Hidden Priors: Toward a Unifying Theory of Systemic Disparate Treatment Law

Abstract

Did the Court’s procedural decision in Wal-Mart Stores, Inc. v. Dukes undermine the substance of the systemic disparate treatment theory of employment discrimination? The answer to that question hinges on understanding the theoretical foundation for what one scholar calls the “most potent and least understood of the various Title VII causes of action.” The current scholarly efforts to understand systemic disparate treatment law can be sorted into two distinct strands – methodological and contextualist. Scholars in the methodological strand question whether statistical techniques currently used by courts are sufficient to support an inference of discrimination. In the contextualist strand, scholars urge a conceptual expansion of the systemic disparate treatment theory that would impose liability on employers for wrongdoing located at the organizational level, rather than simply aggregating individual-level claims. These two strands have advanced independently, with scholars in each strand often overlooking the implications of progression in the other. This Article is the first attempt to unify these two scholarly strands. It does so by exposing the inescapable role of hidden Bayesian priors – preconceptions about background rates of discrimination – in the interpretation of statistical evidence. Taking a Bayesian view, the shortcomings of traditional statistical evidence identified by methodologists are not fatal. Yet, the Bayesian view also provides the conceptual space needed for further development of the organizational approach advanced by contextualists. The Wal-Mart decision presents an opportunity to radically rethink this misunderstood area of antidiscrimination law, and this Article takes the first step in developing of a coherent theory of systemic disparate treatment that embraces Bayesian priors.