The assumption that human behavior is largely under conscious control has taken a theoretical battering in recent years. Although this assault in some ways resembles the previous century's Freudian revolution, there are important differences between the two. Freud's views of unconscious mechanisms were embedded in a theory that never achieved conclusive support among scientists, despite many empirical theory-testing efforts in the middle third of the twentieth century.' Consequently, most psychologists have abandoned Freud's psychoanalytic theory of unconscious mental processes. Theoretical conceptions of conscious control over human behavior were strongly re-established in the last third of the twentieth century, but the dominance of such views has been crumbling during the past two decades. Unlike the Freudian revolution, however, the new science of

Although there is now a considerable literature on how law and legal principles serve as institutionalized models for organizations, there has been far less attention to how organizational practices may serve as institutionalized models for courts. This article offers a theoretical and empirical analysis of legal endogeneity -a subtle yet powerful process through which institutionalized organizational structures and practices influence judicial conceptions of legality and compliance with antidiscrimination law. We argue that, irrespective of their effectiveness, organizational structures, such as grievance procedures, anti-harassment policies, evaluation procedures, and formal hiring procedures, become symbolic indicators of compliance with anti-discrimination laws, first within organizations, but eventually in the judicial realm as well. As organizational structures become increasingly institutionalized, lawyers and judges become more likely to associate them with rationality and fairness and to infer nondiscrimination from the mere presence of those structures. Legal endogeneity has observable manifestations: judges increasingly refer to organizational structures in their opinions, find them relevant to determinations of legal liability, and ultimately defer to those structures by inferring nondiscrimination from their presence. We test legal endogeneity theory by analyzing a random sample of 1024 federal employment discrimination decisions from 1965-1999. We find that legal endogeneity has increased over time; that judicial deference is most likely when plaintiffs lack social and economic clout; and that judicial deference is most likely when the legal theories require judges to rule on organizational attributes that are not directly observable. We suggest that legal endogeneity weakens the impact of law as judges come to view organizational structures as indicators of legal compliance even when those structures are ineffective in combating discrimination.2

Title VII's disparate treatment model of discrimination is premised on the notion that intergroup bias is motivational in origin. This premise, in turn, is based on a number of assumptions regarding the nature of human inference and the respective roles played by cognition and motivation in social judgment and decisionmaking. Applying insights from cognitive psychology, Professor Krieger examines the assumptions about human inference embedded in current disparate treatment theory and questions the premise that discrimination necessarily manifests intent or motive. She suggests that a large number of biased employment decisions result not from discriminatory motivation, as current legal models presume, but from a variety of unintentional categorizationrelated judgment errors characterizing normal human cognitive functioning. Because of the lack offit between the present disparate treatment model and the phenomenon it purports to represent, courts and litigants are presented with a confising array of increasingly ill-defined and questionably premised analytical paradigms. Worse, as currently constructed, it may be exacerbating intergroup tensions and inflating both social and financial adjudication costs. Searching for solutions Professor Krieger explores the legal and policy implications of a cognitive process approach to discrimination and equal employment opportunity and evaluates a variety of modifications to existing equal employment opportunity law. 5. Donohue & Siegelman, supra note 3, at 1019 (citing American Bar Foundation Employment Discrimination Litigation Survey, supra note 3). 6. 411 U.S. 792 (1973). 7. See text accompanying notes 56-67 infra for a discussion of the McDonnell Douglas model. 8. Existing disparate treatment jurisprudence in many ways equates a finding of pretext in plaintiff's favor with a finding that the employer has lied about the reasons for its decision.

A rich theoretical literature describes the disadvantages facing plaintiffs who suffer multiple, or intersecting, axes of discrimination. This article extends extant literature by distinguishing two forms of intersectionality: demographic intersectionality, in which overlapping demographic characteristics produce disadvantages that are more than the sum of their parts, and claim intersectionality, in which plaintiffs who allege discrimination on the basis of intersecting ascriptive characteristics (e.g., race and sex) are unlikely to win their cases. To date, there has been virtually no empirical research on the effects of either type of intersectionality on litigation outcomes. This article addresses that lacuna with an empirical analysis of a representative sample of judicial opinions in equal employment opportunity (EEO) cases in the U.S. federal courts from 1965 through 1999. Using generalized ordered logistic regression and controlling for numerous variables, we find that both intersectional demographic characteristics and legal claims are associated with dramatically reduced odds of plaintiff victory. Strikingly, plaintiffs who make intersectional claims are only half as likely to win their cases as plaintiffs who allege a single basis of discrimination. Our findings support and elaborate predictions about the sociolegal effects of intersectionality.

The first call of a theory of law is that it should fit the facts."-Oliver Wendell Holmes' INTRODUCING BEHAVIORAL REALISM Although they serve different social functions and employ different methods and tools, both law and the empirical social sciences need, use, and produce theories of human behavior. But their respective relationships to these theories differ in significant ways, and for this reason, law and social science often stand in tension with each other when they meet in the courtroom or the case reporter. For its part, law needs, uses, and produces theories of human behavior when judges elaborate constitutional or common law doctrines or interpret ambiguous statutory provisions that implicate human motivation, subjective experience, or choice. Legal actors (judges, jurors, administrative fact finders, dispute handlers, and disputants) also use behavioral theories when they evaluate, litigate, or adjudicate specific disputes, as they attempt, for example, to attribute causation, assess witness credibility, or determine