SOCIAL SCIENCE IN CONSTITUTIONAL LITIGATION

All litigation, including constitutional litigation, resolves issues of law and fact. social science research can help to clarify the facts on which a case may turn; and it can help the resolution of legal issues by laying before the courts data and analyses that bear on the choice of an appropriate legal rule.

Legal lore has it that the rise of social science in the law began with the brandeis brief, in which louis d. brandeis, special counsel for the state of Oregon, successfully bolstered the state's claim in muller v. oregon (1908) that its statute limiting the working hours for women was constitutional. Although in theory the state merely had to show that such a regulation was not unreasonable, previous decisions had struck down laws regulating working hours of other employees as unreasonable invasions of the liberty of contract. The brief supported the reasonableness of the law in part by showing that a great many American states and even more countries abroad had similar statutes. It was an effective if modest social science effort.

More sophisticated techniques are to be found in contemporary constitutional litigation. Sampling, the most powerful tool of social science research, is now firmly established as an appropriate means of gathering evidence. If the survey was conducted without bias and if the technical requirements are met, a sample may be accepted as a reasonably accurate representation of the sampled universe. For instance, in support of a motion for change of venue in a criminal case, a sample survey measures the extent and depth of pretrial prejudice in the community. If a voluminous body of communications is at issue, sampling may be combined with a technique called content analysis. Thus, when the constitutionality of the work of the House Committee on Un-American Activities was litigated, a sample of the committee's public hearings was examined. This approach yielded a numerical statement of the frequency with which the committee asked its witnesses questions that transcended its constitutional authority. Similar content analysis has sometimes been used in support of a motion for change of venue, documenting the charge that a substantial part of the pretrial publicity originated in the prosecutor's office.

Proof of racial or other discrimination in jury selection, employment, and other contexts frequently employs sampling and subsequent statistical analysis. Such proof involves an analysis of the differences between the actual outcome of the selection process and the outcome that would have been expected if discrimination had no role in the process.

In United States v. Hazelwood School District (1977), for instance, the Supreme Court made its own probability computations to determine whether excluding the metropolitan area from the labor market in which a suburban school district hired its teachers would substantially weaken the government's statistical proof that the district had engaged in discrimination. Although the Court's statistical performance in Hazelwood was flawed in certain respects, similar methods in proving discrimination have become accepted in both federal and state courts.

Of particular interest are the cases in which the judicial system itself is charged with discrimination. The two main targets here are the administration of the death penalty and the selection of jurors. Evidence has been mounting, and finally has drawn the attention of the Supreme Court, that the death penalty is administered with bias, discriminating against black offenders who killed white victims. The major technical problem in distilling this evidence is to assure comparability of the homicides under analysis.

In the jury selection area, the statistical analysis of discrimination has had more impact. Despite substantial efforts in this direction, the lower courts have rejected these efforts. In Castandeda v. Partida (1977), for instance, the Court used a standard statistical formula to compute the probability that the disparity between the proportion of Mexican Americans serving on grand juries and their proportion in the county population could have arisen if grand jurors had been selected at random. The majority found the probability to be so minute (about one in a number with 140 zeros) that the discrepancy was sufficient to establish discrimination even though there were problems with the data used to estimate these proportions and even though the majority of jury commissioners were themselves Mexican Americans.

In the trial of Dr. Benjamin Spock and others accused of conspiring to obstruct the draft, the alleged discrimination involved female jurors. The allegation of bias in that case was directed not against the system but against the particular judge who consistently selected juries with significantly fewer women than those of his colleagues, although all drew from the same pool of potential jurors.

At times experimental social science research is offered to aid a court in assessing the consequences of its legal options or in ascertaining facts relevant to the choice of these options. When the Supreme Court in brown v. board of education (1954) held that segregated education was inherently unequal, the Court quoted with approval a lower court's finding that school segregation with the sanction of law produced feelings of inferiority among black children, affecting their motivation to learn. The Court remarked that its conclusion was "amply supported by modern authority." That authority, cited in a footnote, consisted of seven items. Five, such as Gunnar Myrdal's American Dilemma, dealt generally with problems of black education. Two bore more directly on the issue: a statement by thirty-two leading social scientists and an experiment conducted by the psychologist Kenneth Clark. Clark had given sixteen black children in a South Carolina elementary school a sheet of paper on which two dolls were drawn, identical in every respect except that the one was black, the other white. The children were asked, "Which doll would you like to play with?" "Which is the nice and which the bad doll?" "Which doll looks like you yourself?" Ten of the children liked the white doll best; eleven called the black doll the "bad" one; seven of the black children, when asked which doll was like themselves, picked the white one. From these answers and earlier research, Clark concluded "that these children … like other human beings who are subjected to an obviously inferior status in the society in which they live, have been definitely harmed in the development of their personalities.…"

Later, scholars disputed both the evidentiary power of that study and the weight the Justices had attached to it. The study, obviously limited in size and structure, today would hardly survive cross-examination. Most likely its major function was to buttress a position the Justices had reached on their own.

Social science research has provided more solid evidence in litigation over the constitutionality of juries with fewer than twelve members. In the two decisions that affirmed the legality of such juries, the Court cited a number of empirical studies purporting to show that these modifications did not affect the quality of the verdicts rendered by the smaller juries. Subsequently these studies were severely criticized, and five years later ballew v. georgia held five-member criminal juries unconstitutional. Justice harry a. blackmun's opinion repeatedly cited these critical views.

Most social science operations suffer from some imperfection, partly because their subject matter is so complex, and partly because of methodological flaws. Even if such imperfections are minor, courts may hesitate to accept social science findings that threaten to dislodge established rules. One type of effort to compensate for imperfection is "triangulation"—the confluence of evidence from independent studies that approach the same problem from different angles. An example is the series of studies of "death qualified" juries.

At one time, a New York statute allowed New York City to try murder and other crimes of public notoriety before specially selected blue ribbon juries, whose members, among other qualifications, were required to have no objection to the death penalty. When the Court was asked to declare these juries unconstitutional because of alleged bias in favor of the prosecution, it declined by a bare majority on the ground that there was no proof of such bias. Speculation as to how such proof might be established led to the first study which found that jurors who were in favor of the death penalty were indeed more likely to convict, not only in capital trials, but generally. Six other studies followed, with different approaches; each replicated the result.

Witherspoon v. Illinois (1968), decided halfway through these studies, did not reach the issue. Although the Court agreed that merely having scruples about the death penalty was not sufficient cause for eliminating jurors, it dismissed the first few research findings, indicating that the exclusion of jurors with scruples against the death penalty would bias the jury in favor of conviction, as "too tentative and fragmentary." Subsequent efforts to convince other courts that the post-Witherspoon juries, too, were biased in favor of convicting defendants failed until 1983 and 1984 when two federal district courts in habeas corpus proceedings accepted the evidence provided in these studies and invalidated the convictions. Although the federal Courts of Appeals have divided on this issue and the Supreme Court has agreed to review one of these cases, these two decisions mark a preliminary acceptance of proof by triangulation.

The role of social science research in litigation is bound to grow in spite of deep-seated hesitancy on the part of the courts to look at statistical evidence. It is difficult to predict how fast and where the use of social science techniques will increase in constitutional litigation. Much will depend on the resourcefulness of social scientists in developing new research and the initiative of attorneys in presenting evidence that can sharpen the perception of litigated facts and aid courts in judging the consequences of their legal options.

Hans Zeisel

David Kaye
(1986)

Bibliography

Baldus, D. and Cole, J. 1980 Statistical Proof of Discrimination. Colorado Springs, Colo.: Shepard's.

Cahn, Edmond 1962 A Dangerous Myth in the School Segregation Cases. In Kenneth Clark, ed., Confronting Injustice. Pages 329–345. Boston: Little, Brown.

Kaye, David 1980 And Then There Were Twelve: Statistical Reasoning, the Supreme Court, and the Size of the Jury. California Law Review 68:1004–1043.

——1982 Statistical Evidence of Discrimination. Journal of the American Statistical Association 77:773–783.

Lempert, Richard O. 1975 Uncovering "Nondiscernible" Differences: Empirical Research and the Jury-Size Cases. Michigan Law Review 73:643–708.

Loh, Wallace D. 1984 Social Research in the Judicial Process: Cases, Readings and Text. New York: Russell Sage Foundation.

Saks, Michael J. 1974 Ignorance of Science Is No Excuse. Trial 10:18–20.

Walbert, David 1971 The Effect of Jury Size on the Probability of Conviction: An Evaluation of Williams v. Florida. Case Western Law Review 22:529–554.

Zeisel, Hans 1968 Some Data on Juror Attitudes Towards Capital Punishment. Chicago: University of Chicago Center for Studies in Criminal Justice.

——1971 And Then There Were None: The Diminution of the Federal Jury. University of Chicago Law Review 38: 710–724.

——1980 Reflections on Experimental Techniques in the Law. Journal of Legal Studies 2:107–124.

——1985 Say It with Figures, 6th ed. Chap. 14. New York: Harper & Row.

——and Diamond, Shari 1974 "Convincing Empirical Evidence" on the Six-Member Jury. University of Chicago Law Review 41:281–295.