1 Michael A. Rebell & Arthur R. Block, Educational Policy Making and the Courts: An Empirical Study of Judicial Activism (Chicago: University of Chicago Press, 1982) [hereinafter referred to as EPAC].Google Scholar

2 . 347 U.S. 483 (1954).Google Scholar

3 . Chayes, Abram, The Role of the Judge in Public Law Litigation, 89 Harv. L. Rev. 1281 (1976).Google Scholar

4 . Id. at 1282–84; Chayes, Abram, Foreword: Public Law Litigation and the Burger Court, The Supreme Court 1981 Term, 96 Haw. L. Rev. 4, 4–5 (1982).Google Scholar

5 . Frank, M. Johnson Jr., The Role of the Federal Courts in Institutional Litigation, 32 Ala. L. Rev. 271 (1981).Google Scholar

6 . Foreword, Symposium: Judicially Managed Institutional Reform, 32 Ala. L. Rev. 267 (1981).Google Scholar

7 Rebell and Block Indicate that they have also attempted to control for their own biases by using a rigorous methodological discipline and by creating a more extensive data base than ever before applied to the issues in question.Google Scholar

8 LEXIS is a computerized law research system containing the full text of opinions and dissents. Access to these opinions may be made by utilizing a word or phrase. The computer will then list all opinions within a selected set which contain the research word or phrase.Google Scholar

9 The authors report this result at 222. I attempted to replicate their results on two occasions. The following search in November 1982 produced 413 cases: [[PUBLIC SCHOOL or EDUCATION!] w/20 POLICY] and [[[EQUITABLE or INJUNCTIVE] w/20 RELIEF] or [INJUNCTION]] and date [aft 12/31/69 and bef 12/2/77]. The discrepancy here may have been due in large part to my inclusion of “injunction” as a search term. In July 1983, the following search produced 235 cases, but omitted 16 of those on Rebell and Block's list of 65: DATE AFT 12/31/69 (level one—84,900 cases) AND DATE BEF 12/1/77 (level two—41,625 cases) AND PUBLIC SCHOOL OR EDUCATION w/15 POLICY (level three—439 cases) AND EQUIT! OR INJUNCT! w/15 RELIEF (level four—222 cases). For purposes of replication it would have been useful for the authors to have included their specific LEXIS request. The difference in results and the authors' reference to a “carefully selected” sample raises the question of whether purposive selection occurred to obtain the original 236 cases. See infra note 10.Google Scholar

10 In a random sample all members of the population or universe have an equal chance of being selected. Put differently, all possible samples of 65 educational policy making cases would have the same probability of being selected. Cf. Frederick N. Kerlinger, Foundations of Behavioral Research: Educational, Psychological, and Sociological Inquiry 118 (2d ed. New York: Holt, Rinehart & Winston, 1973). If Rebell and Block are interested in generalizing to the population of educational policy decisions, ideally they would utilize random sampling of all such cases, making no systematic structural or subject matter exclusions. They then could be fairly sure that the cases they studied were representative of the larger body of educational cases The exclusions they list, however, suggest they may have drawn a purposive sample. Such sampling is “characterized by the use of judgment and a deliberate effort to obtain representative samples by including presumably typical areas or groups in the sample.” Kerlinger, supra, at 129. Purposive samples may be used to reflect quantitative or qualitative aspects of the data which the researchers consider important. Ole R. Holsti, Content Analysis, in 2 Gardner Lindzey & Elliot Aronson, eds., The Handbook of Social Psychology 596, 654 (2d ed. Reading, Mass.: Addison-Wesley Publishing Co., 1968). Generalizations from such samples should involve great caution. Cf. Kerlinger, supra, at 129 The researchers may also have studied all of the cases of a universe in which they were interested. In this situation their structural criteria and subject matter exclusions define the boundaries of the universe and they would not be justified in applying statements about their data to events outside the universe.Google Scholar

11 Cases involving segregation and aid to nonpublic schools were excluded because such suits “have been subjected to extensive doctrinal analysis and to some empirical study.” EPAC, supra note 1, at 221.Google Scholar

12 The study included only cases which involved a request for “an injunction that would replace, modify, or forbid enforcement of a systemwide educational policy.” Id. at 220. Rebell and Block explain their use of this standard as follows: The use of this criterion admittedly excludes from our sample what might be called individual “test cases” and “declaratory cases.” In the former, a plaintiff seeks only individual relief, but his suit may intend to establish (or may unintentionally establish) a legal precedent that will have a broad impact on school politics. Attempting to include such individual test cases in our sample would have been problematic, since the researcher could not ascertain reform objectives or results without detailed analysis of the background of the case and interviews of individuals involved. In the typical declaratory case, the plaintiff seeks a declaration of the general unlawfulness of a policy, but does not seek judicial orders to compel the defendant to change its rules and practices to conform to the declaration. Although in many such cases the plaintiff may have a reform objective, the declaratory complaint is a much weaker clue to this intent. For these reasons, we decided to concentrate the study on the more reliable sample—injunctive cases—leaving for future research the question as to whether there are any substantial differences in the school reform litigation process resulting from noninjunctive cases Id. at 220–21.Google Scholar

13 To be included a case had to involve a “[d]ismissal order, or liability determination (preliminary or final) made between January 1, 1970, and December 1, 1977.” Id. at 220.Google Scholar

14 Interviews were conducted with attorneys for the principal parties in 60 of the 65 cases. In all, 130 attorneys were interviewed. Ultimately, the case studies resulted in about 1,000 pages of data, which were subjected to analysis. Id. at 225.Google Scholar

15 The New York litigation, dealing with alleged racial discrimination in job selection and retention procedures for school supervisors, spanned eight years and involved 11 reported decisions at the district court level and four reported opinions at the appellate level. The authors provide the case history at 262 n.13 The Colorado litigation, involving the alleged need for bilingual-bicultural education in the Mesa County Valley School District, included two federal district court opinions and two Tenth Circuit opinions. The initial case was reported as Otero v. Mesa County Valley School Dist. No. 51,408 F. Supp. 162 (D. Colo. 1975). The authors provide the complete case history at 281 n.1. The authors conducted 15 interviews in relation to the New York litigation (id. at 233–34) and 11 interviews on the Colorado litigation (id. at 235–36). Data for the court cases also included pleadings, briefs, submissions, transcripts of trials and hearings, background reports, and newspaper accounts. Id. at 74. For the legislative case studies, the authors held 8 interviews in New York (id. at 234–35), 15 interviews in Colorado (id. at 236–37), and also drew on official committee records and attendance at, or tapes of, “all significant hearings and debates” (id. at 74).Google Scholar

16 See Otero v. Mesa County Valley School Dist. No. 51, 408 F. Supp. 162 (D. Colo. 1975).Google Scholar

17 A “coder” in this context refers to someone who has sorted the cases on a given set of categories, using predetermined defintions or criteria for sorting. To the degree that two or more coders agree in their sorting, the category defintions are said to be “reliable.” Although the authors state “that if one applied a formal statistical test to the quantitative comparisons that were specifically relied upon in the EPAC report, one would find that they satisfied the conventionally accepted confidence level of 95%” (id. at 226), there is no indication that this statement refers to comparisons between coders' responses.Google Scholar

18 Mutual exclusivity, or the extent to which a single case could fall into only one category; basis in a single scheme; and exhaustiveness, or the degree to which all the data can be sorted, will generally contribute to the reliability of the category sets. Cf. Ole R. Holsti, Content Analysis for the Social Sciences and Humanities 95–101, 135 (Reading, Mass.: Addison-Wesley Publishing Co., 1969).Google Scholar

19 The following definitions were used: Principle: A statement establishing a right of an individual against the state or against another individual (or, less frequently, the right of an institution to maintain the integrity of its legally defined prerogatives). A principle is expressed as a general rule that should be enforced whenever applicable, regardless of social welfare consequences, except when it is outweighed by a countervailing principle. Policy: A statement concerning collective goals. Policy arguments consider the relative importance or desirability of particular social goals, and/or the relative efficiency and desirability of particular method, for achieving such goals. A policy statement is normally expressed in more specific terms than is a principle, and in a particular context it may be subordinated to competing policy claims that are determined to be better able to serve collective goals more effectively. Rebell & Block, supra note 1, at 23–24.Google Scholar

20 One case, considered a two-issue case, was listed in two categories.Google Scholar

21 Chayes, supra note 3, at 1316.Google Scholar

22 See Douglas Hofstadter, Gödel, Escher, Bach: An Eternal Golden Braid 10 (New York: Basic Books, 1979).Google Scholar