How Economists Entered the ‘Numbers Game’: Measuring Discrimination in the Us Courtrooms, 1971–1989

Abstract

The paper explores why and how economists entered the courtrooms as expert witnesses in employment discrimination cases in the US. The main sources are published legal decisions. I analyze the courts’ and economists’ discourses on the use of a specific method: multiple regression analysis in relation to litigation history, academic debates, and the institutional settings of expertise within the courts. I first show how the early reception of the method in the late 1970s did not involve systematic rejection from the courts but rather a large amount of skepticism. I then illustrate how economic theory underlying the method was progressively introduced in the “judicial tool kit” and how the debates in the courtrooms relate to the debates in academia in the 1980s. By 1989, practical and ethical questions regarding the institutional settings of experts’ testimony took center stage, reflecting the increasing professionalization of forensic economics.

Modeling intervention: Barbara Bergmann’s micro-to-macro simulation projects

“…Bergmann spent a large amount of her time fighting a discrimination case against her own employer in a promotion dispute as well as testifying in other discrimination cases (Small 2022;Chassonnery-Zaïgouche 2020). As a woman and a Jew, she always insisted she was not welcomed in the profession in the early day of her career and never felt "plug-in" the elite of the discipline, a sentiment reflected in her somewhat ambivalent position: recognized and celebrated for some of her works in feminist economics (see Strober 1998 and the whole issue in Feminist economics), ignored when talking about anything else, and in particular macroeconomics, she was rather isolated her entire career.…”

Application of Artificial Intelligence in Digital Games Based on Mathematical Statistics

“…The training process is relatively long. If a long training process is required for each startup, the ease of use of the system will be greatly reduced [22]. In order to avoid repeated training, it is necessary to save the training results locally, so that the text service to be implemented can be quickly loaded when it is just started.…”