Title VII Class Actions and the Disaggregation Defense: Bad Law and Bad Statistics

Abstract

When statistical expertise is offered in a litigation context, the result can be bad law and bad statistics. In this paper, I analyze how statistical evidence is used in the class certification phase of employment discrimination litigation. I analyze how the institutional context of litigation determines how statistical evidence is presented by plaintiffs and defendants and evaluated by the courts. I show that under current case law, defendants can and do introduce statistical evidence that has absolutely no relevance on the issue before the Court, yet is often persuasive in defeating class certification. I then show how both weak and strong statistical power is exploited by experts on each sides of the litigation. Finally I outline a new approach for framing issues of statistical inference and hypothesis testing in a way that more effectively addresses the issue of whether or not systematic bias exists throughout an organization.