Joan C. Williams, Rachel M. Korn & Sky Mihaylo

Volume 72, Issue 1, 337-464

This Article joins other voices in challenging what I will call the “implicit bias consensus” in employment discrimination law, first crystallized in the work of Susan Sturm and Linda Hamilton Krieger. The implicit bias consensus has two basic components. The first is that most employment discrimination today is what Sturm christened “second generation employment discrimination” caused by implicit bias that is uncontrollable and unconscious, subtle and ambiguous. The second component of the consensus is that Title VII is ill-suited to address second generation discrimination.

This Article challenges the implicit bias consensus based on six different datasets from the Workplace Experiences Survey (WES), a simple ten-minute climate survey that provides a fine-grained description of how racial and gender bias play out in everyday workplace interactions. WES data and Williams’s other research offer plaintiffs’ lawyers a simple way to talk about racial and gender bias as falling into five basic patterns. WES data also helps them respond to the common defense argument that studies performed in social psychology labs do not describe what happens at work: the conjunction of lab studies and WES data is more powerful than either type of evidence alone, because lab studies provide objective evidence that the five basic patterns exist in the world, while WES data provide attitudinal evidence that these five patterns exist in today’s workplaces. The Article explores the implications of the WES and the five-patterns model for Title VII, contesting some basic tenets of the implicit bias consensus, notably its description of bias as unconscious and uncontrollable, and the contention that Title VII is ill-suited to address contemporary forms of discrimination. The Article ends by providing a step-by-step guide to how to use bias evidence in litigation, highlighting evidence that will be useful to help plaintiffs’ lawyers establish that a reasonable jury could find bias, particularly in light of the relaxed causation standard articulated in Bostock v. Clayton County.