Discrimination by comparison.

• Author: Goldberg, Suzanne B.

ARTICLE CONTENTS INTRODUCTION I. THE EMERGENCE AND INSTANTIATION OF COMPARATORS IN DISCRIMINATION LAW A. Comparators as the Default Methodology for Observing Discrimination B. Comparators as a Defining Element of Discrimination Law II. THE COMPARATOR DEMAND AS A BARRIER TO DISCRIMINATION CLAIMS A. The Comparator Default and First-Generation Cases 1. No Sufficiently Comparable Coworkers 2. Small Sample Size 3. Uniquely Situated Employees 4. Homogeneous Workplaces 5. Pregnancy, Breastfeeding, and the Nonexistent Comparator B. The Comparator Heuristic's Flaws as Amplified in Second-Generation Cases 1. Intersectionality 2. Identity Performance 3. Structural Discrimination III. ON THE CONCEPTUAL LIMITATIONS OF COMPARATORS A. Comparators as Overinclusive B. Comparators as Underinclusive C. Comparison and Disparate Impact IV. CONTEXT: A METHODOLOGICAL ALTERNATIVE A. The Emergence of the Contextual Model in Stereotyping and Harassment Jurisprudence B. Acts, Statements, and Automaticity C. Reconsidering Comparators in Light of the Contextually Focused Stereotyping and Harassment Jurisprudence V. JUDICIAL LEGITIMACY AND THE STICKING POWER OF COMPARATORS A. The Legitimacy Concerns at Play B. The Comparator Heuristic's Legitimizing Work C. The Call for Experts as a Response to Legitimacy Concerns D. Legitimacy Concerns and the Viability of Second-Generation Discrimination Theories VI. PROSPECTS FOR CHANGE A. Involving Experts in Setting Comparators' Contours B. Considering Hypothetical Comparators C. Moving Beyond Comparators CONCLUSION INTRODUCTION

Contemporary discrimination law is in the midst of a crisis of methodological and conceptual dimensions. The underlying problem is that evaluating allegations of discrimination requires courts and others to see something that is not observable directly: whether an accused discriminator has acted because of a protected characteristic. While this challenge has long been with us, as putative discriminators rarely admit discriminatory intent, (1) the crisis arises because the most traditional and widely used heuristic--comparators, who are similar to the complainant in all respects but for the protected characteristic--is barely functional in today's economy and is largely unresponsive to updated understandings of discrimination.

Some decades ago, when identity-based differentiation was relatively open and notorious and when many workplaces were of a Tayloresque scale with easily comparable jobs, (2) individuals claiming discrimination could often point to counterparts who were treated better. Courts could then deduce, with some confidence, that the protected trait was the reason for the adverse treatment at issue. (3) But in a mobile, knowledge-based economy, actual comparators are hard to come by, even for run-of-the-mill discrimination claims. (4) For the complex forms of discrimination made legible by second-generation theories, the difficulties in locating a comparator amplify exponentially. (5)

This methodological problem has spilled over, conceptually, to constrict the very idea of discrimination. Consider Justice Thomas's statement that a finding of discrimination cannot be made without "a comparison of otherwise similarly situated persons who are in different groups by reason of certain characteristics provided by statute." (6) Justice Kennedy has observed similarly that "one who alleges discrimination must show that she 'received differential treatment vis-a-vis members of a different group on the basis of a statutorily described characteristic.'" (7) On this view, however abusively an employer treats its employees, the bad acts do not present a discrimination problem so long as they are committed in an evenhanded fashion. (8) Their position, in essence, is that discrimination laws and norms do not impose obligations with meaningful abstract value.

Yet this position foreshortens traditional understandings of discrimination even within the Supreme Court's own jurisprudence. As the case law that addresses harassment and stereotyping makes clear, objectionable trait-based acts and statements occur in the absence of a comparator. (9) Indeed, in a well-known stereotyping case, the Court acknowledged the lower court's finding that no comparators existed, yet still found that the plaintiff, Ann Hopkins, was discriminatorily denied partnership at her accounting firm. (10) Likewise, in a much-discussed harassment case, the Court unanimously recognized that discrimination, in the form of sexual harassment, could occur in a work environment where only men were present. (11) At the same time, the Court has acknowledged that the presence of a better-treated comparator does not transform permissible acts into unlawful ones. "Treating seemingly similarly situated individuals differently in the employment context is par for the course," Chief Justice Roberts recently wrote. (12)

Still, the scope of discrimination law continues to shrink. (13) The judicial demand for comparators continues largely unabated outside the harassment and stereotyping contexts, (14) sharply narrowing both the possibility of success for individual litigants (15) and, more generally, the very meaning of discrimination.

In this Article, I argue that we are seeing the transformation of a heuristic device (16) for observing discrimination into a defining element of discrimination and that this collapse presents two serious problems. First, methodologically, comparators' deficiencies have come to outweigh their strengths as devices for discerning discrimination. Specifically, the demand for similarly situated, better-treated others underinclusively misses important forms of discrimination and forecloses many individuals from having even an opportunity to be heard because sufficiently close comparators so rarely exist. (17)

The second problem is conceptual. Since the early 1990s, much of the theoretical work on discrimination has attempted to make legible the many ways in which discrimination occurs beyond the forms of easily recognizable, deliberate exclusion that are based on relatively thin conceptions of protected traits. (18) Yet when comparators are treated as definitional, these theories cannot gain jurisprudential traction because the problems they identify cannot, in effect, be seen by courts.

This Article considers three of the leading theories. (19) The first is intersectionality theory, which recognizes that although the law designates trait-based protections sequentially, employers and others often target individuals because of their identity as a whole, rather than because of individual traits in isolation from one another. (20) In these situations, an employee, such as a black woman or a disabled older man, claims to have experienced discrimination based on a combination of legally protected traits. He or she struggles under a comparator regime in part because it can be difficult to decide who is the proper comparator--is it someone who shares neither of the individual's traits or shares one but not the other? In addition, because intersectional plaintiffs are often few in number relative to all others in a workplace, decisionmakers tend to be skeptical of the comparison's probative value and are typically unwilling to conclude that comparatively worse treatment is attributable to discriminatory intent rather than to the plaintiff's idiosyncratic quirks.

The second theory is identity performance, which conceives of identity traits in a thick way, recognizing that individuals sometimes experience discrimination because of stereotypes about behaviors or personal styles associated with their identity group rather than because of their phenotype. When operationalized, the theory produces cases in which employees and others seek to show that they have suffered trait-based discrimination because they have, for example, a Spanish-inflected accent or a traditionally African hairstyle. (21) Yet a comparator-based approach misses identity-performance theory's point in all but the most limited circumstances. For example, we might imagine an employer refusing to promote one Latino but promoting several others and arguing that it was not ethnicity but personal style (that is, too much Spanish-speaking or too thick an accent) that led to the promotion denial. Unless there is a non-Latino comparator who speaks the same amount of Spanish or has the same accent, the claim will not be legible in an analytic regime that recognizes discrimination only in the presence of a better-treated counterpart. (22)

The third is structural discrimination theory, which focuses on the ways in which the structures and dynamics of workplaces and other environments can effectuate--and obscure--discriminatory intent. Central to this theory are the "patterns of interaction among groups within the workplace that, over time, exclude nondominant groups" based on protected traits but are "difficult to trace directly to intentional, discrete actions of particular actors." (23) Comparators, even if they exist, are unlikely to shed light on the identity traits that motivated the exclusionary interaction patterns in all but the most blatant situations. The judicial insistence on comparators thus renders imperceptible the link between the protected trait and the reduction in opportunities or increase in adverse treatment.

Stepping back, we see that the comparator methodology has left these theories virtually noncognizable in the adjudication context and, by doing so, has depleted antidiscrimination norms of much of their content. Put another way, the synergistic relationship between the law's production of observational tools and those tools' production of law has put comparators in a position to shape and limit what courts can see as discriminatory.

Several payoffs follow from this clarified picture of the comparator methodology's consequences. For one, by putting into stark relief how little work discrimination law is doing in...