Suboptimal social science and judicial precedent.

• Author: Grunwald, Ben K.

INTRODUCTION I. TYPOLOGY OF SOCIAL SCIENCE IN COURTS A. Adjudicative and Legislative Facts B. Social Authority C. Social Frameworks II. LIMITATIONS ON INDIVIDUAL SOCIAL SCIENCE STUDIES, LIMITATIONS ON HUMAN INTUITION A. Methodological Limitations 1. Imprecise Data 2. Nonrandom Sampling Designs 3. Unreported Data Mining 4. Correlation and Causation B. The Biases of Social Scientists Distort the Results of Their Research III. A BODY OF SOCIAL SCIENCE STUDIES AND THE DANGER OF OVERTURNING PRECEDENT A. Small Literatures 1. Overturning Precedent Due to Changes in Scientific Understanding 2. Overturning Precedent Due to Changes in Social Phenomena B. Inconsistent Literatures CONCLUSION INTRODUCTION

The social sciences have developed dramatically over the last century in both breadth and sophistication. (1) These disciplines offer systematic data collection and an analytic methodology to test our empirical intuitions about individual behavior and social institutions. Prior to the development of the social sciences and their application to the legal system, judges could rely only on their personal experiences and untested empirical intuitions when faced with complex questions of social fact. (2) A court's exclusive reliance on personal experience, however, "could continue only so long as its 'best guesses' about [empirical] facts were as good as ... everyone else's." (3) Today, social science research exists on a wide range of legal issues, and courts are faced with the challenge of resolving controversial questions of empirical fact on the basis of complex and sometimes conflicting scientific literatures. Courts have, for example, reviewed social science evidence on racial segregation, (4) maximum work hours, (5) First Amendment rights, (6) jury size] the exclusionary rule, (8) eyewitness identification, (9) and child custody. (10) Yet, despite efforts to encourage the integration of social science into the judicial process, and despite a modest increase in judicial reliance on social science evidence in recent decades, (11) courts remain reluctant to incorporate social science into their decisionmaking. (12)

In this Comment, I explore the role of social science in the development of common law precedent. I begin with the assumption that most judges and legal scholars today would support the use of social science in particular judicial decisions where the research findings are valid, replicated, consistent across studies, and directly applicable to the legal question at hand. I focus instead on the problem of suboptimal social science. In the vast majority of cases only suboptimal evidence is available--that is, evidence that is valuable but not completely valid, consistent, or directly applicable. Judges and legal scholars have long debated the benefits (13) and disadvantages (14) of using this kind of limited empirical research in judicial decisionmaking. And litigants frequently argue that courts should not rely upon specific social science evidence because of limitations in the literature. (15) Yet courts are given relatively little guidance on how to address suboptimal social science in the development of precedent. (16)

In Part I, I provide a brief history and typology of the uses of social science in the courts. I discuss the traditional distinction between adjudicative and legislative facts, (17) as well as other conceptual developments proposed more recently by Monahan and Walker. (18)

In Part II, I survey many of the common limitations of individual studies, which scholars and litigants frequently use to argue that a court should ignore a particular scientific study in developing common law precedent. These limitations include methodological weaknesses and potential biases of the researchers. Given that "courts often cannot avoid basing legal rules on empirical assumptions," (19) I observe that the negative consequences of these limitations will reappear regardless of whether courts choose to rely on social science to ground their decisions in particular cases. Without guidance from social science on complex empirical questions, judges must rely on their own empirical intuitions formed through a process subject to the same set of limitations. Thus, limitations on social science studies should not presumptively preclude their use in the development of judicial precedent.

In Part III, I consider limitations on the value of social science to judicial precedent that inhere in larger bodies of social science research rather than individual studies. I focus on two limiting characteristics in particular: the size of the literature and the consistency of the results. Scholars and litigants have used these characteristics to argue that a court should ignore a scientific literature in resolving a question of precedent. They have argued, for example, that subsequent research may overturn current scientific theory, and that courts are in equipped to assess social science research critically. I show that these limitations are often overstated and describe a number of common institutional mechanisms to address them. In addition, after noting the infeasibility of common solutions proposed by legal scholars, I suggest that existing permanent staff attorney positions in federal and state courts offer a unique and cost-effective method to bring social science expertise into the courthouse.

1. TYPOLOGY OF SOCIAL SCIENCE IN COURTS

   In this first Part, I outline a typology of the various purposes for which courts use social science research, and discuss the relevant procedures for each of these uses. As I note below, the well-established rules that govern the admission of evidence at trial do not appear to apply to social science used to resolve questions of legal precedent. For good reasons, courts have significant discretion in receiving and obtaining social science evidence for this narrow purpose.

   1. Adjudicative and Legislative Facts

      Historically, the use of social science in the development of common law precedent posed a procedural complexity in American courts. As a general matter, judges answer questions of law, and juries answer questions of fact. (20) Factfinders are charged with weighing the evidence (21) presented to the court. Appellate courts may review trial court findings, but only if such findings are "clearly erroneous." (22) This does not mean, however, that appellate courts may not review evidence on their own. Federal judges are permitted to obtain facts independently (without admission into the record) if they are "not subject to reasonable dispute." (23) These facts must be "generally known within the trial court's territorial jurisdiction" or "can be accurately and readily determined from sources whose accuracy cannot reasonably be questioned." (24)

      These general principles of legal practice were challenged in Muller v. Oregon, (25) which scholars often cite as the first use of social science evidence in an American court. (26) In Muller, the owner of a laundry company challenged the constitutionality of a statute prohibiting women from working over ten hours per day in certain industries. (27) As counsel for the State, Louis Brandeis submitted a brief that described medical and social science studies finding a negative effect of long working hours on women's health. (28) Though acknowledging that the underlying empirical studies were not, "technically speaking, authorities," (29) the Court found that the "widespread and long continued belief" that the studies represented was "worthy of consideration." (30)

      The use of social science evidence in Muller raised an evidentiary puzzle for legal practitioners and scholars: Brandeis submitted factual information to an appellate court through a brief. (31) How could Brandeis introduce the social science evidence when facts must be established at the trial court through direct and cross-examination of witnesses? (32) Kenneth Culp Davis provided an answer to this puzzle in a famous law review article that developed a typology of the different uses of social science in administrative and judicial decisionmaking. (33) Prior to the 1940s, legal authorities recognized two rough kinds of evidence: adjudicative facts and constitutional facts. (34)

      Parties introduce evidence of adjudicative facts--concerning the actions or circumstances of immediate relevance to the litigation (35)--through physical evidence or oral testimony to prove that some legal standard is or is not satisfied. The prosecution in a murder trial, for example, might introduce evidence that the defendant was apprehended with the blood of the victim on his hands. This adjudicative fact helps the jury to determine whether the defendant is guilty.

      Legal authorities also recognize a second category of evidence: evidence introduced to prove constitutional facts. (36) Constitutional facts, usually social or economic statistics, are used not to prove certain facts about the events in controversy, but to inform the development of a constitutional rule. (37) The Brandeis brief, for example, introduced evidence of facts to justify a constitutional rule that maximum-hour labor restrictions were permissible. (38)

      Davis argued that constitutional facts were "only one manifestation of a larger category of facts which are utilized for informing a court's legislative judgment on questions of law and policy." (39) Courts rely on facts from this broader category--which Davis called "legislative facts"--to shape not only constitutional rules, but also legal precedent more generally. (40) For example, though DNA identification is an accepted form of forensic evidence today, prosecutors were often required to introduce evidence of its reliability to appellate courts when the technology was first introduced. In United States v. Porter, the prosecution moved to submit the results of a DNA analysis on semen collected from the victim and to submit evidence that the probability of a false match was one in thirty...