fascinating to me because the applicants who are admitted with disciplinary or criminal records generally are admitted because the infraction is minor, isolated or both. In other words, most of those applicants are let in under the assumption that their records do not reflect any cause for concern. Apparently, admissions committees should be paying more attention to random minor-in-possession records than we thought! Yikes!