Students for Fair Admissions v. Harvard: Statistics in the Courtroom

In November 2017, Students for Fair Admissions Inc. filed a complaint alleging that Harvard College violated Title VI of the Civil Rights Act, which prohibited racial discrimination in institutions that received federal funding. The complaint alleged that Harvard College engaged in intentional discrimination against Asian American applicants in its admissions process. Harvard College acknowledged its use of race in the admissions process but maintained that it was only one of many factors the school considered. It also claimed that using race as a “plus factor” was supported by the law. Both sides used a large set of data to plead their cases and both hired econometrics experts to argue their positions, who reached opposite conclusions. The trial judge needed to assess the two experts’ findings and reach a decision about the story the data was telling.

This case can be used in undergraduate and MBA business analytics courses to introduce common issues that may be encountered when conducting statistical modelling. It may also be relevant in organizational behaviour courses to discuss the “analytics” and statistical methods behind providing evidence of discrimination. Students should have already been introduced to multiple regression and should understand the meaning of p-values and coefficients. However, this case can be used effectively without requiring a thorough understanding of logistic regression. This case provides a high-profile example of the use of multiple regression in a legal setting and the analysis of discrimination. A data set is not provided for the case discussion. Students are instead expected to examine the modelling choices underlying two separate logistic regression models and consider the trade-offs and assumptions behind each modelling choice. Students are then asked to decide which model they think is stronger. In the process, they will learn about fundamental econometric modelling issues. Specifically, after working through the case, students will be able to<ul><li>explain the use of multiple regression in high profile legal cases;</li><li>understand why regression analysis necessitates an underlying theory and contextual understanding of the “data generating process,” illustrating the importance of theory-driven modelling choices in statistical analysis;</li><li>identify what an interaction variable is, and what trade-offs are associated with the use of an independent variable;</li><li>define the idea of “statistical power” and give examples of how different modelling assumptions may lead to different levels of statistical power;</li><li>explain how control variables can be used in a regression analysis, and why having more control variables is not necessarily better; and</li><li>identify methods to determine which modelling choices are critical to reach a statistical conclusion.</li></ul>

• Regression, Discrimination