Analysis of Differential Prediction of Law School Performance by Gender Subgroups (TR-98-03)
by Vincent F. Harris, Louis A. Roussos, and Peter J. Pashley

Executive Summary

In the law school admission process, it is essential that the criteria used for admission are fair to all subgroups in the applicant population. One method used to evaluate the fairness of the admission process is to compare the predicted and actual first-year grade-point averages (FYA) for each law school for various subgroups of the applicant population. The purpose of the current study is to address questions of differential prediction between male and female first-year law school students based on data from the 1993, 1994, and 1995 first-year classes of 159 law schools.

Least-squares regression analyses were used to predict first-year average using three traditional predictors: undergraduate grade-point average (UGPA), scale score on the Law School Admission Test (LSAT), and the best predictive linear combination of UGPA and LSAT score. A separate analysis was conducted for each law school included in the study.

The results indicate that a regression model with both LSAT and UGPA as predictor variables performed the best. About 86.8% of schools had predicted minus actual FYA means for females that were less than one point (i.e., a tenth of a standard deviation on the FYA scale) from 0, while about 95.6% of the schools had predicted minus actual FYA means for males that fell within the same range. Only one school experienced a predicted minus actual FYA mean of more than two points. When UGPA was used as the predictor variable, systematic underprediction occurred for males and systematic overprediction occurred for females. In addition, the predicted minus actual FYA mean fell in the range of 1.0 to 1.0 at only 67.9% of the schools for females and 80.5% of the schools for males. When LSAT alone was used, the results were better than when UGPA alone was used, but not as good as when LSAT and UGPA were used in combination. With the model using LSAT alone, 94.3% of the schools had predicted minus actual FYA means for males within a point of zero. For female test takers, 83.0% of the schools had predicted minus actual FYA means that were within a point of zero. For both male and female test takers, the predicted minus actual FYA means were less than two points from zero for all but one school. Overall, the results of this study do not support the concern that the use of LSAT scores or the traditional combination of LSAT scores and UGPAs result in unfair admission decisions with regard to gender.

While considering the results of this study, the reader should keep in mind that the results refer only to subgroup behavior and not to individuals. For example, while results may suggest that UGPAs alone may overpredict FYAs for female law students on average, the performance of many individual female law students will be underpredicted based solely on their UGPAs.


Analysis of Differential Prediction of Law School Performance by Gender Subgroups (TR-98-03)

Research Report Index