A THEORY OF INTRACOURT INFLUENCE Thesis for the Degree of Ph. D. MICHEGAN STATE UNWERSITY GREGORY JAMES, RATHJEN 1972 This is to certify that the thesis entitled A THEORY OF INTRACOUR’I‘ INFLUENCE presented by GREGORY JAMES RATHJEN has been accepted towards fulfillment of the requirements for th)- degree in Boliiical—Science Date August 4, 1972 1 ”NM" 3 "ME & 30” 300K BINUERY INC. M L'UPARY n'woms “RIIGPONJIM! til») Wfiéa') ABSTRAC T A THEORY OF INTRACOURT INFLUENCE By Gregory James Rathjen The purpose of this study is to formulate and apply a theory of intracourt influence on the United States Supreme Court. The basic assumption of the theory is that Supreme Court justices are rational political actors and that each justice is motivated in his decision making by the desire to have Court-made public policy approximate as closely as possible his personal policy preferences. Since the rule structures governing collegial decision making on the Supreme Court severely circumscribe the extent to which all justices are able to have their personal policy preferences reflected in the Court-made public policy across all issue areas, the theory assumes that each justice attempts to influence the decisions of his colleagues regarding public policy. From these basic assumptions the theory posits specific observable (overt) behaviors which are held to be influence strategies employable by the individual justice, Gregory James Rathj en and specifies the circumstances, frequency and intensity of their use. The observable behaviors that are considered influence behaviors are analogous to those suggested by A. O. Hirschman in his book entitled Exit, Voice, and Loyalty (Cambridge: Harvard University Press, 1970). Hirschman contends that in the economic setting, an individual consumer has at his disposal two alternative means to influence a producer with regard to the quality of the product he produces. These are exit and/or "voice. " Hirschman specifies the relationship between these two influence strategies; when and under what circumstances the consumer will use either option, separately or jointly. Hirschman' 3 basic contention is that the use of these two influence strategies is directly related to the extent of quality deterioration perceived by the consumer with regard to the particular product of his concern. Though not completely isomorphic to the Court, Hirsch- man' 3 basic paradigm is considered sufficiently analogous to be appropriately applied in the context of the Supreme Court influence process. On the Court, exit is the act of leaving the majority. Since the justices in the majority determine what is to be Court- made public policy, the act of leaving the majority indicates that an individual justice considers that Court-made public policy Gregory James Rathj en unsatisfactory. "Voice" is the act of expressing one' s dissatisfaction directly through separate opinion writing. The use of either exit or "voice, " or both, is held to be a function of the degree of dif- ference between Court-made public policy and personal policy preference perceived by the individual justice. It was hypothesized that the frequency and intensity (extremity) of "voice" employed by a justice as a member of the majority (before exit) would increase as the difference between Court -made public policy and personal policy preference increased up to the point where exit occurred. It was also hypothesized that "voice" after exit would be greater than "voice" before exit. For after exit "voice, " it was hypothesized that, if the difference between Court -made public policy and personal policy preference increased, the frequency and intensity of "voice" would decrease. The data set used in this study consists of a nonrandomly selected set of 22 category scales consisting of 433 cases decided by the Supreme Court during the 1958- 1969 terms of Earl Warren' 3 Chief Justiceship. These category scales, or cumulative scales, serve as the Operational indicator of the difference between Court- made public policy and personal policy preference. When the hypotheses were tested it was found that "voice" did, indeed, increase approaching exit (both in terms of frequency and intensity) Gregory James Rathjen in the aggregate and when controlling for issue area, time period, and subgroupings within time period. It was also found that "voice" after exit was substantially greater than "voice" before exit. The after exit "voice" hypothesis was found to be unsupported in the aggregate and when controlling for issue area, time periOd, and subgroupings. An alternative explanation for the failure of the after exit "voice" hypothesis is provided. A THEORY OF INTRACOURT INFLUENCE By Gregory James Rathjen A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Political Science 1972 ACKNOWLEDGMENTS Several people have been instrumental in helping me bring this dissertation to completion. To thank them all would be virtually impossible. A special word of acknowledgment, though, must go to Harold J. Spaeth, whose help and guidance over the course of my graduate career has been invaluable. As chairman of my disser- tation committee he quickly and thoroughly responded to the many drafts and re-drafts, responses well deserving of praise and many thanks. Also, to the other members of my committee, David Rohde and Joseph Schlesinger, I express my deepest appreciation for their help and guidance throughout the past year of dissertation writing. Finally, I thank the Department of Political Science, Michigan State University, for the financial assistance which made this dissertation possible. ii TABLE OF CONTENTS LIST OF TABLES . LIST OF FIGURES . CHAPTER I. INTRODUCTION .- II. A THEORY OF INTRACOURT INFLUENCE III. DATA AND OPERATIONALIZATION Iv. DATA ANALYSIS v. SUMMARY AND CONCLUSION BIBLIOGRAPHY APPENDIX: CATEGORY SCALES . iii Page iv 12 41 62 133 144 148 Table 10. LIST OF TABLE S Category Scales . Agreement with Court-made Public Policy and "Voice" Behaviors Before Exit Linear Regression Analysis with Percent "Voice" Behaviors as the Dependent Variable . Non -separate "Voice" Behaviors and Separate "Voice" Behaviors Before Exit . Linear Regression Analysis with Non -separate "Voice" Behaviors and Separate "Voice" Behaviors as the Dependent Variables Joining "Voice" Behaviors and Writing "Voice" Behaviors Before Exit . Linear Regression Analysis with Percent Joining "Voice" Behaviors and Writing "Voice" Behaviors as Dependent Variables Non-separate Writing Behaviors and Separate Writing Behaviors Before Exit Linear Regression Analysis with Percent Non -separate Writing Behaviors and Percent Separate Writing Behaviors as Dependent Variables . "Voice" and "Non-voice" Before and After Exit . iv Page 52 63 65 67 68 69 69 71 71 72 Table 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. "Non-voice" and "Voice" Behaviors After Exit Joining "Voice" Behaviors and Writing "Voice" Behaviors After Exit Linear Regression Analysis with Percent Joining "Voice" Behaviors and Percent Writing "Voice” Behaviors After Exit as the Dependent Variables . . . . . . Scales Constituting the Subsets Freedom, Equality, and New Dealism Agreement with Court -made Public Policy and "Voice" Behaviors Before Exit within the Subsets Freedom, Equality, and New Dealism Linear Regression Analysis with "Voice" Behaviors Before Exit as the Dependent Variable Across the Freedom, Equality, and New Dealism Subsets . . Non-separate "Voice" Behaviors and Separate "Voice" Behaviors within the Subsets Freedom, Equality, and New Dealism Linear Regression Analysis with Non-separate and Separate "Voice" Behaviors as the Dependent Variables . . . . . . Joining "Voice" Behaviors and Writing "Voice" Behaviors within the Subsets Freedom, Equality, and New Dealism Linear Regression Analysis with Joining and Writing "Voice" Behaviors as the Dependent Variables across the Subsets Freedom, Equality, and New Dealism . . Page 74 74 75 78 79 80 82 84 85 86 Table 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Non-separate and Separate Writing "Voice" Behavior within the Subsets Freedom, Equality, and New Dealism Linear Regression Analysis with Non -separate and Separate Writing Behaviors as the Dependent Variables within the Subsets Freedom, Equality, and New Dealism "Non -voice" and "Voice" Before and After Exit within the Subsets Freedom, Equality, and New Dealism . . . . . . "Non -voice" and "Voice" Behaviors After Exit within the Subsets Freedom, Equality, and New Dealism Joining and Writing "Voice" Behaviors After Exit within Subsets Freedom, Equality, and New Dealism Linear Regression Analysis with Joining and Writing "Voice" Behaviors After Exit as Dependent Variables within the Subsets Freedom, Equality, and New Dealism Agreement with Court-made Public Policy and "Voice" Behaviors Before Exit Controlling for Time Period Linear Regression Analysis with "Voice" Behaviors as the Dependent Variable Controlling for Time Period Non -separate and Separate "Voice" Before Exit Controlling for Time Period Joining and Writing "Voice" Behaviors Before Exit Controlling for Time Period vi Page 88 90 91 92 93 94 97 98 100 101 Table 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. Non -separate and Separate Writing Behaviors Before Exit Controlling for Time Period Linear Regression Analysis with Non -separate and Separate "Voice" Behaviors Before Exit as Dependent Variables Controlling for Time Period Linear Regression Analysis with Joining and Writing "Voice" Behaviors as the Dependent Variables Controlling for Time Period Linear Regression Analysis with Non -separate Writing and Separate Writing "Voice" Behaviors Before Exit as Dependent Variables Controlling for Time Period "Non -voice" and "Voice" Before and After Exit Controlling for Time Period "Non-voice" and "Voice" After Exit Controlling for Time Period . . . Joining and Writing "Voice" Behaviors After Exit Controlling for Time Period Linear Regression Analysis with Joining and Writing "Voice" Behaviors After Exit as the Dependent Variables Controlling for TimePeriod...... Agreement with Court-made Public Policy and "Voice" Behaviors Before Exit within Subgroups During the Pre- and Post-Goldberg TimePeriods Linear Regression Analysis with "Voice" Behaviors as the Dependent Variable within Subgroups during the Pre- and Post -Goldberg TimePeriods vii Page . 102 103 . 104 . 105 . 106 . 108 . 109 . 110 . 112 . 113 Table 41. 42. 43. 44. 45. 46. 47. 48. 49. Non-separate and Separate "Voice" Behaviors Before Exit within Subgroups during the Pre- and Post-Goldberg Time Periods . Joining and Writing "Voice" Behaviors Before Exit within Subgroups during the Pre- and Post-Goldberg Time Periods . . Non ~separate and Separate Writing Behaviors Before Exit within Subgroups during Pre- and Post-Goldberg Time Periods Linear Regression Analysis with Non -separate and Separate "Voice" Behaviors as the Dependent Variables within Subgroups during the Pre- and Post-Goldberg Time Periods Linear Regression Analysis with Joining and Writing "Voice" Behaviors as the Dependent Variables within Subgroups during the Pre- and Post -Goldberg Time Periods . Linear Regression Analysis with Non-separate and Separate Writing Behaviors as the Dependent Variables within Subgroups during the Pre- and Post-Goldberg Time Periods "Non-voice" and "Voice" Before and After Exit within Subgroups during the Pre- and Post-Goldberg Time Periods . . "Non -voice" and "Voice" After Exit within &Ibgroups during the Pre- and Post- -Goldber.g Time Periods . . . . . . . . . . Joining and Writing "Voice" Behaviors After Exit within Subgroups during the Pre- and Post-Goldberg Time Periods viii Page 116 117 118 120 121 122 124 125 127 Table Page 50. Linear Regression Analysis with Joining and Writing Behaviors as Dependent Variables within Subgroups during the Pre- and Post-Goldberg Time Periods . . . . . . . . . . 128 Figure LIST OF FIGURES "Voice" Before Exit "Voice" After Exit "Voice" Before and After Exit Extremity of "Voice" Before and After Exit A Hypothetical Cumulative Scale Operationalization of Justice H' s Behaviors Operationalization of Justice C' s Behaviors Percent "Voice" Before Exit Page 28 30 31 33 44 45 45 64 CHAPTER I IN TRODUC TION Prior to the behavioral revolution in Political Science, the dominant mode of judicial study was characterized by a high degree of legalistic scholarship and historicism. 1 Judicial scholars of this tradition viewed the individual justice as a non- political entity. From this perspective, the justice was perceived as being primarily constrained by the legalistic nature of his decision making; he was to reach decisions only on the merits of the case according to appropriate statute, common law holding, or constitutional dictate. Such a posture, commonly referred to as mechanical jurisprudence, viewed the judicial decision as a simple deductive process; a judge did no more than deduce a decision by comparing fact and law. Cast in such a non -political, mechanical light the judiciary is of little interest or relevance to a science which seeks to explain the nature of political phenomena. Indeed, for those embracing this conception of the judicial role, there was little concern to develop a theoretical framework designed to explain and predict judicial decision making. Rather, scholars of this persuasion focused primarily on the logical and normative aspects of judicial doctrine. Such endeavors sought to clarify what a justice had said rather than determine what led him to decide or to write as he did. Within the last decade and a half, however, the study of the judiciary has been profoundly influenced by a redefinition of the judicial function. Such a redefinition, inspired by a school of legal philosophy variously referred to as "legal realism, " "socio- logical jurisprudence" or "political jurisprudence" (a philosophy which originated some 40 years before its application to the study of the judiciary), conceptualizes the judiciary as an integral part of the political process. From this perspective a justice can in no way be assumed to mechanically reach decision free of personal value biases or political preference. The implications of this redefinition in studies of the American judiciary, particularly the Supreme Court, only recently began to manifest themselves. 2 Initially scholars introduced psychological factors as a significant dimension in the judicial decision. Relying on the social -psychological model of Stimulus- Organism-Response (S-O-R), judicial researchers attempted to provide an explanation of judicial decision making. Although variations existed, most scholars viewed the case presented for decision as the stimulus to the individual justice and the subsequent vote on the merits as the response. From this basic model there followed attempts to demonstrate empirically that certain back- “ ground factors, 3 role conceptions, 4 or attitudes5 intervened between stimulus and response, and, in turn, determined (or greatly influenced) the individual response. On the surface, at least, this approach appears warranted when applied to the behavior of the individual justice on the Supreme Court. Consistent with the S-O-R model, a Supreme Court justice, considering a case in isolation, may, indeed, evaluate the question brought to the Court for interpretation within the con- text of his own personal attitude system, role conception, or past history. From this evaluative perspective, he may well respond in a manner consistent with the predispositions his relevant attitude dictates. However neatly this approach explains the manifest behavior of the individual justice, it suffers from certain deficiencies. Most notably it fails to account for the collegial, group decision- making process on the Supreme Court. By focusing on the individual justice, this approach, of necessity, treats as inconsequential this important characteristic. To assume that an individual justice' 3 attitude system has behavioral consequences only within the limited parameters of isolated decision making serves to severely limit the explanatory power of this approach. With but a few exceptions, the behavioral approach, though clearly rejecting mechanical jurisprudence, has yet to fully embrace the implications of the legal realist' 3 "political" jurisprudence. J. Woodford Howard in his article entitled "On the Fluidity of Judicial Choice"6 draws much the same conclusion. He contends that the behavioral approach infers "individual attitude from a form of group behavior, and with insufficient attention to the group inter- action which intervenes between attitude and action and qualifies both. "7 Howard argues that this lack of attention to the collegial nature of Supreme Court decision making in present behavioral research inspires serious reservations "for the precise reason that intervening variables operating in a collegial court mediate sig- nificantly between individual attitude and behavior. "8 Attempting to deal with this apparent gap in judicial theory, Howard categorizes a set of factors which he contends characterize the nature of collegial influences on individual behavior. Such major factors as socialization, strategy, and personal perception, in conjunction with other less salient factors as equivocation, pragmatism, and cross -pressure are held to adequately characterize the nature of his "fluidity" variable moderating individual behavior. 9 To support his contention that fluidity of choice exists as an intervening variable, Howard relies primarily on the private papers of Justice Murphy during his tenure on the Roosevelt and Vinson Courts. From this purely descriptive analysis, he con- cludes that ”the fluidity of choice (on the Roosevelt and Vinson Courts) serves as a reminder that judging, like most American decision -making, is situational and that causation is apt to be more than the simple mirroring of precedent or principle or personal belief systems. "10 In an earlier work, Walter J. Murphy offers similar insights into the nature of the collegial aspect of Supreme Court decision making. 11 Murphy delineates what he perceives to be the available strategies which a policy oriented justice or group of justices might use in order to accomplish a certain policy goal or set of policy goals. The focus of his study is the simple question: "How can a Justice of the Supreme Court most efficiently utilize his resources, official and personal, to achieve a particular set of policy objectives?"12 Discussing such tactics as persuasion on the merits, increasing personal regard, use of sanctions, and bargaining, Murphy quite adequately suggests what could, indeed, be the actual range of alternatives whereby the individual justice might be able to maximize the effects of collegiality in his efforts to accomplish his personal goals. Similar to Howard' s, Murphy' s categories and supporting evidence are derived primarily from the papers of past justices. Both Howard' 3 article and Murphy' 5 book are replete with examples of intracourt bargaining, manipulation, negotiation, and compromise. With little difficulty one can conclude, as Howard himself contends, that such examples vividly point out the inadequa - cies of present behavioral research in providing a full explanation of Supreme Court decision making. Despite the skill with which both authors provide evidence of these inadequacies, neither pro- vide an adequate theoretical or empirical alternative. As has been indicated, both Howard and Murphy rely on the private papers of former justices for supporting evidence. This reliance on private papers as the primary data source raises substantial questions regarding the validity and reliability of any conclusions drawn. As Goldman points out, the use of private papers has several distinct drawbacks. 13 The sample is biased by the fact that not all justices keep extensive accounts of their par- ticipation in Court deliberations. Of those justices who do, most stipulate that public access be delayed until their colleagues who served with them have left the bench. Such stipulations, though ethically admirable, cause considerable time lag between what is reported and eventual scholarly analysis. Since the sample tends to be biased, there is no way for the scholar to be assured that the reportage is complete or objective. In certain instances the papers are edited by the individual justice himself. Such editing further minimizes the completeness and objectivity likely to emerge from analysis. Though clearly bringing the validity of the empirical generalizations of the behavioralists into question, Murphy and Howard, through their choice of data source, must be challenged on similar grounds. What is apparent is that Murphy and Howard sacrifice the advantages of the behavioralist' s methodological rigor in order to control for an apparent theoretical inadequacy. The gain that is accomplished by both is clearly offset by the losses incurred by the use of such debatable supporting evidence. Even the gain of theoretical clarity suggested by Murphy and Howard is open to question. Both provide little more than categorization. Mere categorization cannot be considered theory building. Theory construction entails the abstracting of certain salient features of phenomena, delineating certain assumptions and stipulate that public access be delayed until their colleagues who served with them have left the bench. Such stipulations, though ethically admirable, cause considerable time lag between what is reported and eventual scholarly analysis. Since the sample tends to be biased, there is no way for the scholar to be assured that the reportage is complete or objective. In certain instances the papers are edited by the individual justice himself. Such editing further minimizes the completeness and objectivity likely to emerge from analysis. Though clearly bringing the validity of the empirical generalizations of the behavioralists into question, Murphy and Howard, through their choice of data source, must be challenged on similar grounds. What is apparent is that Murphy and Howard sacrifice the advantages of the behavioralist' s methodological rigor in order to control for an apparent theoretical inadequacy. The gain that is accomplished by both is clearly offset by the losses incurred by the use of such debatable supporting evidence. Even the gain of theoretical clarity suggested by Murphy and Howard is open to question. Both provide little more than categorization. Mere categorization cannot be considered theory building. Theory construction entails the abstracting of certain salient features of phenomena, delineating certain assumptions and then offering testable hypotheses which follow from these abstracted assumptions. 14 On this basis, Murphy and Howard both fail to advance a cogent theory of intracourt influence. Perhaps it is unfair to criticize both authors for failing to accomplish what neither specifically set out to do. The most that can be said is that both provide interesting speculations about the nature of the intracourt influence process, leading one to conclude that such factors must be considered in any comprehensive theory of Supreme Court decision making. What emerges from this discussion is the apparent need to extend present work beyond the 5-0 -R model and descriptive and speculative analysis to a theoretical framework designed to explain the political dimensions of the judicial decision. The pur- poSe of this present study is to do just that: to formulate and apply a theory of intracourt influence on the United States Sipreme Court. The theory presented relies on recent developments in the rational -calculus school of political science. Rational -calculus theories are noted for their conception of man as essentially a rational being; as a calculating person who analyzes relative costs and benefits of alternative activities in relation to specified self - interests or private goals. In contradistinction to sociological theories of politics, this approach assumes an individual' 3 goals as given, and, as a consequence, seeks to determine how these goals are satisfied rather than how or why these goals arose. This theory is presented in Chapter II. FOOTNOTES 1Glendon Schubert, Judicial Policy Making (Chicago: Scott, Foresman, and Company, 1965). pp. 159-161. 2For extensive discussion of the nature of political science efforts in the study of the judicial process see, Schubert, Ibid. , pp. 158-183; Glendon Schubert, "Ideologies and Attitudes, _A_c-a-demic and Judicial, " 29 Journal of Politics, February, 1967, pp. 3-40; Joel Grossman and Joseph Tannenhaus, Frontiers of Judicial Research (New York: John Wiley and Sons, Inc. , 1969), pp. 3-44; andi'Social Science Approaches to the Judicial Process, " 79 Harvard Law Review, June, 1966, pp. 1551-1628. 3Cf. Earnst Nagel, "Political Party Affiliation and Judges' Decisions, " 55 American Political Science Review, December, 1961, pp. 843 -850; John Schm—id‘hauser, “Judicial Behavior and the Sectional Crisis of 1837 -1860, " 23 Journal of Politics, November, 1961, pp. 615-640; and, David Danelski, _A_ Supreme Court Justice Is Appointed (New York: Random House, 1964). 4Joel Grossman, ”Dissenting Blocs on the Warren Court: A Study of Judicial Role Behavior, " 30 Journal of Politics, Novem- ber, 1968, pp. 1068-1090; Theodore Becker, Political Behavior and Modern Jurisprudence (Chicago: Rand McNally, 1964): and, Kenneth Vines, "The Judicial Role in the American States, " in Grossman and Tannenhaus, op. cit., pp. 461-485. 5Glendon Schubert, The Judicial Mind (Evanston: North- western University Press, 1965); Harold Spaeth, An Introduction to Supreme Court Decision Making (San Franciscozfiandler Publishing Company, 1972); and, S. Sidney Ulmer, "Supreme Court Behavior and Civil Rights, " 13 Western Political Quarterly, June, 1960, pp. 288-311. 10 11 6J. Woodford Howard, "On the Fluidity of Judicial Choice, " 62 American Political Science Review, March, 1968, pp. 43 -56. 7Ibid., p. 43. 8Ibid., p. 44. 91bido! ppo 45-47. 10Ibid., p. 51. 11Walter J. Murphy, Elements of Judicial Strategy (Chicago: The University of Chicago Press, 1964). 12Ibid. , p. 3. 13Sheldon Goldman, "Behavioral Approaches to Judicial Decision Making: Toward a Theory of Judicial Voting Behavior, " 2 Jurimetrics Journal, March, 1971, p. 153. 14David Rohde, Strategy and Ideology: The Assignment of Majority Opinions in the United States Suprene Court (unpublished Ph.D, Dissertation, University of Rochester, 1971), p. 6. CHAPTER II A THEORY OF INTRACOURT INFLUENCE The primary assumption of this theory of intracourt influence is that the individual justice is motivated in his Supreme Court decision making by the desire to have Court -made public policy approximate as closely as possible his personal policy preferences. Consistent with rational -calculus theory, it is further assumed that the individual justice behaves rationally; that is, in light of his decisional motivation, the individual justice acts to accomplish this goal as efficiently and effectively as possible. For the individual justice this is assumed to entail con- sideration of both the means and ends involved. In any particular case, the justice is assumed to first evaluate the known policy alternatives and to select the particular alternative which most closely approximates his specified preference. Further, the individual justice, as a rational decision maker in a collegial con- text, is assumed to calculate the relative costs and benefits of alternative influence strategies (strategies designed to increase the 12 13 probability that Court -made policy will approximate more closely his personal policy preferences), and to select an alternative influence strategy which will maximize net benefit relative to his specified goal. In short, as discussed in Chapter I, it is maintained that individual decision making on the Supreme Court involves more than an automatic stimulus -response pattern of judicial behavior. Individual justices are assumed to attempt to influence their col- leagues in order to maximize the utility derived from Court-made public policy. Before proceeding, it is appropriate at this point to dis - cuss the rationale underlying the major motivational assumption. There are four major reasons for assuming that personal policy preferences function as the primary motivating factor in Supreme Court decisions: 1. The nature of the appointment process to the Supreme Court is such that a justice is not subject to the constraints imposed by electoral responsibility. Once appointed, he holds office until he dies or resigns. 2 Both the lack of electoral responsibility and the lack of effective removal power serve to minimize the effects of external pressures likely to bear on his public policy decisions. 14 The Supreme Court justice is unlikely to be influenced by the desire for higher office. Schlesinger has maintained that much of the behavior of political actors can be under- stood in terms of their political ambitions. 3 Most political settings have a number of officials seeking higher public office. A political actor with such ambition is likely to reach public policy decisions in his present position in light of his perceptions of how those decisions will affect his chances for higher office. Such "progressive" ambi- tions clearly influence the extent to which personal policy preferences serve as the motivating factor in public decisions. The fact that the office of Supreme Court justice is one of the most highly respected offices in America today minimizes this constraint on the justice' 3 public behavior. Beyond the fact that high prestige lessens the influence of ambition, the fact that the only office of apparent higher status is the presidency serves to pro- vide a rather limited opportunity structure for such ambitions. Both these factors further support the assump- tion that Supreme Court justices are primarily motivated in their decision making by personal policy preferences. 15 A political actor may not have "progressive" ambitions in the higher office sense as used by Schlesinger, but none- theless may be influenced by ambition. The division of labor which characterizes most American political bodies provides for the possibility of an individual political actor desiring to increase his power within the specific decision making arena. Such internally progressive ambitions, similar to ambition for higher office, are likely to affect public policy decisions to the extent that any public policy decision will be perceived as affecting chances for internal political advancement. On the Supreme Court, however, such internally progressive ambitions are unlikely to manifest themselves. On the Court there is no division of labor, per se. All justices sit for all cases and the work load tends to be evenly distributed among them all. With the possible exception of an Associate Justice' s desire to become Chief Justice, there exist few incentives to inspire internally progressive ambitions. In addition to these factors, the Supreme Court enjoys a high degree of autonomy vis -a -vis other decision ~making bodies. The Supreme Court is the court of last resort. As such it reaches final decision on most matters brought 16 before it. This is most evident with regard to decisions based on constitutional interpretation. Decisions made on this ground are altered only by the Court itself or by constitutional amendment. 4 Although it is reasonable to contend that the Court is more autonomous in its decision making than other decision -making bodies, it must be kept in mind that such autonomy is not without its limits. Congress can affect the decisions of the Court either through new statutes (providing the particular decision was based on statutory interpretation), by threatening to limit or actually limiting the Court' 8 jurisdiction, or, as already mentioned, by constitutional amendment. The executive branch can limit the Court' S autonomy most directly through the manner in which the Court' 3 dictates are enforced or not enforced. Despite these limitations, the Court' 3 greater autonomy relative to other branches supports the contention that the individual justice is more likely than other political actors to be primarily motivated in his decision making by personal policy preferences. To prevent misinterpretation, it is important to point out that the foregoing justifications are not intended to give the impres- sion that Supreme Court justices are entirely free from constraints. 17 Nor are they meant to be interpreted as eliminating the possibility of other motivational factors. The justifications are presented simply to lend credence to the belief that, as stated, the primary motivational assumption is a viable one. The extent to which, in fact, it is a viable assumption remains to be seen. If one accepts the assumption as viable, how, then, does this assumed motivation affect the use of influence strategies by the individual justices in the process of creating Court-made public policy? To answer this, it is first necessary to outline the rule structures governing the creation of Court -made public policy and to clarify what is meant by Court -made public policy. David Rohde delineates four stages of Supreme Court decision making which are helpful in this regard. 6 These stages ("decision points") are: "the vote on whether or not to accept a case for decision, the vote on the. merits, the assignment of the majority opinion, and the bargaining over the content of the majority opinion. "7 At the first stage, members of the Court decide which of the cases petitioned to be heard possess sufficient significance to warrant further consideration and ultimate decision. For the most part, whether a case is accepted for review is governed by the "rule of four"; namely, if at least four justices vote that a case is of sufficient merit to be considered by the Court, 18 the case is accepted for review. Once the Court has accepted a case for review, members of the Court read briefs and the record of the case and hear oral argument. Upon completion, the justices vote on the merits of the case. The result at this stage is particularistic; that is, the decision applies only to the litigants at hand. The Court reaches its decision by a simple vote. The litigant wins if a majority of the justices hearing the case vote in his favor and loses if a majority vote to the contrary. At the completion of the vote on the merits, the Chief Justice (or the senior associate justice if the Chief Justice is not in the majority) assigns to himself or to another justice in the majority the task of writing the opinion of the Court. At this stage justices bargain with (or seek to influence) each other with regard to the contents of the majority opinion. This bargaining takes place since, in order for an opinion to become the opinion of the Court, at least four justices besides the author must agree to what has been written. In this final stage, Court-made public policy is created. The decisions at this final stage go beyond the particular litigants. Ideally, the statement of the Court binds the lower courts in their disposition of similar cases. At the same time, depending upon the nature of the decision, the opinion informs individuals, groups, and the legislative and executive branches of the constitutional and statutory limitations imposed on 19 future behaviors. As such, the opinion of the Court is the vehicle for communicating the policy position of those justices constituting the majority in the case at hand. Court-made public policy, however, is not viewed as manifest only in the opinion of an isolated case. Court-made public policy is an on-going enterprise wherein the Supreme Court makes conscious choices among alternative solutions to identifiable public problems. 8 The range of alternative solutions and the full extent of the public problem posed within a particular policy area are not likely to be fully presented in any one case. As such, a fully comprehensive Court -made public policy within a particular policy area is unlikely to emerge in the policy position espoused in any one isolated majority opinion. Often it takes numerous cases decided in a particular issue area before the Court shapes an integrated body of public policy within that issue area. The policy stated in a single majority opinion is viewed as merely a narrow segment of a much broader issue area. Thus, the policy implied within the confines of any one case must be seen in light of its relation to other cases within that issue area. Further, Court -made public policy, as an on -going enterprise, is not static: that is, it is sub- ject to "occasional review and marginal adjustments, individual . . . 9 exceptions, and sometImes major changes." 20 It is this flexible, non -static nature of issue wide Court -made public policy which permits an assumption regarding the fourth stage of Supreme Court decision making. It is assumed that "the bargaining over the content of the majority opinion" is likewise an on -going enterprise; a process which is not limited to the specific majority opinion at hand. Though not successful in securing the desired congruence between personal policy preference and the majority Opinion in one isolated case, the individual justice can still seek to accomplish his goal in future cases arising within that issue area. The nature of the influence strategies available to the individual justice, whether in the case at hand or in cases yet to be heard, are numerous. As discussed in Chapter I, Murphy has categorized these available influence strategies under the headings of 1) persuasion on the merits, 2) increasing personal regard, 3) the use of sanctions, and 4) bargaining. 10 These types of strate- gies as suggested by Murphy, for all intents and purposes, are empirically non -observable. The influence strategies discussed above are viewed as internal to the dynamics of the Court; strategies which may, indeed, be employed, but are essentially covert behaviors. The influence strategies posited in this theory are overt strategies; that is, 21 strategies publicly employed by the justice to increase the probability that in future cases Court-made public policy will more closely approximate his personal policy preferences. Drawing this distinc - tion allows for empirical observation of the overt behaviors and thus opens hypotheses regarding their use and effectiveness to verification. The observable behaviors deemed to be influence strategies are analogous to the consumer' s influence strategies posited by 1 A. O. Hirschman in Exit, Voice, and Loyalty. 1 Hirschman' s concern is with the ways in which individual consumers can influence producers regarding the quality of the products purchased. Hirsch- man assumes that the rational consumer attempts to influence the producer to maintain a certain preferred level of product quality. For unspecified reasons, the quality of a firm' 3 product may deteriorate in quality. He assumes that such deterioration is neither so "compelling or durable" as to prevent return to the previous quality level, provided that remedial action is taken. Hirschman contends that the producer takes such remedial action to correct quality deterioration if he is made aware of such quality deterioration. Producers are informed about these deteriorations through consumers' choice of action. The consumer has two possible influence strategies: exit and voice. In choosing the exit 22 strategy, the consumer simply stops buying the firm' s deteriorated product. As a result of the exit of consumers the producer is notified of quality deterioration by the resulting losses of revenue. Through such a market mechanism, Hirschman contends that the producer seeks remedial action to correct the quality level of his product. The consumer may also choose the "voice" strategy. In this strategy, the consumer expresses his dissatisfaction directly to management. Through the choice of such a non -market (political) strategy, the consumer effectively influences the producer to alter the product quality in order to minimize consumer discontent. Ideally, by employing either option, the consumer is able to influence the quality of the particular product. 12 Hirschman attempts to specify the relationship between these two basic strategies; the conditions underlying their employ, either separately or jointly and under what circumstances these strategies are likely to be most effective. Hirschman' 3 basic con- tention is that use of these two influence strategies is directly related to the extent of quality deterioration perceived by the con- sumer. Needless to say, this theory does not posit the existence of a one -to -one correspondence between the behavior of Hirschman' s rational consumer in the market setting and the behavior of the 23 individual justice while on the bench. Certain similarities, however, warrant consideration of Hirschman' 3 basic paradigm. It is readily apparent that Hirschman' 5 consumer is motivated in his behavior vis -5 -vis the producer in much the same way as is the individual justice. Both are assumed to be motivated by the desire to see the the object of their concern approximate as closely as possible their personal preferences regarding that object: the consumer con- cerned with the quality of the products purchased; the justice con- cerned with the nature of Court —made public policy. Both are assumed to be rational actors seeking the most effective and efficient means of accomplishing their goal. Much of the similarity ends at this point. The market setting in which the consumer employs the influence strategies sug- gested by Hirschman cannot be considered identical to the collegial decision making setting of the Court. This lack of total identity, however, does not negate the utility of Hirschman' s paradigm. Hirschman' 8 basic concern is with influence: with the ability of the consumer to influence the producer. It is not unreasonable to contend that although the setting may vary, the nature of the influence process and the strategies available to participants within that process are similar. Rather than construe the Court setting as isomorphic to the market setting, it is assumed that the influence 24 strategies available to the rational consumer and the overt influence strategies available to the individual justice are analogous. In the context of this theory, what is meant by "exit" and "voice" on the Court? Exit is the act of leaving the majority. Since, in the decision -making process of the Supreme Court, the justices constituting the majority determine the public policy em- bodied in the majority Opinion, the act of leaving the majority at the vote on the merits stage indicates that the individual justice considers the Court -made policy unsatisfactory. This view of dissent behavior assumes that the individual justice' 3 vote on the merits is premised on the public policy likely to be manifest in the Opinion of the Court. "Voice" is the act of directly expressing dis— satisfaction through separate opinion writing. This entails a range of behaviors which includes individual writing, joint writing, or joining in a concurrence or dissent. A concurrence is the act of "voice" undertaken prior to exit (when a justice is a member of the majority), 13 while a dissenting opinion is the act of "voice" under- taken after exit (when a justice is in the minority). Similar to the use of exit, the act of "voice" is viewed as an indicator of dissatis- faction with the public policy created by those supporting the opinion of the Court. This view of separate opinion writing assumes that the individual justice' 8 opinion writing behavior is premised on the public policy expressed in the majority opinion. 25 In other words, the exit and "voice" behaviors of the individual justice are a function of the public policy embodied in the opinion of the Court. The basic assumption of this theory is that the individual justice is motivated in his decision making by the desire to have Court ~made public policy approximate as closely as possible the individual justice' 3 personal policy preferences. AS such, the justice' 3 use of these overt influence strategies is a function of the difference between Court -made public policy and the personal policy preference of the individual justice. Assuming this function, it is now necessary to outline more specifically the nature of the expected relationships. Clearly, if there is no difference between the Court -made public policy and the justice' 8 personal policy preference in a particular issue area, the justice, maximally satisfied with respect to his most preferred policy, will not employ overt influence strategies: _i_._g. , he will simply vote with the majority and join the majority opinion. 14 On the other hand, if there exists a difference between the Court-made public policy and the justice' 3 personal preference, the justice will resort to overt influence behaviors; that is, he will resort to "voice" or exit, or both. On first analysis it would appear that when a difference exists between Court -made public policy and personal policy preference, the justice would simply resort to exit and exit alone. This, however, is not the case. 26 Hirschman hypothesizes that where the costs of exit are high an individual consumer will choose the exit strategy only after considering the prospects of effective use of "voice. "15 In other words, when the individual is reasonably certain that "voice" may well lead to the return of previous satisfactory quality levels, the decision to exit will be postponed. Such a decision is based on the desire to "do something" about the product of concern, and only by continuing to purchase the product can this be accomplished. Where the costs of exit are high, the individual will not exit until he has exhausted the "voice" strategy. With "voice" considered an alterna- tive to exit, Hirschman' s theory hypothesizes that the rational consumer will employ increasing levels of "voice" as quality deterioration becomes greater, and, if preferred changes are not forthcoming, will then, and only then, choose the exit strategy. On the Court, the costs of exit are high. As mentioned earlier, a justice in the minority is not directly involved in the creation of Court-made public policy and thus to exit means, for the individual justice, the loss of direct impact on the nature of the policy created. In light of the justice' 3 decisional motivation, such exclusion from the bargaining over the contents of the majority opinion makes exit a costly strategic choice. Not only does the exiting justice lose the direct opportunity to formulate the majority 27 opinion, he also violates the Court' 3 informal norm toward consensus. 16 Since exit is costly, it is assumed that the individual justice will choose "voice" as an alternative to exit; that is, the individual justice will utilize exit only after the exhaustive use of "voice" while a member of the majority. Consistent with Hirschman, it is assumed that in each issue area for each justice there exists a specific difference between Court-made public policy and the justice' 3 personal policy prefer- ences, along a range of possible differences, where the justice will choose exit: a point of difference where the justice perceives the exit strategy as the only viable alternative. Where a difference exists, but is less than the difference equal to the exit point, then as that difference between Court -made public policy and the justice' 8 personal policy preference approaches the point of exit the justice will resort to increasing levels of "voice. " In other words, the frequency of an individual justice' s use of "voice" prior to exit will increase as the difference between Court-made public policy and personal policy preference approaches exit. 17 Graphically, the expected relationship would look as follows: 28 I u I E “ : I I I” I 32 I I a? = 63 5 DUI I am ' It" E I o I Em DIFFERENCES BETWEEN COURT- MADE PUBLIC POLICY 5 PERSONAL POLICY PREFERENCE Figure 1. --"Voice" Before Exit. Once a justice has resorted to the exit strategy, one might expect the justice to cease using "voice" as an influence strategy. This expectation, however, does not hold. There is no reason to assume that a justice' S concern with the majority policy position simply dissipates once he chooses to withdraw directly from the majority bargaining process. In fact, it is more reason- able to contend that "voice" will be greater after exit than before. Having made such a costly strategic choice, the justice is left with little else but the dissenting opinion as a vehicle for expressing his policy preference. In light of the on -going nature of the public P01icy enterprise, it is reasonable to assume that the justice, though 29 not part of the majority, will still seek to influence Court-made public policy and, thus, will continue to resort to "voice. " Thus, when exit has occurred, the justice will resort to "voice. " Rather than an increasing "voice" function, however, it is hypothesized that as the difference between Court -made public policy and personal policy preference increases beyond exit the justice will resort to decreasing levels of "voice. " This decline in the frequency of "voice" after exit is based upon the contention that the greater the difference between Court -made public policy and personal policy preference manifest, the less likely will the indi- vidual justice' 3 net benefit be maximized by further influence attempts. As specified in the rationality assumption, the decision to use an influence strategy rests on the justice' 3 calculation of the relative costs and benefits of a particular strategy. Where the costs of a particular strategy outweigh the benefits derived from its use, it is assumed that the rational justice will not use such a strategy. In short, what is being suggested is that where the policy differences are great, the costs of "voice" outweigh the probability that "voice” will serve to alter that difference (that "voice, " at this point, will be effective in altering the policy of the Court in the future). Graphically, the hypothesized relationship between policy 30 differences and frequency of use of "voice" after exit is given in Figure 2 . ‘4 F""'""""""" use or "voucs' REQUENCY OF THE EXIT 4 DIFFERENCES BETWEEN COURT- NADE PUBLIC POLICY 6 PERSONAL POLICY PREFERENCE Figure 2. -- "Voice" After Exit Taken together, these hypotheses regarding changes in the use of "voice" before and after exit (including the hypothesis of greater "voice" after exit) look as follows: _———-‘_ 31 -3.-.- .. use OF ‘voucs' FREQUENCY OF THE h“-- - ---- O EXIT DIFFERENCEs BETWEEN COURT- MADE PUBLIC POLICY 8 PERSONAL POLICY PREFERENCE Figure 3. -- "Voice" Before and After Exit. Not only is it hypothesized that the frequency of use of "voice" strategies is a function of the difference between Court- made public policy and personal policy preference, but it is also hypothesized that a relationship exists between the nature of the specific "voice" strategies employed and policy differences. As specified in the definition, "voice" includes a range of behaviors or strategies. These available strategies can be transitively ordered along an extremity dimension. The extremity dimension, in this context, refers to the degree of activity (or cost) required by 32 the individual justice in carrying out that strategy. For instance, the act of joining another justice' 8 separate opinion is considered a less extreme overt influence strategy than writing one' s own separate opinion. The specifics entailed in this distinction will be clearer when discussing the Operationalization of this hypothesis. The rationale underlying the transitive ordering of these strategies is that, along the range of possible differences between Court -made public policy and personal policy preference, specific strategies are more likely to be employed at certain points than are other strategies. That is, the probability that a justice will utilize a more extreme "voice" strategy varies with changes in the dif- ference between Court-made public policy and personal policy preference. Where that difference is small the probability that a moderate "voice" strategy will be used is higher than the probability that a more extreme "voice" strategy will be employed. Simply, as the difference increases between Court -made public policy and the justice' 8 personal policy preference the extremeness of the specific voice strategy used prior to exit will increase. Similar to the previous hypothesis regarding the frequency of "voice" after exit, and for the same rationale, it is hypothesized that as the difference increases beyond the exit point the extremeness of the .nhc )Ihu-‘lLHIIIIIII 33 "voice" strategies will decrease. Graphically these hypotheses would look as follows: :4. O I I" I 3? g )- I 22 ' 2'“ I In I 0:8 I I—m I as : .> IL 0“ EXIT —-—-) DIFFERENCES BETWEEN COURT- MADE PUBLIC POLICY 8 PERSONAL POLICY PREFERENCE Figure 4. --Extremity of "Voice" Before and After Exit. If, in fact, such hypothesized relationships emerge as accurately reflecting the use of these overt influence strategies, why might one expect their use to affect Court -made public policy? The basic motivational assumption is that a justice desires Court- made public policy to approximate as closely as possible his personal policy preferences. It would appear that exit and "voice" would have little or no impact on a majority' 3 public policy position if at least five justices composing the majority opinion coalition 34 received maximum utility from the public policy as fashioned. This situation, however feasible, is unlikely. In certain respects this is analogous to a firm' 3 attempts to maximize profits. Hirsch- man contends that this effort is constrained by the need to minimize discontent. If a firm is sensitive to voice, this "concern with voice (that is, with minimizing hostility and discontent) can be expected to qualify the concern with maximum profits. "19 On the Court, the majority are assumed to be sensitive to "voice" and, thus, are likely to qualify Court-made public policy in order to minimize discontent. What is being suggested is that other factors besides the degree of congruence between Court -made public policy and per- sonal policy preference impinge upon the amount of utility a justice derives from a particular public policy. These other factors are exogenous to the dynamics of intracourt decision making. The most salient factor is the individual justice' 3 recognition of the fact that the power and authority of the Court is based, to a certain extent, upon the degree to which its decisions are viewed as the decisions of the Court as a whole. Policy made by the Court states only general guidelines; the actual implementation of policy depends upon executive enforcement and lower court execution. The greater the degree of discontent manifest on the Court, the greater the possibility 35 that these general guidelines will blur into uncertainty about what the Court "really intended. " It amounts to a small victory for a justice in the majority to have his personal policy reflected in the opinion if that policy has minimal impact in the larger political setting. As Howard contends, "Justices frequently compromise personal opinion in order to maximize their collective force and to safeguard the power and legitimacy of the Court among its reference groups. "20 In effect, there would be little accomplished if the majority' 3 public policy is an empty policy as a result of its multiple interpretations. Another factor is the Court' s concern with preserving its domain of authority. Many students of the Court contend that separate opinion writing is engaged in by disgruntled justices in an attempt to appeal to other decision -making bodies in the hope that these bodies will alter the decision of the majority through non- judicial means. 21 The extent to which the possibility exists that the Court' 3 majority opinion policy stance will be undermined alters the extent to which a majority justice gains maximum utility from the Court -made public policy. These exogenous factors do not, however, undermine the basic motivational assumption. The key phrase in that assumption is "as closely as possible. " These factors are merely "as possible" 36 constraints which must be considered in the motivational determinates of a justice' 3 decision making. They serve to underscore the validity of the assumption that "voice" and exit are effective overt influence strategies in altering the policy of the Court. Clearly, it would be of interest to ascertain the relative effectiveness of these influence strategies in effectuating change in Court -made public policy over time. This, however, is not within the scope of this study. Such a concern must remain in the realm of future research. The concern of the present study is with verification of the hypotheses presented regarding the uses of the various overt influence strategies suggested. FOOTNOTES 1The following discussion is derived from a similar discussion by David W. Rohde in Strategy and Ideology: The Assignment of Majority Opinions in the United States Supreme aurTTunpublished Ph. D. Dissertation,4University of Rochester, 1971), Chapter 2 . 2Technically, a Supreme Court justice can be impeached. This tactic, however, has yet to be successfully employed. Samuel Chase in 1804 was subject to an impeachment trial, but was acquitted by the Senate. Since Chase, the House has never approved any of the numerous impeachment resolutions. In light of this fact, impeach- ment can be considered of negligible concern to the justice. 3Joseph A. Schlesinger, Ambition and Politics: Political Careers in the United States (Chicago: Rand McNally, 1966). 4The latter only occurred three times in American history: the Eleventh, the Sixteenth and the Twenty ~sixth Amendments. As such, alteration of Court -made public policy by constitutional amend- ment can be discounted as a serious threat to the Court' s autonomy. 5Whether or not the primary assumption can be fully justi- fied is, in theory, unnecessary. As Friedman contends: the relevant question to ask about the "assumptions" of a theory is not whether they are descriptively "realistic, " for they never are, but whether they are sufficiently good approximations for the purpose in hand. . . . And this question can be answered only by seeing whether the theory works, which means whether it yields sufficiently accurate predictions. [Milton Friedman, Essays in Positive Economics (Chicago: University of Chicago fiess, 1953, p. 143 Thus, the extent to which the assumptions are viable depends more upon the fruit they bear than upon their relation to reality. 37 1968) and E the Suprel 1972, p. . r I I Judicial I Judicial S in Grossr York: Jo the on -gc Burger C stance. suspectsI held that" could be impeach the origiL I regard I I instices and, as Persuad \Iurphy,% (inaCtior legitima junior Q Colleagfl USe of t. a Separ. Increas "39 of 38 6David W. Rohde, "Policy Goals and Opinion Coalitions in the Supreme Court, " 26 Midwest Journal of Political Science, May, 1972, p. 208. 7Ibid. , p. 208. 8Richard Wells and Joel Grossman, "The Concept of Judicial Policy Making, " in Jahnige and Goldman' s The Federal Judicial System (New York: Holt, Rinehart and Winston, Inc. , 1968) and Joel Grossman, "A Model for Judicial Policy Analysis, " in Grossman and Tannenhaus, Frontiers of Judicial Research (New York: John Wiley and Sons, Inc. , 1969). 9Wells and Grossman, Ibid. , p. 298. A recent example of the on -going, non-static nature of Court ~made public policy is the Burger Court' 8 revisions of the Warren Court' s Miranda policy stance. Though not overruling Miranda' 3 rules for interrogation of suspects, the Burger Court in Harris v. New York (401 US 22 [1971]) held that a statement made without the required Miranda warnings could be admitted, not as evidence of guilt, but as evidence to impeach a defendant' s testimony. Clearly a marginal adjustment of the original Court -made public policy established in Miranda. 10Murphy, op. cit. , supra footnote 10, Chapter I. With regard to "persuasion on the merits, " Murphy suggests that justices do not always approach a case from a fixed perspective and, as a result, are open to change their minds when adequately persuaded. Grounds for such a persuasion strategy, according to Murphy, include: marshalling precedent, whether judicial action (inaction) is appropriate, whether attainment of certain goals is legitimate, and on the basis of morality. In "increasing personal regard, " Murphy discusses the basic strategies of charming or winning the friendship of one' s junior colleagues, and enhancing one' 8 personal standing with colleagues by bestowing praise, or giving advice. The "use of sanctions" strategy, for Murphy, entails the use of threats of withdrawing one' 8 vote or the threat of publishing a separate opinion. Using these threats, the justice is assumed to increase his power relative to the contents of the majority opinion. The fourth strategy,"bargaining," is an extension of the use of sanctions strategy. This strategy differs from the above in that it is ‘ of an 8551 justice is desired p Harvard 1 formulat; category under thi of exit. With res‘ the "VOIC with the tion of t}- ThOUgh r pOIicy, t a“empts preponde the (11fo Preferep in Ihe ex i“113mm EIIOrt a cit \w p. dimens used he em Ofabe 39 that it is viewed as less explicit. It is designed to take advantage of an assumed norm of consensus on the Court. In this context, the justice is able to employ a "watchful waiting" strategy until his desired positions become incorporated in the majority opinion. 11A. O. Hirschman, Exit, Voice, and Loyalty (Cambridge: Harvard University Press, 1970). 2 1 The preceding is a summary of Hirschman' 8 basic formulation found in Chapters 1, ,2, and 3, Ibid. 3Concurrence without Opinion is included under the "voice" category. Though technically not a voice behavior, it is included under this general category since it is an influence behavior short of exit. 4In certain instances a justice may be maximally satisfied with respect to his most preferred policy position and still choose the "voice" option. In such cases a justice may choose to argue with the dissenters in the case or point out his personal interpreta - tion of the majority position with which he fundamentally agrees. Though not strictly indicating dissatisfaction with Court -made public policy, these "voice" behaviors are still considered influence attempts employed in the context of the earlier assumption. The preponderance of these types of "voice" should be evident where the difference between Court-made public policy and personal policy preference is assumed to be least. This is subsequently suggested in the extremity hypothesis. 15Hirschman, op. cit., pp. 60-62. 16Particularly relevant in this regard is a portion of Canon 19 from the Canons of Judicial Ethics which states: "It is of high importance that judges constituting a court of last resort should use effort and self -restraint to promote solidarity of conclusion and the consequent influence of judicial decision. " Quoted in Murphy, 22. cit., p. 177. Though the term "approaching exit" implies a time dimension (as does the definition of the term "exit" itself), it is used here only for ease of eXplication. In this theory such implicit time references refer more specifically to a psychological (for want of a better term) dimension for the individual justice than to any longimdin public p01 the indivi' completel dent upor.I differenc» policy pre'I less of tr. need not function ll public pc' and the : 40 longitudinal or sequential ordering of cases. The Court-made public policy in one case at time t1 may be quite unsatisfactory to the individual justice while a subsequent case at time t2 may be completely to his liking. "Voice" in either of these cases is depen- dent upon the individual justice' s perception of the degree of difference between Court -made public policy and his own personal policy preference manifest in the particular case at hand, regard- less of the sequence. 18Though presented here as a linear relationship, this need not be the case. Any relationship which shows an increasing function between "voice" and the difference between Court-made public policy and personal policy preference would be satisfactory. 19Hirschman, op. cit., p. 65. 20Howard, op. cit., supra footnote 5, Chapter 1, p. 46. 2 1Murphy, op. cit., p. 60; and S. Sidney Ulmer, "Dissent and the Social Backgrounds of Supreme Court Justices, " 32 Journal of Politics, August, 1970, p. 581. techniqu individu made p1 The the making rank 0: each jt t0 an i unfavc IUStic "ans.“ area COtis. up It CHAPTER III DATA AND OPERATIONALIZATION In order to test these hypotheses one must select a technique which will provide an operational indicator of the individual justice' s personal policy preferences vis a -vis Court- made public policy. Cumulative scaling is such an indicator. The theory of cumulative scaling, as applied to judicial decision making, assumes that cases within a particular issue area can be rank ordered cumulatively along a specified dimension such that each justice votes consistently favorably in each case ordered up to an identifiable case, after which he votes consistently unfavorably. The case is the stimulus or "question" put to the justice and the dichotomous pro or con vote is the response or "answer" to the question. If the cases within a particular issue area have been successfully ordered from, for example, the most conservative stimulus to the most liberal stimulus, each justice, in theory, should vote positively to the next more liberal stimulus up to the point at which he can no longer vote positively to the next 41 more libe continue ‘ and : holds (1): Previous Providec regard: area. agreem Words ’ his OI‘c' libera] policy depem Skit p‘ 42 more liberal stimulus. At that point he should vote negatively and continue to do so to all stimuli more liberal. In summary: . one seeks to arrange respondents (for example, judges) and stimuli (for example, cases) in a matrix in an effort to determine whether persons who respond affirmatively to a weak stimulus do in fact respond affirmatively to all stronger stimuli--and, in addition, whether persons who respond negatively to a strong stimulus will also respond negatively to all weaker ones. If a single well -structured set of attitudes is shared by all or virtually all respondents, a continuum of stimuli representing varying degrees of intensity should reveal an identifiable point at which each respondent ceases to react affirmatively and begins to react negatively. 1 If the ordinal relationship assumed by cumulative scaling holds (that is, if each case ordered is equal to or greater than the previous case ordered along the specified continuum), then one is provided with an indicator of the range of agreement among justices regarding changes in Court -made public policy in a particular issue area. More importantly, it provides an indicator of each justice' 3 agreement with the changes in Court -made public policy. In other words, as a majority of justices agree with each successive stimu- lus ordered from conservative stimuli to liberal stimuli or from liberal stimuli to conservative stimuli, the Court -made public policy changes, either becoming more liberal or more conservative depending on the directionality. In terms of the hypotheses, then, an individual justice' s exit point is operationalized as the first case along a cumulative _, «we-www— scale in V that exit j differenc personal along the is assun: and pers matter c in Figurl the y-axl eXtent al ”it poi}I that it 1 many sv That is “813’ n: from C, SerVati the exr area a PUblic 43 scale in which the justice is in the minority. As cases approach that exit point along the cumulative scale, it is assumed that the difference between Court -made public policy and the justice' s personal policy preference is increasing. If, in successive cases along the cumulative scale, the justice remains in the minority, it is assumed that the difference between Court-made public policy and personal policy preference continues to increase. Testing the hypotheses, then, appears to be a simple matter of placing the cumulatively ordered cases along the x-axes in Figures 3 and 4 in Chapter II. A justice' 3 exit point is placed at the y-axis intersection and testing is a matter of calculating the extent and extremity of voice strategies used before and after the exit point. Cumulative scales, however, are constructed such that it is likely that two exit points for each justice will exist. In many scales there is unanimous agreement at both ends of the scale. That is, although the stimuli in successive cases are, say, increas- ingly more liberal, the policy position of the majority may range from completely favorable (pro) agreement with the extreme con- servative stimulus to completely unfavorable (con) agreement with the extreme liberal stimulus. Therefore, in one particular issue area a justice may exit from the majority when the Court -made public policy becomes either too liberal or too conservative. One must, 131 purposes quandar: exit poir present of the C What is joint ab to the jc is equiv PUI‘pose Justice 44 must, then, specify which exit point is considered _tl_1_e_ exit point for purposes of analysis. Rather than choose one or the other of the exit points, this quandary is resolved by viewing the two exit points as a singular exit point, or a joint absolute exit point. For the purposes of the present analysis, whether a justice agrees with a majority position of the Court at either end of the cumulative scale is irrelevant. What is important is the placement of each case relative to the joint absolute exit point; that is, positioning each case in relation to the joint absolute exit point such that the position of each case is equivalent to the distance from each separate exit point. For purposes of illustration the following hypothetical scale is presented: Cases 1 2 3 4 5 6 7 10 Justice A + + + + + + + + + - B + + + + + + + + - - C + + + + + + + - - - D + + + + + + - - - - E + + + + + - — - - _ F + + + + - - - - - - G + + + - - - - - - - H + + - - - - - - — - I + - - - - - - - - - C? '1‘ “I ‘7 '1‘ “P “.D ‘T “3 °.’ 03 co b so to VII 00 N H C Figure 5. --A Hypothetical Cumulative Scale In this 11 of points the beha as follou Fi behaVic dent Va DUbhc ing the judici 45 In this illustration, Justice H' s absolute exit point is the combination of points between cases 2 and 3 and cases 5 and 6. Placement of the behaviors of Justice H in each of the cases along the x-axis is as follows: +1 +2 I -3 -10 -9 -8 -7 -6 ‘ -5 -4 Exit Point Figure 6. --Operationalization of Justice H' s Behaviors. For Justice C a similar ordering is as follows: +1 +2 +3 +4 +5 I +6 -10 -9 -8 l -7 Exit Point Figure 7. --Operationalization of Justice C' s Behaviors. This procedure for ordering the individual justice' s behavior in each case along the x-axis operationalizes the indepen- dent variable: the increasing differences between Court-made public policy and the justice' 8 personal policy preference approach- ing the exit point and moving beyond the exit point. As cumulative scaling is presently employed by other judicial researchers, the behaviors of concern are simply the justice' 5 justice i: 10. 11. 12. 13. 14. 15. 16. 17, 18. 19, 20. 21. 22_ 14~17 Constit 46 justice' 3 pro and con vote. The range of behaviors available to the justice in any one case are much broader than this. These include: NH 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. Joining the opinion of the Court Writing the opinion of the Court Joining the opinion of the Court and joining a non -separate concurring Opinion ' Joining the Opinion of the Court and writing a non -separate concurring opinion Joining the opinion of the Court and joining a separate concurring opinion Joining the opinion of the Court and writing a joint concurring opinion Joining the opinion of the Court and writing and joining a non -separate concurring opinion Concurring in the result without opinion Writing a separate joint concurring opinion Joining a separate concurring opinion Writing a separate concurring opinion Writing and joining a separate concurring opinion Joining a per curiam Opinion Joining a per curiam opinion and writing a joint concurring opinion Joining a per curiam opinion and joining a non -separate concurring opinion Joining a per curiam opinion and writing a non -separate concurring opinion Joining a per curiam opinion and joining a separate concurring opinion Dissenting without Opinion Writing a joint dissenting opinion Joining a dissenting opinion Writing a dissenting opinion Joining and writing a dissenting opinion For the purposes of this analysis, behaviors 3 - 12 and 14 - 17 constitute the dependent variable before exit; that is, they constitute the generalized, non -specific "voice" strategy discussed 47 in Chapter 11. Taken together, behaviors 19, 20, and 21 above constitute the dependent variable after exit. This, too, is the generalized, non -specific "voice" strategy previously discussed. Once the ordering of behaviors along the x-axis is established, testing of the first hypothesis is a simple matter of determining the percentage of these before and after exit "voice" behaviors employed at each interval. Testing of the second hypothesis uses the same ordering of behaviors along the x-axis but with proportionate tabu- lation of the specific behaviors listed above. The transitive order- ing of the available behaviors on the extremity dimension as dis- cussed in Chapter II follow the ordering of strategies given above. That is, behavior 3 is less extreme than behavior 4; behavior 4 is less extreme than behavior 5, and so on. Operationally, the hypotheses can be stated as follows: H : Along a set of cumulatively scaled cases, the justices' 1 II ' II - - - frequency use of v01ce behavmrs w111 Increase as the difference between Court -made public policy and personal policy preference approaches exit. H : Along a set of cumulatively scaled cases, the justices' use of specific "voice” behaviors will become more extreme as the difference between Court-made public policy and personal policy preference approaches exit. H3: The justices' "voice" after exit will be greater than "voice" before exit. unlikely test, if the hypo creatior cient if scales, random decided WarI‘en cElses 1 these P uhiVEI‘s pPESQn. Ufl‘IEr 48 H : Along a set of cumulatively scaled cases, the justices' frequency use of "voice" behaviors will decrease as the difference between Court ~made public policy and personal policy preference increases after exit. H5: Along a set of cumulatively scaled cases, the justices' use of specific "voice" behaviors will become less extreme as the difference between Court -made public policy and personal policy preference increases after exit. Testing these hypotheses on one scale for each justice is unlikely to reveal the expected relationships. Furthermore, such a test, if positive, would not be sufficient to warrant a contention that the hypotheses were supported. Thus, testing necessitates the creation of several scales. Support for these hypotheses is suffi- cient if the expected relationship holds for all justices across all scales. The data set which is used in this study consists of a non- randomly selected set of 22 category scales consisting of 433 cases decided by the Supreme Court during the 1958-1969 term of Earl Warren' 8 Chief Justiceship. The original categorization of these cases into specific policy areas and the cumulative scaling within these policy areas was done by Harold J. Spaeth. 2 From Spaeth' s universe of 73 category scales, 22 scales were selected for the present analysis. Scale selection was based upon four factors: 1) the number of cases which constitutes the scale (scales with fewer th-E acceptab [CR's] c represe: during ti salience consists Opinion, Spaeth 1 cases u only ca cases a multipl decisio inada rescal Spaeth to 0rd Chron 49 fewer than ten cases were excluded); 2) whether the scale exhibits acceptable levels of reliability (Coefficients of Reproducibility [CR' 8] of . 90 or better) ; 3) an effort to create a sample of scales representing a cross section of policy areas decided by the Court during the time period; and 4) a subjective evaluation of the relative salience of the issue in American politics. The 22 scales selected from Spaeth' 8 original data set consisted of 582 cases. These included cases decided by full opinion, per curiam opinion and memorandum cases. In addition, Spaeth treated as separate stimuli each case wherein a number of cases were dispensed with under a single opinion. In this analysis only cases decided by a full opinion and orally argued per curiam cases are included. Unlike Spaeth' 8 work, this study treats multiple -case, single Opinion decisions as a single case. These decision rules regarding inclusion of cases in the data set resulted in a data reduction of 149 cases: a final data set of 433 cases. As a result of this data reduction, the category sets were rescaled. This rescaling followed the basic method employed by Spaeth, with one exception. 3 It is standard procedure for Spaeth to order cases in which a number of cases have the same marginal chronologically. For instance, if, in a particular scale, there were three cases, each with a 7-2 marginal, Spaeth would place the case decided 6 most recl ma rgina‘.I | situatior. coalition Court w‘; distinct justices on the rr Opinion it is p05 will dif ing of q Opinior ship th the sax the ca: With a Would no dif: Opinic 50 decided earliest following the last 8 -1 case and would place the most recently decided case preceding the first case with a 6 -3 marginal. In this study a different approach is used. In the same situation, cases are ordered according to the size of the opinion coalition. An opinion coalition consists of those members of the Court who agree with the majority opinion. An opinion coalition is distinct from a decision coalition in that the latter consists of those justices in agreement with the disposition of the case at the vote on the merits stage. Since an Opinion becomes a majority opinion if five members of the majority agree with that Opinion, it is possible, and often the case, that an opinion coalition will differ from the decision coalition. With this in mind, the order- ing of cases with the same marginals is based upon the size of the opinion coalition in each case and not upon the longitudinal relation- ship that may exist. In the previous example, the three cases with the same 7 -2 marginal would be ordered along the scale such that the case with the largest opinion coalition would follow the last case with an 8-1 marginal and the case with the smallest opinion coalition would precede the first case with a 6-3 marginal. Where there is no difference between the size of the decision coalition and the Opinion coalition, Spaeth' s procedure is used. extent to reliabili . .I Coefficn Reprodu oiconsi or bette The M1 a partic indicati result I the CR Scales surpas with the MI by -06 methc pOSe 51 Once rescaled, each scale was analyzed to determine the extent to which each satisfied minimally acceptable levels of reliability. Two measures were employed for this purpose: the Coefficient of Reproducibility (CR)5 and the Minimal Marginal Reproducibility (MMR). 6 The CR provides a measure of the degree of consistency in the cumulative scale. Conventionally a CR of . 90 or better is considered satisfactory evidence of a reliable scale. The MMR is a measure of the empirical lower limit of the CR for a particular scale. Comparing the CR with the MMR provides an indicator of the extent of improvement in a scale' s reliability as a result of scaling. These two measures, and the comparison between the CR and the MMR, are given in Table 1. The CR' 8 for all the scales exceed the minimum level of . 90, with the majority of scales surpassing this level by . 05 or better. Comparing the actual CR' 3 with the MMR indicates that the majority of the scales improve on the MMR by . 10 or better, with only three scales improving on MMR by .06. These measures indicate that the scales used in this study are quite sound. 7 Before proceeding to tests of the hypotheses, a number of methodological problems need to be dealt with. In each scale with a CR less than 1. 0, nonscale responses pose a problem in the Operationalization. Should the nonscale 52 Table 1. --Category Scales. b c CR - Scal a es N 11 CR MMR MMR 1. Contempt of Court 12 101 1. 00 . 84 . 16 2. First Amendment 20 172 1. 00 . 93 . 07 3' Legmanye . 19 162 .98 .92 . 06 Investigation 4. Security Risks 16 142 . 98 . 92 .06 5. Federal Internal Security Legislation 18 157 1. 00 . 87 . 13 6. Obscenity (State) 11 99 . 96 . 82 . 14 7. Sit-ins 14 126 .97 . 90 .07 8. Protest 13 116 .95 . 81 .14 9. Reapportionment 22 195 . 99 . 87 . 12 10. Bugging 10 84 . 94 .83 . 11 11. Search and Seizure (1958-61) 12 105 .96 .79 .17 12. Search and Seizure (1962-66) 13 117 .96 .80 .16 13. Search and Seizure (1967 -68) 15 128 . 92 . 80 .12 14. Self-incrimination 32 278 . 99 .81 . 17 15. Anti -trust 57 484 . 97 .88 .09 16. Mergers 27 217 1.00 . 89 . 11 1 7 . Transportation Regulation 30 264 . 97 . 75 . 22 18' Pubhc um.“ 23 197 .99 .79 .20 Regulation 19. Pre -emption 21 181 . 97 . 70 .27 20. Taxation of Gifts 15 127 . 96 . 90 .06 21. Comity 15 132 .97 .77 .20 22 . Judicial Administration 18 153 . 93 . 82 . 1 1 433 3, 737 aNumber of Cases in the Scale bNumber of Behaviors in the Scale cCoefficient of Reproducibility d Minimal Marginal Reproducibility for Justices my 1'- response should It occur? as indicz ceived a differ justices proced is inap‘ assume sized L functin Court- a meas behavit able to such 11 scale, do not their . Since 81151le1 Study. 53 response behaviors be excluded or included in the analysis, or should multiple exit points be created wherever nonscale responses occur? For the present study, nonscale responses are considered as indications that the individual justice voting inconsistently, per— ceived that particular stimulus from a different perspective (along a different continuum than hypothesized) than the rest of the justices. Viewed in this light, the most reasonable methodological procedure is to exclude them from the analysis. Theoretically, it is inappropriate to include behaviors in the analysis that are assumed to be motivated by other considerations than those hypothe - sized in the scale. The theory holds that the use of "voice" is a function of the difference between personal policy preference and Court -made public policy. Testing this necessitates constructing a measure that provides an indicator of this difference. Where behaviors are not directly related to this difference it is unreason- able to include those behaviors as part of the analysis. As long as such nonscale responses are randomly distributed throughout the scale, it is safe to say that these different individual perceptions do not seriously affect the overall validity of the scale, nor will their exclusion directly affect the results of the analysis. Further, since nonscale responses constitute only 1. 5% of the total behaviors analyzed, their exclusion does not constitute a serious threat to this study. justice I justice, cumula: this Ch; situatio solutior exiting ‘ exit po _- the jusl In this I | assumi 01‘ morl Thus, the x-: 31 ins acres. along than u are a 54 Another methodological problem is posed when an individual justice does not exit. Occasionally a justice, particularly a moderate justice, will agree with the majority in all cases along a particular cumulative scale. In the hypothetical scale presented earlier in this Chapter, Justice E illustrates this situation. Where this situation exists, there is no "real" exit point. Several alternative solutions present themselves: 1) exclude from the analysis all non- exiting justices, 2) consider either end of the scale as the justice' 3 exit point, or 3) consider the justice' 3 breakpoint (the point where the justice shifts from a pro vote to a con vote) as the exit point. In this study the last alternative is used. This is premised on the assumption that, had there been stimuli either more conservative or more liberal at the breakpoint, the justice would have exited. Thus, behaviors on either side of the breakpoint are placed along the x-axis before exit up to a hypothetical exit point. There were 31 instances of this absence of an exit point for individual justices across the 22 scales; that is, only 16% of the placements of cases along the x-axis necessitated following this decision rule. Since the scales were constructed across terms rather than within terms, non -participations as a result of changes in Court personnel create further methodological problems. If there are a large number of non -participations for an individual justice 55 in a particular scale, how valid is it to include these behaviors in the overall analysis? There is no clear answer to this problem. In the operationalization, the rule of thumb applied was to exclude the behaviors of an individual justice if no clearly discernible exit point or breakpoint existed in a particular scale. This decision rule resulted in a 4% reduction in the number of behaviors analyzed: from 3, 737 to 3, 582. In the present operationalization, each case is considered a unit of distance from the exit point. Certain difficulties arise when comparing across scales due to the variation in the size of the N in the various scales. Table 1 shows the range in scale N' s from a low of 10 cases in the Bugging scale to a high of 57 in the Antitrust scale. To avoid possible bias introduced by this variation, each unit of distance was normalized within each scale. The exit point is set as the zero point and each distance before and after exit is denoted by a normalized score between zero and one based on the N of the particular scale. 8 In so doing, the distances from the absolute exit point in any one scale or for any one justice can reasonably be compared relative to the absolute exit point in any other scale, all scales, or any other justice or all justices. From these normalized distances, intervals are established in order to 56 insure equal n' s at each distance from the exit point within which the frequency tabulation is calculated. The methodological problems discussed so far have been primarily mechanical. A more substantive methodological question remains to be dealt with; that is, the problem of assuming an interval measure from what is ostensibly an ordinal measure. An ordinal level measurement is characterized by the fact that among a set of objects sharing a similar basis of classification, each object so classified can be specified as equal to, greater than, or less than every other object in the class. Interval level measure- ment, while having the same properties of ordinal measurement, is further characterized by the fact that the distances or intervals between objects have meaning similar to intervals in a numeric system. That is, the interval between, say, 1 and 3 is equivalent to the interval between 7 and 9. Most of the more sophisticated statistical manipulations assume interval level measurement. For this reason, numerous statisticians argue strongly that these statistics must be employed only if an interval level of measurement is attained. 9 In the social sciences such a level of measurement is rarely obtained. For this reason, many social scientists believe that, under certain circumstances, sophisticated statistical tech- niques can be used on data characterized as less than interval level 10 measurement. 57 Rather than be statistical purists, the latter group holds that the distinction between levels of measurement should be an issue of importance in light of the "practical meaning" of the researcher' 3 work rather than an issue at the statistically theo- retical level. 11 As Hays contends, There are very many instances where the level of measure- ment may not reach the level supposed by the statistical technique, and yet the method itself may be quite adequate for showing what the experimenter wants to know. More to the point of this study is Blalock' s contention that: it may turn out that it is no more misleading to make use of dubious assumptions about the level of measurement than it is to make use of data involving . . . ordinal scales that obscure differences in the amount of variation. 13 In this analysis, a ”dubious assumption" about the level of measurement is made; that is, that the ostensibly ordinal measure - ment provided by the cumulative scale can be construed as an interval measure of the differences (distances, intervals) between Court -made public policy and the personal policy preferences of individual justices. The rationale for this assumption rests on the fact that the actual content of cases and the changes in policy mani- fest along the scale reveals more than the cumulative ordering itself implies. As both Tufte and Hays contend, level of measure- ment obtained by the data depends on what is being measured and 58 on what one wants to say about the "real properties" underlying the particular measurement. 14 By assuming this, more sophisticated statistical techniques are made available (specifically, regression analysis) which, hope- fully, will reveal greater differences in the amount of variation than would have been accomplished were the ordinal level of mea - surement left intact. It goes without saying that certain hazards are involved in making such an assumption and that caution must be exercised in the interpretation of the results. FOOTNOTES 1Joseph Tannenhaus, "The Cumulative Scaling of Judicial Decisions, " 79 Harvard Law Review, June 1966, p. 1586. 2Harold J. Spaeth, " A Theory and Methodology for the Explanation and Prediction of Supreme Court Decisions, " an unpublished manuscript. 3Harold J. Spaeth, "Unidimensionality and Item Variance in Judicial Scaling, " 10 Behavioral Science, July 1965, p. 300. 4 . Thls example, of course, assumes the absence of non- scale responses. 5The formula for the Coefficient of Reproducibility is: number of inconsistencies CR = 1 - number of responses In calculating the CR, decisions with a single dissent are excluded in order to avoid artificially inflating the CR. See Glendon Schubert, Quantitative Analysis of Judicial Behavior (Glencoe: The Free Press, 1959), p. 287. 6A3 calculated in this study, the Formula for the Minimal Marginal Reproducibility Coefficient is: 1 1 ma MMR = N Z ma + mb 1—9N where N is the total respondents, m is the more populous of the two marginal frequencies for each row, and m is the less populous. See Glendon Schubert, The Judicial Mind (Evanston: Northwestern 59 60 University Press, 1965), pp. 79-81. Schubert contends that MMR should be calculated independently for both respondents and for cases, since the higher MMR will be the lower limit for CR in any given scale. Since MMR for cases tends to approximate . 67, it was felt unnecessary to report in light of the fact that all the MMR' s for justices were higher than the approximate . 67. 7CS (Coefficient of Scalability) is excluded due to its relative meaninglessness as a measure of scale reliability. For instance, in the illustration discussed earlier, this normalizing of distance from the exit point would look as follows: normalized score . 5 . 4 . 3 . 2 . 1 0 . l . 2 Justice H +1 +2 -3 -10 -9 -8 -7 -6 -5 -4 Justice C +1 +2 +3 +4 +5 +7 -10 -9 -8 +6 Exit Point For a discussion of those who support this position see, Bela 0. Baker, et al., "Weak Measurements vs. Strong Statistics, " 26 Educational and Psychological Measurement, Summer, 1966, pp. 291-309. Specifically, they cite V. L. Senders, Measurement and Statistics (London: Oxford University Press, 1958) and S. Siegel, Nonparametric Statistics (New York: McGraw ~Hill, 1956) as examples of statisticians holding strongly to this view. 10Blalock, Hays, Tufte, and Labovitz, among others, sup- port this position. See, Hubert Blalock, Causal Inferences in Nonexperimental Research (Chapel Hill: University of North Carolina Press, 1964); William Hays, Statistics for Psychologists (New York: Holt, Rinehart, and Winston, 196—3); Edward R. Tufte, "Improving Data Analysis in Political Science, " 21 World Politics, July, 1969, pp. 641-654; and, Sanford Labovitz, "The Assignment of Numbers to Rank Order Categories, " 35 American SocioloLical 61 Review, June, 1970, pp. 515-524. Also see S. S. Stevens, "Mathematics, Measurement, and Psychophysics" in S. S. Stevens (Ed. ), Handbook of Experimental Psychology (New York: Wiley, 1951). 11Tufte, Ibid., p. 645. 12Hays, Ibid., p. 74. 13Blalock, Ibid., p. 94. 14Hays, Ibid., p. 75; and Tufte, Ibid., p. 645. CHAPTER IV DATA ANALYSIS Having specified the Operationalization and discussed the data set used in this study, it is appropriate to proceed to tests of the hypotheses. The first hypothesis of concern states that along a set of cumulatively scaled cases, individual justices will use increas - ing levels of "voice" as the difference between Court-made public policy and personal policy preference becomes greater; that is, as the differences approach the specific difference where exit occurs, individual justices will use increasing levels of "voice" behavior before that exit point. Table 2 presents the aggregate data for all justices across all scales. With the exception of a decrease of . 3% from interval 5 to interval 4, the hypothesized relationship is supported by the data. "Voice" behaviors do, indeed, increase as the dif- ference between Court -made public policy and personal policy preference approaches exit: from 8. 5% "voice" at the point where 62 63 this difference is assumed to be least to 21. 1% "voice" just prior to exit where the difference between Court -made public policy and the justices' personal policy preferences is assumed to be greatest. This relationship is shown graphically in Figure 8. Table 2. --Agreement with Court-made Public Policy and "Voice" Behaviors Before Exit. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Agreement 494 513 493 512 451 (91 . 48%) (91. 77%) (87. 26%) (84. 21%) (78. 98%) "Voice" 46 46 72 96 120 (8.52 ) (8.23 ) (12.74 ) (15.79 ) (21.02 ) N=540 N=559 N=565 N=608 N=571 ( 100% ) ( 100% ) ( 100% ) ( 100%) ( 100% ) Since the use of "voice" behavior is considered a function of the difference between Court -made public policy and personal policy preferences, a more powerful technique which can be applied to this data is linear regression. To test the hypothesized relation- ship using linear regression, differences between Court-made public policy and personal policy preference is the independent 64 30%T 25%”: '5“. PERCENT “VOICE“ IO“ ' 5x1- o + : L : + r 5 4 3 2 I EXIT NORMALIZED INTERVALS INDICATINC DIFFERENCES BETWEEN COURT-MADE PUBLIC POLICY 6 PERSONAL POLICY PREFERENCE Figure 8. --Percent "Voice" Before Exit. 65 variable (the predictor variable) and the percentage of "voice" behavior is the dependent variable. The results of the linear regression analysis are given in Table 3. Table 3. -- Linear Regression Analysis with Percent "Voice" Behaviors as the Dependent Variable. 2 Standard Dependent A B R R Error Variable (Intercept) (Slope) of the Estimate "Voice" 23.69 -3. 53 -. 54 . 29 10.67 aThe intercept, slope, R, R2, and Standard Error of the Estimate for the dependent variable "Agreement with Court-made Policy" is excluded since it is merely a linear transformation of the ”voice" percentages. The regression analysis further confirms the hypothesis. The regression equation (Y = A + BX) indicates that, on the average, for every interval increase in the difference between Court -made public policy and personal policy preference approaching the exit point, the mean percentage of "voice" increases by 3. 53%. A correlation coefficient of -. 54 indicates that 29% of the "voice" variance is accounted for by this single independent variable. 1 This is considered sufficient to indicate that the linear relationship 66 is substantial. A visual purview of the scattergram reveals that the scatter of the points around the line of best fit is more homoscedastic than heteroscedastic. This indicates that the line of best fit is not simply a function of extreme scores. On the whole, then, the first hypothesis is supported. Of course, before accepting the hypothesis as verified, it is necessary to control for the effects that certain factors such as the type of issue area, time period, and bloc formations on the Court during different time periods, might have on the aggregate result. This will be undertaken once the other hypotheses have been tested in the aggregate. The second hypothesis holds that individual justices' use of specific "voice" behaviors will become more extreme as the difference between Court -made public policy and personal policy preference approaches exit. Since the n' s in many of the specific categories are small, meaningful interpretations would be dubious at best. More valid interpretations can be drawn if the categories are collapsed. Table 4 presents the percentage of "voice" behaviors divided into two categories: 1) those "voice" behaviors employed while a justice agrees with the majority opinion [non -separate "voice"] , and 2) those "voice" behaviors employed while a justice concurs with the 67 Table 4. --Non -separate "Voice" Behaviors and Separate "Voice" Behaviors Before Exit. Normalized Intervals Indicating Difference Between Court -made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 l N“; arate 22 18 20 28 30 "VIOice" ( 4.07%) ( 3.22%) ( 3.54%) < 4.61%) ( 5.25%) Separate 24 28 52 68 90 "Voice" ( 4.44 ) ( 5.01 ) ( 9.20 ) (11.18 ) (15.76 ) N=540 N=559 N=565 N=608 N=571 ( 8.52%) ( 8.23%) (12.74%) (15.79%) (21.02%) majority decision but not with the majority opinion [separate "voice"] . Viewing the latter category as the more extreme set of behaviors, it is clear that while the use of the less extreme set of "voice" behaviors remains relatively constant approaching exit, the use of the more extreme set increases. The hypothesized relationship is shown more clearly in the regression analysis. Table 5 shows the results of the regression analysis for non -separate and separate "voice" behaviors before exit. "Voice" in conjunction with agree- ment with the majority opinion increases by . 56% for every interval closer to the exit point, while the more extreme behavior set increases by 3. 19 percent for every interval. 2 68 Table 5. --Linear Regression Analysis with Non-separate "Voice" Behaviors and Separate "Voice" Behaviors as the Dependent Variables. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Non- separate 5.18 - .56 -.20 .04 3.77 "Voice" se‘ifiaf‘e ,, 18.58 -3.19 -.63 .40 5.66 mm In Table 6 the data is categorized on the basis of whether or not the "voice" behavior used involved the actual writing of an opinion or simply the joining of another justice' 8 written opinion. Though use of both sets of behaviors increases approaching exit, the magnitude of change is greater for the more extreme behavior set. For joining behaviors the magnitude of change from the least difference interval to the greatest difference interval is 4. 44%, while for the more extreme set of behaviors (the writing behavior set) this magnitude is 8.05%. As in the previous data set, the expected relationship is clearly shown in the regression analysis. Table 7 presents the regression data for the joining behavior set and the writing behavior set. For the joining behavior 69 Table 6. --Joining "Voice" Behaviors and Writing "Voice" Behaviors Before Exit. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Joining 12 15 2 1 22 38 ( 2.22%) ( 2. 68%) ( 3. 71%) ( 3. 62%) ( 6. 66%) Writing 34 3 1 5 l 74 82 (6.30) (5.55) (9.03) (12.17) (14.36) N=540 N=559 N=565 N=608 N=571 ( 8.52%) ( 8.23%) (12.74%) (15.79%) (21.02%) Table 7. --Linear Regression Analysis with Percent Joining "Voice" Behaviors and Writing "Voice" Behaviors as Dependent Variables. 2 Standard Dependent A B R R Error Variables (Intercept) (Slope) of the Estimate Joining 6.24 - .94 -. 40 . 16 3.07 Writing 17.79 -2. 84 -.46 .21 7.88 70 set the regression coefficient is -. 94, while for the writing behavior set the coefficient is -2. 84. In other words, the more extreme set of behaviors is characterized by a far greater rate Of "voice" increase approaching exit than the less extreme behavior set. Table 8 presents data which distinguish between the two different types of writing behavior before exit: that is, between writing behaviors employed while a justice agrees with the majority opinion [non -separate writing behavior] and writing behaviors employed while a justice agrees with the majority decision but not with the majority opinion [separate writing behaviors] . In this instance, the data reveal that the less extreme behavior set (non- separate writing behaviors) does not vary to any considerable degree across intervals approaching exit, while the more extreme behavior set varies, for the most part, in an increasing fashion approaching exit. This difference is shown in the regression analysis in Table 9. From the data presented in the collapsed category analysis, it is evident that the second hypothesis is supported. Justices do, indeed, employ more extreme "voice" behavior more frequently than less extreme "voice" behaviors as the difference between Court -made public policy and personal policy preference approaches exit. 71 Table 8. -- Non -separate Writing Behaviors and Separate Writing Behaviors Before Exit. Normalized Intervals Indicating Differences Between Court —made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 N°’;;parate 19 16 15 26 24 Writing ( 3.52%) ( 2.86%) ( 2.65%) ( 4.28%) ( 4.20%) Separate 15 15 36 48 58 Writing ( 2.78 ) ( 2.68 ) ( 6.37 ) ( 7.89 ) (10.16 ) N=540 N: 559 N=565 N=608 N=571 (6.30%) ( 5.55%) ( 9.03%) (12.17%) (14.36%) Table 9. --Linear Regression Analysis with Percent Non-separate Writing Behaviors and Percent Separate Writing Behaviors as Dependent Variables. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Non- separate 4.00 - .38 - . 15 .02 3.41 Writing separétfa 13.36 -2.40 - .51 .26 5.84 Writing 72 The third hypothesis of concern states that "voice" will be employed more frequently after exit than before exit. Testing this hypothesis is a simple matter of comparing the percentages of "voice" employed before and after exit. Table 10 presents this comparative data. It is clear that "voice" after exit is employed with substantially greater frequency than "voice" before exit: a difference of 84.63%. It is safe to conclude that the third hypothesis is supported. Table 10. -- "Voice" and "Non -voice" Before and After Exit. Behaviors Before Exit After Exit "Non -voice" 2, 463 14 (86. 63%) ( 2. 00%) "Voice" 380 685 (13.37 ) (98.00 ) N = 2, 843 N = 699 ( 100% ) ( 100% ) The fact that there exist only 14 instances of dissent without ”voice" out of a possible 699 dissenting behaviors places support for the fourth hypothesis in serious jeOpardy. No matter what the distribution of the 14 "voiceless" dissents across the normalized intervals, any conclusions drawn would be precarious 73 at best and pretentious if made at all. The fourth hypothesis states that along a set of cumulatively scaled cases, an individual justice' 3 frequency use of "voice" behaviors will decrease as the difference between Court -made public policy and personal policy preference increases after exit. Table 11 presents what data there are relevant to this hypothesis. It is evident that no clear pattern emerges and that the n' s in each cell are too small for meaningful interpretation. For the latter reason, no regression analysis was attempted on this data set. It can safely be concluded that the fourth hypothesis is unsupported. The final aggregate hypothesis to be tested deals with the extremity of "voice" behaviors employed after exit. The fact that the previous hypothesis is not supported does not preclude the possibility that the extremity of voice behavior used may be related to increasing differences between Court -made public policy and personal policy preferences as hypothesized; that is, a justice' 3 use of more extreme "voice" behaviors will decrease as this difference becomes greater. As in the testing of the before exit extremity hypothesis, collapsing categories serves as a more meaningful way to present the data. Table 12 presents the data collapsed by categories into 74 Table 11. -- "Non -voice" and "Voice" Behaviors After Exit. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Behaviors Preferences Exit Least Difference Greatest Difference 1 2 3 4 "Non- 8 0 4 2 voice" ( 3. 60%) ( 0. 00%) ( 2. 25%) ( l. 47%) "Voice" 214 163 174 134 (96.40 ) ( 100%) (97.75 ) (98.53 ) N=222 N=163 N=178 N=136 ( 100% ) ( 100% ) ( 100% ) ( 100% ) Table 12. --Joining "Voice" Behaviors and Writing "Voice" Behaviors After Exit. Normalized Intervals Indicating Differences Between Court-made Public Policy and Personal Policy Behaviors Preferences Exit Least Difference Greatest Difference 1 2 3 4 Joining 102 72 78 47 (45. 95%) (44. 17%) (43. 82%) (34.56%) Writing 1 12 9 1 96 87 (50.45 ) (55.83 ) (53.93 ) (63.97 ) N=222 N=163 N=178 N=136 (96.40%) ( 100% ) (97.75%) (98.53%) 75 two divisions: 1) those "voice" behaviors characterized by a dissenting justice joining another justice' 8 dissenting opinion, and 2) those "voice" behaviors wherein a justice actually writes a dis- senting opinion himself. On first analysis it appears as though the more extreme set of "voice" behaviors (the latter set) increases as the difference between Court-made public policy and personal policy preference increases. This result is directly opposite to the relationship hypothesized. This difference does not prove very substantial in the regression analysis. In fact, both behavior sets tend to increase after exit with the joining behavior set increasing at a greater rate than the writing behavior set. Table 13 shows the regression analysis for the joining behavior set and writing behavior set. Table 13. -- Linear Regression Analysis with Percent Joining "Voice" Behaviors and Percent Writing "Voice" Behaviors After Exit as the Dependent Variables. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Joining 45.05 .68 .029 .0008 142.65 Writing 51.02 .29 .013 .0002 217.54 76 On the average, for the joining behavior set, the mean percent increase for each normalized interval increasing after exit is . 68%, while for the writing behavior set this increase is . 29%. The "goodness of fit" for the regression lines indicates that these relationships are not very substantial: only . 08% of the variance accounted for by the joining behavior set and . 02% accounted for by the writing behavior set. In fact, the slopes of the lines for both dependent variables are, for all practical purposes, parallel to the x-axis. Furthermore, the points on the scattergram are more heteroscedastic than homoscedastic. These factors indicate that the difference between Court-made public policy and personal policy preference has little or no influence on the extremity of the "voice" behavior employed. Clearly the final hypothesis is unsupported by the data. As mentioned earlier, before passing judgment on the extent to which these hypotheses are supported or unsupported by the aggregate data, it is necessary to control for the possibility that these results are artifacts of a specific type of issue area, of a particular segment of the overall time period, or of one unique subgroup or bloc of justices on the Court during those time periods. The remainder of this Chapter will deal with controlling for these factors. 77 In controlling for the possible effects that different types of issue areas may have on the hypothesized relationships, the individual justices' behaviors are divided into three mutually exclusive subsets of the total 22 scale data set. These three category subsets are labeled Freedom, Equality, and New Dealism. Those scales constituting the Freedom set are characterized by issue areas primarily pertaining to the guarantees of the Bill of Rights. The Equality set contains issue areas wherein persons are subject to some type of political, economic, or racial discrimina- tion, as well as issues pertaining to equal protection of the laws. The final category, New Dealism, is comprised of scales pertain- ing to issues of economic activity, often in the context of govern- mental regulation. The scales constituting each subset are pre- sented in Table 14. This subsetting of the data into the Freedom, Equality, and New Dealism categories is based upon analysis and categoriza- tion of the original 73 scale data set undertaken by Harold Spaeth. By correlating the rank order of the justices on each pair of cate- gory scales and computer analyzing the resulting correlation coefficients by a number of data reduction techniques, Spaeth found that over 80% of the Court' s decisions were explained by three factors which, in light of the content of the various scales 78 comprising each factor, could be appropriately labeled Freedom, Equality, and New Dealism. 4 Table 14. --Scales Constituting the Subsets Freedom, Equality, and New Dealism. Freedom Equality New Dealism First Amendment Protest Contempt of Court Legislative Sit -ins Self -incrimination Investigation Search and Seizure Anti -trust Federal Internal (1967 -196 9) Mergers Security Reapportionment Transportation Legislation Bugging Regulation Security Risks Comity Public Utility Search and Seizure Regulation (1962 -1966) Judicial ' Pre -emption Administration Taxation of Gifts The scales which comprise each subset in Table 14 are so categorized on the basis of Spaeth' 3 original analysis. Twenty (20) of the scales selected for this study were characterized by sufficient discrimination across factor loadings to be placed into one of the three categories . Two scales, Obscenity (State) and Search and Seizure (1958-1961), were part of the 20% unexplained variance in Spaeth' 3 analysis (they did not load satisfactorily on any one of the three major factors) and, as a result, are excluded from the analysis. 79 Table 15. --Agreement with Court-made Public Policy and "Voice" Behaviors Before Exit within the Subsets Freedom, Equality, and New Dealism. Normalized Intervals Indicating Differences Between Court-made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Freedom Agreement 85 146 141 119 104 (85. 86%) (95. 42%) (88.68%) (86. 86%) (75. 36%) "Voice" 14 7 18 18 34 (14.14 ) ( 4.58 ) (11.32 ) (13.14 ) (24.64 ) N= 99 N=153 N=159 N=137 N=138 ( 100% ) ( 100% ) ( 100% ) ( 100% ) ( 100% ) Equality Agreement 80 111 85 118,.-. 76 (86. 96%) (85. 38%) (80. 19%) (76. 13%) (76. 00%) "Voice" 12 19 21 37 24 (13.04 ) (14.62 ) (19.81 ) (23.87 ) (24.00 ) N: 92 N=130 N=106 N=155 N=100 ( 100% ) ( 100% ) ( 100% ) ( 100% ) ( 100% ) New Dealism Agreement 311 229 244 253 253 (96. 58%) (93. 85%) (92. 78%) (88. 77%) (82. 95%) "Voice" 11 15 19 32 52 ( 3.42 ) ( 6.15 ) ( 7.22 ) (11.23 ) (17.05 ) N=322 N=244 N=263 N=285 N=305 ( 100% ) ( 100% ) ( 100% ) ( 100%) ( 100% ) 80 Table 15 presents the data for each subset relevant to the first hypothesis. With the exception of a 9. 56% decrease in "voice" behaviors from the 5th interval to the 4th interval in the Freedom subset, the hypothesis is supported. Though not per- fectly supported, it is safe to conclude that the use of "voice" behavior before exit increases as the difference between Court- made public policy and personal policy preferences becomes greater across all three types of issue areas. Further evidence for this is provided by the results of the regression analysis. Table 16 shows the results of the regression analysis by subset Freedom, Equality, and New Dealism. Table 16. -- Linear Regression Analysis with "Voice" Behaviors Before Exit as the Dependent Variable Across the Freedom, Equality, and New Dealism Subsets. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Freedom "Voice" 1 23. 30 -3. 51 -. 42 . 18 10. 60 Equality "Voice" l 29.26 -3. 53 -.35 . 12 13.29 New Dealism "Voice" 17. 62 -2. 92 -. 45 . 20 8. 07 81 The regression indicates that, on the average, for every interval closer to the exit point prior to exit, "voice" increases by 3. 51% for Freedom, 3. 53% for Equality, and 2. 92% for New Dealism. The amount of variance accounted for by the independent variable in each subset is sufficient to indicate that the relationship is substantial: 18% for Freedom, 12% for Equality, and 20% for New Dealism. It is interesting to note that although the slopes of the lines do not vary to any considerable extent, the Y intercepts vary from a high of 29.62 for Equality, 23.30 for Freedom, to a low of 17. 62 for New Dealism. Perhaps this indicates a difference in the relative salience of each issue area during the time period. In order to test whether the extremity hypothesis is sup- ported across issue areas, collapsed categories are again employed. Table 17 presents the data for "voice" behaviors used while a justice agrees with the majority opinion (non-separate "voice") and for "voice" behaviors employed while a justice agrees with the majority decision but not with the majority opinion (separate "voice"), controlling for issue area. Only in the New Dealism subset does a clearly evident increasing pattern emerge for the less extreme "voice" behavior set ("voice" in conjunction with agreeing with the majority opinion). Comparing this increasing pattern with the pattern that emerges for the more extreme behavior set in the New 82 Table 17. --Non-separate "Voice" Behaviors and Separate "Voice" Behaviors within the Subsets Freedom, Equality, and New Dealism. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Freedom Non- separate 11 2 2 3 7 "Voice" (11. 11%) ( 1. 31%) ( 1. 26%) ( 2. 19%) ( 5. 07%) Separate 3 5 16 15 27 "Voice" ( 3.03 ) ( 3.27 ) (10.06 ) (13. 14 ) (19. 57 ) N= 99 N=153 N=159 N=137 N=138 (14. 14%) ( 4. 58%) (11. 32%) (13. 14%) (24. 57%) Equality Non- separate 4 11 6 13 9 ”Voice" ( 4. 35%) ( 8. 46%) ( 5. 66%) ( 8. 39%) ( 9.00%) Separate 8 8 15 23 15 ”Voice" ( 8.70 ) ( 6.15 ) (14.15 ) (14.84 ) (15.00 ) N= 92 N=130 N=106 N=155 N=100 (13. 04%) (14. 62%) (19. 81%) (23. 81%) (24. 00%) New Dealism Non- separate 4 5 7 9 14 "Voice" ( 1.24%) ( 2.05%) ( 2. 66%) ( 3. 16%) ( 4. 59%) Separate 7 10 12 23 38 "Voice" ( 2.17 ) ( 4.09 ) ( 4.56 ) ( 8.07 ) (12.46 ) N=322 N=244 N=263 N=285 N=305 ( 3. 42%) ( 6. 15%) ( 7.22%) (11. 23%) (17. 05%) 83 Dealism subset, it is evident that the magnitude of change is greater for the separate "voice" behavior than for the non-separate "voice" behavior: 3. 35% for the less extreme set as compared with 10.29% for the more extreme set. In all but the Equality subset the more extreme "voice" behavior set increases in the manner hypothesized. But for a decrease of 2. 55% from interval 5 to interval 4, the Equality subset reveals the expected pattern. The regression analysis provides more definite support for the hypothesis. The results of the regression within each subset are presented in Table 18. It is clear that across all three types of issue areas the more extreme "voice" behavior set increases at a more rapid rate than the less extreme behavior set as the difference between Court— made public policy and personal policy preference becomes greater. The data comparing "voice" behaviors characterized by joining or writing are given in Table 19. As before, the New Dealism subset is the only subset which reveals a consistent increase in the percent of the less extreme behavior set approach- ing exit. In comparing the magnitudes of change for the less extreme behavior set with the writing behavior set, it is evident that the increase for the more extreme behaviors in the New Dealism sub- set is the greater increase: 3. 68% for the joining behaviors as compared to 9. 96% for the writing behaviors. This, however, is 84 Table 18. -- Linear Regression Analysis with Non-separate and Separate "Voice" Behaviors as the Dependent Variables. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Freedom Non- separate 2.78 .38 .07 .005 8.39 "Voice" Separate , . ,, 19.58 -3.69 -. 55 .30 7.95 'VOIce Equality Non- separate 9.20 - 65 - 11 .01 8.30 "Voice" 598?“? ,, 18.27 -2. 43 -. 34 . 12 9. 92 01ce New Dealism Non- separate 4.41 - 46 - 18 .03 3. 50 "Voice" separate 12.96 -2. 40 -. 48 .23 6. 07 "Voice" 85 Table 19. --Joining "Voice" Behaviors and Writing "Voice" Behaviors within the Subsets Freedom, Equality, and New Dealism. Behaviors Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Preferences Least Difference 5 4 3 2 1 Freedom Joining Writing 4 3 3 8 15 (4.04%) ( 1.96%) ( 1.89%) ( 5.84%) (10.87%) 10 4 15 10 19 (10.10) (2.61) (9.43) (7.30) (13.77) N: 99 N: 153 N=159 N: 137 N: 138 (14.14%) ( 4.58%) (11.32%) (13.14%) (24.57%) Equality Joining Writing 2 5 5 6 5 ( 2.17%) ( 3.85%) ( 4.72%) ( 3.87%) ( 5.00%) 10 14 16 30 19 (10.87 ) (10.77 ) (15.09 ) (19.35 ) (19.00 ) N= 92 N=130 N=106 N=155 N=100 (13.04%) (14.62%) (19.81%) (23.81%) (24.00%) New Dealism Joining Writing 4 4 7 7 15 ( 1.24%) ( 1.64%) ( 2.66%) ( 2.46%) ( 4.92%) 7 12 12 25 37 ( 2.17 ) ( 4.92 ) ( 4.56 ) ( 8.77 ) (12.13 ) N= 322 N=244 N= 263 N=285 N=305 ( 3.42%) ( 6. 15%) ( 7.22%) (11.23%) (17.05%) Greatest Difference Exit 86 offset by the fact that none of the writing behavior sets shows a perfectly increasing pattern as hypothesized. In the New Dealism subset the percent writing behaviors declines by . 36% from inter- val 4 to interval 3, while in the Equality subset the percent writing behavior declines by . 10% from interval 5 to interval 4 and by . 35% from interval 2 to interval 1. In the Freedom subset no clear pattern emerges at all in either of the "voice" behavior sets. The mixed results evident in Table 19 are also apparent in the regres- sion analysis. These data are presented in Table 20. Table 20. -- Linear Regression Analysis with Joining and Writing "Voice” Behaviors as the Dependent Variables across the Subsets Freedom, Equality, and New Dealism. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Freedom Joining 10.51 -2.08 -. 50 .25 5.08 Writing 12.79 -1.43 -.22 .05 9.22 Equality Joining 5. 39 - . 55 -. 14 . 02 5. 94 Writing 22. 08 -2. 53 -. 27 . 07 12. 51 New Dealism Joining 4.87 - .74 -. 30 .09 3.41 Writing 12.50 -2. 12 -. 35 . 12 6. 42 87 Both the Equality and New Dealism subsets reveal the expected results. On the average, the writing behavior set increases by 2. 53% and 2. 12% approaching exit for Equality and New Dealism respectively as compared to . 55% and . 74% increases approaching exit for the joining behaviors. On the other hand, the Freedom subset indicates the opposite results; the writing behavior set increases by 1. 43% approaching exit, while the joining behavior set increases by 2. 08%. Both data analyses leave little doubt that the eXpected relationship is only partially supported. Table 21 presents the data divided into non -separate writing behaviors and separate writing behaviors. In contrast to the previous results, the data in this set reveal, for the most part, the expected relationship. Again, with the exception of the data in the New Dealism subset, no clear pattern emerges for the less extreme behavior sets. In the New Dealism subset, the mag- nitude of change from interval 5 to interval 1 is 2. 69% for the non -separate writing behaviors and 7. 2% for the separate writing behaviors, indicating a greater increase for the more extreme behavior set. Despite a decline of 2. 24% from interval 3 to inter- val 2 in the Freedom subset and a decline from interval 5 to interval 4 of 3. 76% in the Equality subset, the general pattern for the more extreme behavior set is to increase approaching exit 88 Table 21. --Non-separate and Separate Writing "Voice" Behavior within the Subsets Freedom, Equality, and New Dealism. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Freedom Non- separate 9 2 2 2 5 Writing ( 9.09%) ( 1.31%) ( 1.26%) ( 1.46%) ( 3.62%) Separate 1 2 13 8 14 Writing ( 1.01 ) ( 1.31 ) ( 8.18 ) ( 5.84 ) (10.15 ) N= 99 N=153 N=159 N=137 N=138 (10.10%) ( 2.61%) ( 9.43%) ( 7.30%) (13.77%) Equality Non- separate 3 9 5 12 7 Writing ( 3.26%) ( 6.92%) ( 4.72%) ( 7.74%) ( 7.00%) Separate 7 5 11 18 12 Writing ( 7.61 ) ( 3.85 ) (10.38 ) (11.61 ) (12.00 ) N= 92 N=130 N=106 N=155 N=100 (10.87%) (10.77%) (15.09%) (19.35%) (19.00%) New Dealism Non- separate 4 5 6 9 12 Writing ( 1.24%) ( 2.05%) ( 2.28%) ( 3.16%) ( 3.93%) Separate 3 6 6 16 25 Writing ( .93 ) (2,46 ) (2.28 ) ( 5.61 ) ( 8.20 ) N=322 N=244 N=263 N=285 N=305 ( 2.17%) ( 4.92%) ( 4.56%) ( 8.77%) (12.13%) 89 across all three subsets. The results of the regression analysis, given in Table 22, more clearly Show this result. In all three instances the more extreme behavior set increases, on the average, at a greater rate approaching exit than the less extreme behavior set. Though not perfectly supported across all three subsets, it is reasonable to conclude that the second hypothesis is supported when controlling for types of issue areas. Table 23 presents the data apprOpriate to the third hypothe- sis. It is clear that "voice" after exit is greater than ”voice" before exit across all three issue areas. The fact that the "voiceless" dissents are distributed relatively equally across all three subsets indicates that the fourth hypothesis is unlikely to be supported despite controlling for issue area. Table 24 presents what relevant data there are in this regard. It is evident that no clear pattern emerges and that the fourth hypothesis is unsupported across all three subsets. The final hypothesis to be tested controlling for issue area deals with the extremity of Specific voice behaviors employed after exit. As before, collapsing the categories serves as the more meaningful way to present the data. In Table 25 the data are collapsed into categories characterized by "voice" behaviors 90 Table 22. --Linear Regression Analysis with Non-separate and Separate Writing Behaviors as the Dependent Variables within the Subsets Freedom, Equality, and New Dealism. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Freedom Non- separate 1.41 .63 . 12 .014 7.09 Writing separgt? 11.39 -2. 06 -. 42 .18 6. 42 WrItIng Equality Non- separate 8.05 - .69 -. 11 .012 8.36 Writing Separate . . 14.03 -1.84 -.28 .08 8.85 WrItIng New Dealism Non- separate 4. 11 - . 48 -. 19 . 04 3. 74 Writing separatt? 8. 39 -1. 63 -. 39 .15 5. 47 WrItmg 91 Table 23. --"Non -voice" and "Voice" Before and After Exit within the Subsets Freedom, Equality, and New Dealism. Before Exit After Exit Freedom ”Non-voice" 595 4 (86.73%) ( 2.05%) "Voice" 91 191 (13.27 ) (97.95 ) N = 686 N = 195 ( 100% ) ( 100% ) Equality "Non-voice" 470 4 (80.62%) ( 2.68%) ”Voice" 113 145 (19.38 ) (97.32 ) N = 583 N = 149 ( 100% ) ( 100% ) New Dealism "Non-voice" 1290 6 (90.91%) ( 1.97%) "Voice" 129 298 ( 9.09 ) (98.03 ) N = 1419 N = 304 ( 100% ) ( 100% ) 92 Table 24. -- "Non -voice" and "Voice" Behaviors After Exit within the Subsets Freedom, Equality, and New Dealism. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy . Preferences BehaVIOrs Exit Least Difference Greatest Difference 1 2 3 4 Freedom "Non-voice" 4 0 0 0 ( 5.06%) ( 0.00%) ( 0.00%) ( 0.00%) "Voice" 75 22 58 36 (94.94 ) ( 100% ) ( 100%) ( 100%) N: 79 N=22 N: 58 N: 36 ( 100% ) ( 100% ) ( 100%) ( 100% ) Equality "Non-voice" 0 0 3 1 ( 0.00%) ( 0.00%) ( 9.09%) ( 2.04%) "Voice" 13 54 30 48 ( 100%) ( 100%) (90.91 ) (97.96 ) N: 13 N= 54 N: 33 N= 49 ( 100% ) ( 100% ) ( 100%) ( 100% ) New Dealism "Non-voice" 4 0 l 1 ( 3.08%) ( 0.00%) ( 1.33%) ( 2.56%) "Voice" 126 60 74 38 (96.92 ) ( 100%) (98.67 ) (97.44 ) N=130 N= 60 N= 75 N= 39 ( 100%) ( 100% ) ( 100% ) ( 100% ) 93 Table 25. --Joining and Writing "Voice" Behaviors After Exit within Subsets Freedom, Equality, and New Dealism. Normalized Intervals Indicating Differences Between Court-made Public Policy and Personal Policy Behaviors Preferences Exit Least Difference Greatest Difference 1 2 3 4 Freedom Joining 38 15 23 18 (48. 10%) (68. 18%) (39. 66%) (50. 00%) Writing 37 7 35 18 (46.84 ) (31.82 ) (60.34 ) (50.00 ) N= 79 N= 22 N= 58 N= 36 (94.94%) ( 100% ) ( 100% ) ( 100% ) Equality Joining 3 19 12 10 (23. 08%) (35. 19%) (36. 36%) (20. 14%) Writing 10 35 18 38 (76.92 ) (46.30 ) (54.55 ) (77.55 ) N: 13 N= 54 N= 33 N= 49 ( 100% ) ( 100% ) (90.91%) (97.96%) New Dealism Joining 61 25 35 17 (46. 92%) (41. 67%) (46. 67%) (43. 59%) Writing 65 35 39 21 (50.00 ) (58.33 ) (52.00 ) (53.85 ) N=130 N= 60 N= 75 N= 39 (96.92%) (100%) (98.67%) (97.44%) 94 wherein a justice joins a dissenting opinion and "voice" behaviors wherein a justice writes his own dissenting opinion. No discernible pattern emerges in any one of the three data sets, either for the joining behavior set or for the writing behavior set. On first analysis, the regression analysis appears to reveal a more clear cut pattern. Table 26 shows the results of the regression analysis. Table 26. - -Linear Regression Analysis with Joining and Writing ”Voice" Behaviors After Exit as Dependent Variables within the Subsets Freedom, Equality, and New Dealism. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Freedom Joining 59. 88 -1. 89 . 07 . 005 35. 10 Writing 33. 68 3. 83 . 14 . 02 32.66 Equality Joining 41.87 -3.77 -. 13 .02 36.07 Writing 58. 83 2. 56 . 08 . 006 32. 16 New Dealism Joining 48.39 - .58 -.03 .001 21. 11 Writing 48. 82 . 93 . 05 . 003 21. 48 95 In all three data sets the slopes of the lines for the writing behavior set are positive and the slopes for the joining behavior sets are negative. From this alone, it is clear that the extremity hypothe- sis, as stated, is unsupported. One might be led to conclude that across issue areas the frequency use of the more extreme "voice" behaviors increases rather than decreases after exit. This con- clusion, however, would be erroneous in light of the fact that only two of the six regression lines explain 2% of the variance, while the remainder explain less than 1%. A visual purview of the scattergrams reveals more heteroscedasticity than homoscedas- ticity; that is, the regression lines are primarily a function of extreme scores. In effect, the hypothesis as stated is unsupported and the pattern which apparently emerges is more an artifact of extreme scores than an actual function of the difference between Court -made public policy and personal policy preference. It is apparent from the preceding analysis that the hypotheses supported in the aggregate are similarly supported when controlling for issue area. Likewise, those hypotheses unsupported in the aggregate remain so within issue areas. What is evident thus far is that the theory and the hypotheses deduced from it hold before exit and not after exit. 96 To control for time period, the aggregate data set is divided into two mutually exclusive subsets; one data set con- sisting of all cases decided between 1958 and the end of the 1961 term and a second data set consisting of all cases decided from 1962 to 1969. For convenience, these two time periods are labeled Pre -Goldberg and Post-Goldberg, respectively. The rationale for this particular time period dichotomy is the fact that a substantial change in the Court' 3 internal make -up occurred with the retirement of Felix Frankfurter and the appointment of Arthur Goldberg. For the first time in Warren's tenure as Chief Justice, the Court was, for all practical purposes, characterized by a liberal majority. This change in the Court' 3 internal dynamics serves as an appropriate division point to control for the possible effects that different time periods (as well as changes in personnel) might have on the hypothesized relationships. Table 27 presents the tabular data appropriate to the first hypothesis. Though both time periods reveal, for the most part, increasing use of "voice" approaching exit, it is apparent that this relationship is more clearly demonstrated in the Post-Goldberg time period. This is shown as well in the regression analysis presented in Table 28. Whereas the mean percent "voice" increase for each interval approaching exit is 3. 01% during the earlier time 97 Table 27. --Agreement with Court-made Public Policy and "Voice" Behaviors Before Exit Controlling for Time Period. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy . Preferences Behav1ors Least Difference Greatest Difference Exit 5 4 3 2 l Pre -Goldberg Agreement 157 161 133 149 172 (91. 81%) (93. 06%) (87. 50%) (88. 69%) (80. 75%) "Voice" 14 12 19 19 41 (8.19 ) (6.94 ) (12.50 ) (11.31 ) (19.25 ) N=17l N=l73 N=152 N=168 N=213 ( 100%) ( 100% ) ( 100% ) ( 100%) ( 100% ) Post -Goldberg Agreement 336 349 356 358 275 (91. 55%) (91. 12%) (87. 04%) (82. 30%) (77. 90%) "Voice" 31 34 53 77 78 (8.45) (8.88 ) (12.96) (17.70 ) (22.10) N=367 N=383 N=409 N=435 N=353 ( 100%) ( 100% ) ( 100% ) ( 100% ) ( 100%) 98 Table 28. -- Linear Regression Analysis with "Voice" Behaviors as the Dependent Variable Controlling for Time Period. Standard Dependent A B R R2 Error Variable (Intercept) (Slope) of the Estimate Pre ~Goldberg "Voice" 20.13 -3.01 -. 46 .21 8. 18 Post -Goldberg "Voice" 27. 34 -4. 35 -. 58 . 34 8. 83 period, "voice" increases by 4. 35% in the Post-Goldberg data set: a difference of 1. 34%. As indicated by the percent of variance accounted for in each time period, it is evident that the hypothesized relationship holds more strongly once Goldberg joins the Court: 21% of the "voice" variance accounted for during the Pre -Goldberg time period as compared with 34% of the "voice" variance accounted for in the Post -Goldberg period. Regardless of the relative strength of the relationship in either subset, it is evident that the hypothesis is supported; justices use increasing levels of "voice" as the dif- ferences between Court -made public policy and personal policy preference becomes greater during both time periods. 99 Tables 29, 30, and 31 present the tabular data relevant to the extremity hypothesis controlling for time period. Though not perfectly increasing, the more extreme behavior sets (separate "voice, " writing, and separate writing) during the Pre -Goldberg time period tend to increase to a greater extent than the less extreme behaviors. In fact, with the exception of the joining behavior set, the less extreme behavior sets tend to decrease: from 4. 09% to 2. 82% for both non -separate "voice" and non-separate writing. The tabular results for the extremity hypothesis are more substantial for the Post-Goldberg data. Although the less extreme behaviors tend to increase approaching exit, the pattern is not consistent and the magnitude of change from interval 5 to interval 1 is not as sub- stantial as the increase manifest in the more extreme behavior sets: 2. 71% for non-separate "voice" as compared to 10. 94% for separate "voice, " 3. 23% for joining as compared to 10. 43% for writing, and 1. 83% for non-separate writing as compared to 8. 60% for separate writing. Besides the inconsistent patterns for the less extreme behavior sets and the differences in the magnitudes of "voice" change, the more extreme behavior sets all evince consistent increasing patterns during the Post -Goldberg time period. 100 Table 29. --Non-separate and Separate "Voice" Before Exit Controlling for Time Period. Normalized Intervals Indicating Differences Between Court-made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Pre -Goldb erg Non- separate 7 3 4 5 6 "Voice" ( 4.09%) ( 1.74%) ( 2.63%) ( 2.98%) ( 2.82%) Separate 7 9 15 14 35 "Voice" ( 4.09 ) ( 5.20 ) ( 9.87 ) ( 8.33 ) (16.43 ) N=171 N=173 N=152 N=168 N=213 ( 8.19%) ( 6.94%) (12.50%) (11.31%) (19.25%) Post ~Goldberg Non- separate 15 15 16 23 24 "Voice" ( 4.09%) ( 3.92%) ( 3.91%) ( 5.29%) ( 6.80%) Separate 16 19 37 54 54 "Voice" ( 4.36 ) ( 4.96 ) ( 9.05 ) (12.41 ) (15.30 ) N=367 N=383 N=409 N=435 N=353 ( 8.45%) ( 8.88%) (12.96%) (17.70%) (22.10%) 101 Table 30. --Joining and Writing "Voice" Behaviors Before Exit Controlling for Time Period. Court -made Public Policy and Personal Policy Normalized Intervals Indicating Differences Between Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Pre - Goldberg Joining 1 6 7 6 17 ( . 59%) ( 3. 47%) ( 4. 61%) ( 3. 57%) ( 7. 98%) Writing 13 6 12 13 24 (7.60) (3.47) (7.90) (7.74) (11.27) N=171 N=173 N=152 N=168 N=213 ( 8. 19%) ( 6. 94%) (12. 50%) (11. 31%) (19. 25%) P0 st - Goldberg Joining 10 8 14 16 21 ( 2. 72%) ( 2. 09%) ( 3. 42%) ( 3. 68%) ( 5. 95%) Writing 21 26 39 61 57 (5.72) (6.69) (9.54) (14.02) (16.15) N=367 N=383 N=409 N=435 N=353 ( 8.45%) ( 8.88%) (12.96%) (17.70%) (22.10%) 102 Table 31. --Non -separate and Separate Writing Behaviors Before Exit Controlling for Time Period. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Pre-Goldberg Non- separate 7 3 l 5 6 Writing ( 4.09%) ( 1.73%) ( .6670) ( 2.98%) ( 2.82%) Separate 6 3 11 8 18 Writing (3.51 ) ( 1.73 ) (7.24 ) (4.76 ) (8.45 ) N=171 N=173 N=152 N=168 N=213 ( 7.60%) ( 3.47%) ( 7.90%) ( 7.74%) (11.27%) Post - Goldberg Non- separate 12 13 14 2 1 18 Writing ( 3.27%) ( 3.39%) ( 3.42%) ( 4.61%) ( 5. 10%) Separate 9 13 25 40 39 Writing (2.45) (3.39) (6.11) (9.22) (11.05) N=367 N=383 N=409 N=434 N=353 (5.72%) (6.79%) (9.54%) (14.02%) (16.15%) 103 Similar results are evident from the regression analyses presented in Tables 32, 33, and 34. Table 32. -- Linear Regression Analysis with Non -separate and Separate "Voice" Behaviors Before Exit as Dependent Variables Controlling for Time Period. _ Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Pre ~Goldberg Non- separate 2. 66 . 005 . 002 . 000004 . 066 "Voice 889‘???“ ., 16.86 -2. 86 -. 57 .33 5. 94 01ce Post- Goldberg Non- separate 7. 21 - . 94 -. 31 . 10 4. 22 "Voice" sepf‘ra'fe ,, 19.57 -3. 32 -. 59 .35 6. 59 Voice In the Pre -Goldberg data set only the joining behavior set, as one of the less extreme behavior sets, reveals a negative regression coefficient which is greater than its corresponding more extreme behavior set. The other two less extreme behavior sets are 104 Table 33. -— Linear Regression Analysis with Joining and Writing "Voice" Behaviors as the Dependent Variables Controlling for Time Period. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Pre -Goldberg Joining 8.43 -1.48 -.39 . 15 5.09 Writing 11.09 -1. 37 -.27 .07 6. 90 Post-Goldberg Joining 5.58 - .75 -. 32 . 10 3.09 Writing 21.39 -3.47 -. 50 .25 8.73 characterized by positive, though not substantial, regression coefficients. In general, the Pre -Goldberg data set supports the hypothesis that extremity use of specific "voice" behaviors will increase approaching exit. The Post-Goldberg behaviors support the extremity hypothesis as well. Despite the fact that all of the less extreme "voice" behaviors increase approaching exit, the fact that the slopes of the lines for the more extreme behavior sets are substantially greater than those of the less extreme behavior sets lends support to the hypothesis. It is interesting to note that in all three tables the regression coefficients for the more extreme 105 Table 34. --Linear Regression Analysis with Non-separate Writing and Separate Writing "Voice" Behaviors Before Exit as Dependent Variables Controlling for Time Period. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Pre - Goldberg Non- separate 2. 09 . 005 . 002 . 000004 . 055 Writing Separate . . 9.00 -1.37 -. 32 . 10 5.87 Wr1t1ng Post- Goldberg Non- separate 5.55 - .65 -.23 .05 4.01 Writing separaft‘? 15.28 2.74 -. 53 .28 6.44 Wr1t1ng Post-Goldberg behavior sets are consistently and substantially greater than the more extreme Pre —Goldberg behavior sets. Further, the extent of the variance accounted for by the more extreme data set is consistently greater for the Post—Goldberg time period than it is for the Pre -Goldberg time period. Although both time periods provide reasonable support for the hypothesis, 106 it is evident that the Post-Goldberg time period provides the more substantial evidence in this regard. Table 35 presents the data comparing "voice" before exit and ”voice" after exit, controlling for time period. Table 35. -- "Non-voice" and "Voice" Before and After Exit Controlling for Time Period. Behaviors Before Exit After Exit Pre -Goldberg "Non-voice" 772 7 (88.03%) ( 2.77%) "Voice" 105 246 (11.97 ) (97.23 ) N = 877 N = 253 ( 100% ) ( 100% ) Post -Goldberg "Non-voice" 1674 7 (85. 98%) ( 1.58%) "Voice" 273 437 (14.02 ) (98.42 ) N = 1947 N = 444 ( 100% ) ( 100% ) It is clear that both time periods show substantial differences between ”voice" before exit and "voice" after exit, both in the 107 predicted direction. The difference between "voice" before exit and "voice" after exit during the two time periods are similar: a difference of 85. 26% for Pre -Goldberg and an 84. 04% difference for Post-Goldberg. The third hypothesis is supported across time periods; ”voice" after exit is greater than "voice" before exit during both time periods. As evident in Table 35, "voiceless" dissents are equally divided between time periods. Thus controlling for time period is not likely to improve the chances of the fourth hypothesis being supported. Table 36 presents the relevant data. As before, the patterns are not consistent and the n' s are too small for meaning- ful interpretation. Again the fourth hypothesis remains unsupported. Table 37 shows the data relevant to the after exit extremity hypothesis. Neither time period reveals a consistent after exit behavior pattern for either the more extreme behavior set or the less extreme behavior set. Both of the more extreme behavior sets (the writing behavior sets) tend to increase, though in rather erratic fashion. The results of the regression analysis, presented in Table 38, show similar findings. Writing behaviors during both time periods increase as the difference between Court-made public policy and personal policy preference becomes greater. This increasing function, however, is not particularly substantial as 108 Table 36. --"Non -voice" and "Voice" After Exit Controlling for Time Period. Normalized Intervals Indicating Differences Between Court-made Public Policy and Personal Policy , Preferences Behav1ors Exit Least Difference Greatest Difference 1 2 3 4 Pre -Goldberg "Non-voice" 7 0 0 0 ( 7.07%) ( 0.00%) ( 0.00%) ( 0.00%) "Voice" 92 50 70 34 (92.93 ) ( 100% ) ( 100% ) ( 100%) N: 99 N= 50 N= 70 N= 34 (100%) (100%) (100%) (100%) Post -G oldb erg "Non-voice" 1 0 4 2 ( .83%) ( 0.00%) ( 3.70%) ( 1.96%) "Voice” 120 113 104 100 (99.17 ) ( 100%) (96.30 ) (98.04 ) N=121 N=113 N=108 N=102 ( 100% ) ( 100% ) ( 100%) ( 100% ) 109 Table 37. --Joining and Writing "Voice" Behaviors After Exit Controlling for Time Period. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy . Preferences Behav1ors Exit Least Difference Greatest Difference 1 2 3 4 Pre -Goldb erg Joining 44 27 30 14 (44. 44%) (54. 00%) (42. 86%) (41. 18%) Writing 48 23 40 20 (48.49 ) (46.00 ) (57. 14 ) (58.82 ) N= 99 N= 50 N= 70 N= 34 (92.93%) ( 100% ) ( 100% ) ( 100% ) Post -Goldb erg Joining 58 37 48 33 (47. 93%) (32. 74%) (44. 44%) (32. 35%) Writing 62 76 56 67 (51.24 ) (67.26 ) (51.85 ) (65.69 ) N=121 N=113 N=108 N=102 (99.17%) ( 100%) (96.30%) (98.04%) 110 Table 38. -- Linear Regression Analysis with Joining and Writing "Voice" Behaviors After Exit as the Dependent Variables Controlling for Time Period. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Pre -Goldb erg Joining 48. 17 .79 .03 .0009 30.27 Writing 44. 91 1. 34 . 05 . 003 34. 83 Post -G oldberg Joining 47.26 -1.53 -.06 .004 27.23 Writing 47.72 2.08 .09 .008 27.48 indicated by an R2 of . 003 and . 008 for the writing behaviors in the Pre -Goldberg and Post-Goldberg time periods respectively. The joining behavior set during the Pre -Goldberg time period increases as well, though to a lesser extent than the more extreme ”voice" behavior set. During the Post-Goldberg time period, the joining behavior set decreases by 1. 53% for every interval beyond the exit point. Despite the tendency toward support for a relationship contrary to the hypothesized one, the relationships are not sub- stantial, accounting for minimal amounts of the variance. As in the previous data sets, the fifth hypothesis remains unsupported. 111 To control the hypotheses for bloc formations (or subgroupings) on the Court, the data sets used in the previous analysis are further sub -divided into four discrete data sets: a Liberal and Conservative subgrouping for each time period. During the Pre -Goldberg time period Justices Warren, Black, Brennan, and Douglas constitute the Liberal bloc and Justices Harlan, Clark, Stewart, Whittaker, and Frankfurter constitute the Conservative bloc. For the Post -Goldberg time period the Liberal bloc is comprised of Justices Warren, Black, Brennan, Douglas, Goldberg, and Fortas, while the Conservative subgroup during this time period is comprised of Justices Harlan, Clark, Stewart, and White. 6 The data relevant to the first hypothesis is presented in tabular form in Table 39. It is evident from these data that the behavior of the Conservatives during both time periods and the Liberals inthe Post -Goldberg time period support the hypothesized relationship; the percentage of "voice" behaviors employed before exit increases as the difference between Court -made public policy and personal policy preference becomes greater. Despite the fact of a substantial decrease in the percent of "voice" employed from the third interval to the second interval (a decrease of 12. 47%), the general trend forthe Liberal "voice" behavior during the 112 Table 39. --Agreement with Court-made Public Policy and "Voice" Behaviors Before Exit within Subgroups During the Pre- and Post-Goldberg Time Periods. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Pre -Goldberg Liberal 1 12 63 33 58 75 Agreement (90. 32%) (90. 00%) (76. 74%) (89. 23%) (82. 42%) Liberal 12 7 10 7 16 "Voice" ( 9.68 ) (10.00 ) (23.26 ) (10.77 ) (17.58 ) N=124 N= 70 N= 43 N= 65 N= 91 ( 100% ) ( 100% ) ( 100% ) ( 100% ) ( 100% ) Conservative 45 98 100 91 97 Agreement (95. 74%) (95. 15%) (91. 74%) (88. 35%) (79. 51%) Conservative 2 5 9 12 25 "Voice" ( 4.26 ) ( 4.85 ) ( 8.26 ) (11.65 ) (20.49 ) N: 47 N=103 N=109 N=103 N=122 ( 100% ) ( 100% ) ( 100% ) ( 100% ) ( 100% ) Post -Goldberg Liberal 259 217 191 214 164 Agreement (90. 24%) (90. 04%) (88. 43%) (88. 80%) (81. 19%) Liberal 28 24 25 27 38 "Voice" ( 9.76 ) ( 9.96 ) (11.57 ) (11.20 ) (18.81 ) N=287 N=241 N=216 N=241 N=202 ( 100% ) ( 100% ) ( 100% ) ( 100% ) ( 100% ) Conservative 77 132 165 144 111 Agreement (96.25%) (92. 96%) (85. 49%) (74. 23%) (73. 51%) Conservative 3 10 28 50 40 "Voice" ( 3.75 ) ( 7.04 ) (14.51 ) (25.77 ) (26.49 ) N= 80 N=142 N=193 N=194 N=151 ( 100% ) ( 100% ) ( 100% ) ( 100% ) ( 100% ) 113 Pre -Goldberg time period is increasing as hypothesized. This is shown more clearly in the regression analysis presented in Table 40. Table 40. --Linear Regression Analysis with "Voice" Behaviors as the Dependent Variable within Subgroups during the Pre- and Post-Goldberg Time Periods. Standard Dependent A B R R2 Error Variable I(Intercept) (Slope) of the Estimate Pre -Goldberg lefé'a) ,, 18.94 —1.60 -.25 .06 9.16 01ce C°r,’:,er."atfive 21.09 -4.13 -.69 .48 6.37 01ce Post - Goldberg ”bf”? ,, 22.73 -3.02 -.50 .25 7.68 V01ce C°’,‘,Ser."at,fve 34.25 -6.35 -.69 .48 9.95 V01ce The regression equations indicate that, on the average, as the difference between Court-made public policy and personal policy preference becomes greater, ”voice" increases by 1. 60% and 4. 13% during the Pre ~Goldberg time period for the Liberals and 114 Conservatives respectively and by 3. 02% for the Liberals and 6. 35% for the Conservatives during the Post-Goldberg time period. With the exception of the Pre -Goldberg Liberals, the amount of variance accounted for is substantial, particularly for the Con- servative subgroup in both time periods. What is apparent from these data is the fact that the differences evident between the Pre -Goldberg and Post -Goldberg data sets in the previous analysis are primarily the result of Conservative "voice" during the Pre -Goldberg time period and a combination of increased "voice" for both subgroupings during the Post -Goldberg time period. It is also interesting to note the fact that the slope of the "voice" line for the Conservatives is greater during the Post -Goldberg time period than it is during the Pre- Goldberg time period. This may be a function of the loss of a Conservative majority once Goldberg joined the Court in 1962. The reverse of this pattern, however, does not occur when com- paring the Post-Goldberg Liberals with the Pre -Goldberg Liberals; the slope of the "voice" line for the Pre -Goldberg "minority" Liberals is smaller than the slope for the Post-Goldberg "majority" Liberals. The latter fact minimizes the possible conclusion that "voice" is more likely to be employed by members of a subgroup when that subgroup is in the minority. 7 Despite these differences, 115 it is reasonable to conclude that the first hypothesis is supported when controlling for subgroupings. Tables 41, 42, and 43 present the tabular data relevant to the extremity hypothesis. In the Pre -Goldberg data set only one of the less extreme behavior sets (the Conservative joining behavior set) manifests a consistently increasing pattern approach- ing exit, while the remainder of the less extreme behaviors do not reveal any recognizable pattern across intervals. On the other hand, in four of the six more extreme behavior sets during this time period a discernible increasing pattern is evident (that is, in the Liberal and Conservative separate "voice, " Conservative writing, and Conservative separate writing, more extreme behaviors). Tabular non-support for the extremity hypothesis is found in the Pre -Goldberg Liberal writing and separate writing more extreme data sets. This result is to be expected in light of the weakness of the Liberal Pre -Goldberg ”voice" behaviors relative to the other subgroups in support of the first hypothesis. During the Post-Goldberg time period a clearer tabular pattern emerges with respect to the extremity predictions. In three of the less extreme behavior sets (Liberal non -separate "voice" and non -separate writing as well as Conservative non- separate writing) no meaningful pattern emerges, while for the 116 Table 41. --Non-separate and Separate "Voice" Behaviors Before Exit within Subgroups during the Pre- and Post-Goldberg Time Periods. Normalized Intervals Indicating Differences Between Court-made Public Policy and Personal Policy Behaviors Preferences Least Difference Greatest Difference Exit 5 4 3 2 1 Pre -Goldberg Liberal Non- 7 0 4 l 5 separate "Voice" ( 5. 65%) ( 0. 00%) ( 9.30%) ( 1. 54%) ( 5. 50%) Liberal 5 7 6 6 1 1 Separate "Voice" ( 4.03 ) (10.00 ) (13.95 ) ( 9.23 ) (12.08 ) N=124 N: 70 N= 43 N= 65 N= 91 ( 9. 68%) (10.00%) (23. 26%) (10. 77%) (17. 58%) Conservative Non- 0 3 0 4 1 separate "Voice" ( 0.00%) ( 2.92%) ( 0.00%) ( 3.88%) ( .82%) Conservative 2 2 9 8 24 Separate "Voice" ( 4.26 ) ( 1. 94 ) ( 8. 26 ) ( 7. 77 ) (19. 67 ) N= 47 N=103 N=109 N=103 N=122 ( 4.26%) ( 4.85%) ( 8.26%) (11. 65%) (20.49%) Post - Goldberg Liberal Non- 15 10 7 8 14 separate "Voice" ( 5.23%) ( 4. 15%) ( 3.24%) ( 3.32%) ( 6. 93%) Liberal 13 14 18 19 24 Separate "Voice" ( 4.53 ) ( 5.81 ) ( 8. 33 ) ( 7. 88 ) (11. 88 ) N=287 N=241 N=216 N=241 N=202 ( 9.76%) ( 9.96%) (11.57%) (11.20%) (18.81%) Conservative Non- 0 5 9 15 10 separate "Voice" ( 0.00%) ( 3. 52%) ( 4.66%) ( 7.73%) ( 6.62%) Conservative 3 5 19 35 30 Separate "Voice" ( 3.75 ) ( 3.52 ) ( 9.85 ) (18.04 ) (19.87 ) N: 80 N=142 N=193 N=194 N=151 ( 3. 75%) ( 7. 04%) (14. 51%) (25. 77%) (26.49%) 117 Table 42. --Joining and Writing "Voice" Behaviors Before Exit within Subgroups during the Pre- and Post-Goldberg Time Periods. _ .C' -.—4.._.___.._ fl ._ -. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy B . Preferences ehavmrs Least Difference Greatest Difference Exit 5 4 3 2 1 Pre - Goldberg Liberal 1 4 4 3 6 Joining ( .81%) ( 5.71%) ( 9.30%) ( 4.62%) ( 6.59%) Liberal 11 3 6 4 10 Writing ( 8.87 ) ( 4.28 ) (13. 95 ) ( 6.15 ) (10. 99 ) N=124 N= 70 N= 43 N= 65 N= 91 ( 9. 68%) (10. 00%) (23. 26%) (10. 77%) (17. 58%) Conservative 0 3 3 3 1 1 Joining ( 0.00%) ( 2. 92%) ( 2. 75%) ( 2. 92%) ( 9. 02%) Conservative 2 2 6 9 14 Writing ( 4.26 ) ( 1. 94 ) ( 5. 51 ) ( 8.74 ) (11.48 ) N= 47 N=103 N=109 N=103 N=122 ( 4.26%) ( 4.85%) ( 8.26%) (11. 65%) (20.49%) Post — Goldberg Liberal 9 8 9 7 13 Joining ( 3. 14%) ( 3. 32%) ( 4. 17%) ( 2. 91%) ( 6.44%) Liberal 19 16 16 20 25 Writing ( 6.62 ) ( 6.64 ) ( 7.41 ) ( 8.29 ) (12.38 ) N=287 N=241 N=216 N=241 N=202 ( 9.76%) ( 9.96%) (11.57%) (11.20%) (18.81%) Conservative 1 0 5 9 8 Joining ( 1.25%) ( 0.00%) ( 2.59%) ( 4.64%) ( 5.96%) Conservative 2 10 23 41 32 Writing ( 2.50 ) ( 7.04 ) (11.92 ) (21.13 ) (21.19 ) N= 80 N=142 N=193 N=194 N=151 ( 3. 75%) ( 7. 04%) (14. 51%) (25. 77%) (26. 49%) 118 Table 43. --Non -separate and Separate Writing Behaviors Before Exit within Subgroups during Pre- and Post-Goldberg Time Periods. Behaviors Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Preferences Least Difference Greatest Difference 5 4 3 2 1 Pre - Goldberg Liberal Non- 7 0 1 1 5 separate Writing ( 5. 65%) ( O. 00%) ( 2. 33%) ( 1. 54%) ( 5. 50%) Liberal 4 3 5 3 5 Separate Writing ( 3.23 ) ( 4. 29 ) (11. 63 ) ( 4. 62 ) ( 5. 50 ) N=124 N= 70 N= 43 N= 65 N= 91 ( 8. 87%) ( 4. 29%) (13. 95%) ( 6. 15%) (10. 99%) Conservative Non- 0 2 0 4 1 separate Writing ( 0.00%). ( 1. 94%) ( 0. 00%) ( 3. 88%) ( . 82%) Conservative 2 0 6 5 13 Separate Writing ( 4. 26 ) ( 0. 00 ) ( 5. 51 ) ( 4. 85 ) (10. 66 ) N= 47 N=103 N=109 N=103 N=122 ( 4. 26%) ( 1. 94%) ( 5. 51%) ( 8. 74%) (11. 48%) Post - Goldberg Liberal Non- 12 8 6 6 10 separate Writing ( 4. 18%) ( 3. 32%) ( 2. 78%) ( 2. 48%) ( 4. 95%) Liberal 7 8 10 14 15 Separate Writing ( 2.44 ) ( 3.32 ) ( 4.63 ) ( 5.81 ) ( 7.43 ) N=287 N=241 N=216 N=241 N=202 ( 6. 62%) ( 6.64%) ( 7.41%) ( 5.29%) (12. 38%) Conservative Non- 0 5 8 15 8 separate Writing ( 0. 00%) ( 3. 52%) ( 4. 15%) ( 7.73%) ( 5. 30%) Conservative 2 5 15 26 24 Separate Writing ( 2. 50 ) ( 3. 52 ) ( 7.77 ) (13.40 ) (15. 89 ) N= 80 N=142 N=193 N=194 N=151 ( 3. 75%) ( 7. 04%) (11. 92%) (21. 13%) (21. 19%) 119 more extreme behavior counterparts to each of these a clearly increasing pattern is apparent. In the remaining three less extreme behavior sets an increasing pattern emerges, but in all three instances the magnitude of increase from interval five to interval one is substantially less than the magnitude of the per- centage increase in the counterpart more extreme behavior set: 6. 62% for non -separate "voice" as compared to 16. 12% separate ”voice" in the Conservative subgroup; 4. 71% for the joining behavior set as compared to 18. 69% for the more extreme writing behavior set in the Conservative subgroup; and 3. 30% for the join- ing set as compared to 5.76% for the writing set in the Liberal subgroup. In sum, all six of the more extreme behavior sets show a tabular pattern which is clearly increasing across intervals approaching exit, while in the less extreme behavior sets only three increasing patterns are apparent (increasing patterns which are less substantial than the increasing pattern for each counterpart more extreme behavior set). The results of the regression analyses relevant to the extremity hypothesis are presented in Tables 44, 45, and 46. These data support the extremity hypothesis in all but one instance. This instance occurs during the Pre -Goldberg time period where the Liberal joining behavior set increases, on the average, by 1. 61% 120 Table 44. -- Linear Regression Analysis with Non-separate and Separate "Voice" Behaviors as the Dependent Variables within Subgroups during the Pre- and Post-Goldberg Time Periods. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) Of the Estimate Pre - Goldberg leeral NOR‘ , ,, 3. 50 .23 .05 .003 6. 86 separate V01ce Liberal Separate "Voice" 14' 06 4° 49 - 39 .15 5. 23 Conservanve Non; 1 99 _ 17 _. 09 .003 3.06 separate V01ce Conservative 19 10 _3 96 _ 70 49 5 94 Separate "Voice" ° ° . . Post — Goldberg leeral N0?‘ , ,, 6.19 - .57 -.20 .04 4. 05 separate 'Vo1ce Liberal Separate "Voicen 15- 07 '2- 41 " 52 '27 5' 90 Conservanxe Non; 8. 74 _1. 49 _. 46 . 21 4, 36 separate V01ce Conservat1ve 24. 82 _4. 69 -. 70 . 49 7, 27 Separate "V oice" 121 Table 45. -- Linear Regression Analysis with Joining and Writing "Voice" Behaviors as the Dependent Variables within Subgroups during the Pre- and Post-Goldberg Time Periods. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the . Estimate Pre - Goldberg leerf‘l, 8. 67 -1.16 -. 26 . 07 6.50 Jom1ng leerél, 8.89 - .09 -.02 .0004 8.15 Wr1t1ng Conservative . . 8.25 -1.74 -. 57 .33 3.76 Jom1ng consef‘fatlve 12.85 -2. 39 -. 49 .24 6. 43 Wr1t1ng Post - Goldberg leerf‘l, 5.03 - .41 -. 18 .03 3.06 Jom1ng ”be“? 17.47 -2. 51 -. 39 . 15 8. 53 Wr1t1ng Conservative . . 6.39 -1.27 -. 54 .29 2.95 Jommg Conservative 27.25 -4.91 -.65 .42 8.61 Writing 122 Table 46. -- Linear Regression Analysis with Non -separate and Separate Writing Behaviors as the Dependent Variables within Subgroups during the Pre- and Post-Goldberg Time Periods. Standard Dependent A B R R2 Error Variables (Intercept) (Slope) of the Estimate Pre -Goldberg Liberal NOn‘ .. 2.21 .23 .07 .005 6. 52 separate Wr1t1ng Liberal Separate Writing 6 68 - 32 -. 09 . 008 5. 79 Conservative Non- 1 99 _ 17 _ 09 008 3 06 separate Writing ' ' ' Conservative . - . -. . . 12 Separate Writing 10 86 2 22 48 23 6 Post-Goldberg leeral N‘m‘ . . 4.13 - .25 -.09 .008 3.60 separate Wr1t1ng Liberal Separate Writing 12. 47 -2. 14 -. 42 . 18 6. 87 Conservative Non- 7 68 _1 26 _ 39 15 4 46 separate Writing ' ' ° ' ' C°nservatwe 19.48 -3. 65 -. 70 .49 5. 51 Separate Writing 123 for each interval approaching exit, while the more extreme behavior set, the writing behaviors, increases by only . 09% for each interval. On the whole, then, considering both sets of results, the extremity hypothesis is supported, for the most part, within subgroups across both time periods. Table 47 presents data comparing "voice" and "non -voice" . 1'-“ behaviors before and after exit within subgroupings during the time periods under consideration. In all eight cases the percentage of "voice" after exit is substantially greater than the percentage of "voice" before exit: a difference of 81. 80% for Pre —Goldberg Liberals, 89. 05% for Pre -Goldberg Conservatives, and 85. 60% and 81.69% for Post-Goldberg Liberals and Conservatives respec- tively. With little doubt the third hypothesis remains supported when controlling for bloc formations on the Court during both time periods. Data relevant to the fourth hypothesis is shown in Table 48. Considering the low n' s in each of the "non -voice" cells and the lack of a discernible pattern across the intervals after exit within all subgroupings, it is reasonable to conclude that the fourth hypothesis is unsupported. "Voice" does not decline after exit as the difference between Court-made public policy and personal policy preference becomes greater when controlling for bloc forma - tions on the Court. 124 Table 47. -- "Non-voice" and "Voice" Before and After Exit within Subgroups during the Pre- and Post -Goldberg Time Periods. Behaviors Before Exit After Exit Pre - Goldberg Liberal 341 7 "Non-voice" (86.77%) ( 4. 97%) Liberal 52 134 "Voice" (13.23 ) (95.03 ) N = 393 N = 141 ( 100% ) ( 100% ) Conservative 431 0 "Non-voice" (89.05%) ( 0.00%) Conservative 53 112 "Voice" (10.95 ) ( 100% ) N = 484 N = 112 ( 100% ) ( 100% ) Post - Goldberg Liberal 1045 4 "Non-voice" (88. 04%) ( 2. 44%) Liberal 142 160 "Voice" (11.96 ) (97.56 ) N = 1187 N = 164 ( 100% ) ( 100% ) Conservative 629 3 "Non-voice" (82.76%) ( 1.07%) Conservative 131 277 "Voice" (17.24 ) (98.93 ) N = 760 N = 280 ( 100% ) ( 100% ) 125 Table 48. -- ”Non -voice" and "Voice" After Exit within Subgroups during the Pre- and Post -Goldberg Time Periods. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy Behaviors Preferences Exit Least Difference Greatest Difference 1 . 2 3 4 Pre - Goldberg Liberal 7 0 0 0 "Non-voice" (15.56%) ( 0.00%) ( 0.00%) ( 0.00%) Liberal 38 22 47 27 "Voice" (84.44 ) ( 100% ) ( 100% ) ( 100% ) N = 45 N = 22 N = 47 N = 27 ( 100%) ( 100% ) ( 100% ) ( 100% ) Conservative 0 0 0 0 ”Non-voice" ( 0.00%) ( 0.00%) ( 0.00%) ( 0.00%) Conservative 54 28 23 7 "Voice" ( 100% ) ( 100% ) ( 100% ) ( 100% ) N = 54 N = 28 N = 23 N = 7 ( 100% ) ( 100% ) ( 100% ) ( 100% ) Post-Goldberg Liberal 1 0 1 2 "Non-voice" ( 2.22%) ( 0.00%) ( 3.45%) ( 4.88%) Liberal 44 49 28 39 "Voice" (97.78 ) ( 100% ) (96.55 ) (95.12 ) N = 45 N = 49 N = 29 N = 41 ( 100% ) ( 100% ) ( 100% ) ( 100% ) Conservative 0 0 3 0 "Non-voice" ( 0.00%) ( 0.00%) ( 3.80%) ( 0.00%) Conservative 76 64 76 61 "Voice" ( 100% ) ( 100% ) (96.20 ) ( 100% ) N = 76 N = 64 N = 79 N = 61 ( 100%) ( 100%) ( 100%) ( 100% ) 126 The final hypothesis of concern, the after exit extremity hypothesis, is similarly unsupported by the data. The collapsed category data within subgroupings across time periods is presented in tabular form in Table 49. Despite the apparent percentage increase in the after exit behaviors of the Pre -Goldberg Conserva- tives and the Post-Goldberg Liberals, the alternative conclusion (that more extreme "voice" will increase after exit) is not sup- ported by the regression analysis. The results of the linear regression are shown in Table 50. In the case of the Pre -Goldberg Conservatives and both of the Post-Goldberg subgroupings, the more extreme behaviors tend to increase while the less extreme behaviors decrease. This relationship, however, is not very sub- stantial as shown by the minimal amounts of variance accounted for by each: .04% for both behavior sets for the Pre -Goldberg Con- servatives, .6% and . 1% for each set of behaviors for the Post- Goldberg Liberals, and . 2% and .4% for each behavior set for the Post-Goldberg Conservatives. In light of this it is clear that ”voice" after exit is, for all practical purposes, unrelated to increasing differences between Court —made public policy and personal policy preference within subgroupings. In conclusion, it is apparent that the before exit hypotheses are supported in the aggregate as well as when controlling for issue 127 Table 49. --Joining and Writing ”Voice" Behaviors After Exit within Subgroups during the Pre- and Post-Goldberg Time "T4? ..-u_...- __...__.fl._.-r__._..q.. ‘— ’ Periods. Normalized Intervals Indicating Differences Between Court -made Public Policy and Personal Policy . Preferences Behav1ors Exit Least Difference Greatest Difference 1 2 3 4 Pre -Goldberg Liberal 15 13 20 11 Joining (33. 33%) (59. 09%) (42. 55%) (40. 74%) Liberal 23 9 27 16 Writing (51. 11 ) (40. 91 ) (57.45 ) (59.26 ) N= 45 N= 22 N: 47 N= 27 (84.44%) ( 100% ) ( 100% ) ( 100% ) Conservative 29 14 10 3 Joining (53. 70%) (50. 00%) (43. 48%) (42. 86%) Conservative 25 14 13 4 Writing (46. 30 ) (50. 00 ) (56. 52 ) (57. 14 ) N: 54 N: 28 N: 23 N: 7 ( 100% ) ( 100% ) ( 100% ) ( 100% ) Post -Goldberg Liberal 18 20 11 11 Joining (40. 00%) (40. 82%) (37. 93%) (26.83%) Liberal 26 29 17 28 Writing (57.78 ) (59.18 ) (58.62 ) (68.29 ) N= 45 N= 49 N= 29 N= 41 (97. 78%) ( 100% ) (96. 55%) (95. 12%) Conservative 40 17 37 22 Joining (52 . 63%) (26.56%) (46. 84%) (36. 07%) Conservative 36 47 39 39 Writing (47.37 ) (73. 44 ) (49. 36 ) (63. 93 ) N= 76 N= 64 N= 79 N= 61 ( 100% ) ( 100% ) (96.20%) ( 100% ) 128 Table 50. -- Linear Regression Analysis with Joining and Writing Behaviors as Dependent Variables within Subgroups during the Pre- and Post-Goldberg Time Periods. Standard Dependent A B R R2 Error Variables (Intercept) (S1099) 0f the Estimate Pre -Goldberg Liberal. 33.59 2.80 .12 .014 29.05 Jom1ng leerafl, 46.13 1. 75 .08 .006 26.06 Wr1t1ng Conservat1ve 55. 43 _ . 64 _. 02 , 0004 32.41 Jom1ng Conservative 44. 57 . 64 . 02 . 0004 32, 41 Wr1t1ng Post- Goldberg ”ber,“, 48.90 -217 -. 08 .006 33.61 Jom1ng leerél, 44.52 2. 79 .10 . 01 33.73 Wr1t1ng Conservative 45. 04 _ . 71 -. 04 .0015 16. 99 Jom1ng Conservatwe 52. 63 . 94 . 05 . 004 20, 41 Wr1t1ng 129 area, time period, and subgroupings within time periods. It is also evident that "voice" after exit is substantially greater than "voice" before exit in each of the data sets analyzed. Finally, both of the after exit hypotheses are clearly unsupported in the aggregate and across all controls. These results will be discussed more fully in Chapter V. FOOTNOTES are the result of reversing the positions of the normalized intervals along the x-axis. Where, in Figure 8, the interval closest to the y-axis is the interval most distant from the exit point before exit, in the regression analysis, the reverse is the case; that is, the interval closest to exit is closest to the y-axis. This was done for ease of computer analysis. 1 The negative slope and negative correlation coefficients 1 rte-5 - -. - .-._ 2As discussed in Chapter II, non -separate "voice" is often employed by justices for purposes other than influence. In this light the absence of a relationship between non -separate "voice" and differences between Court -made public policy and personal policy preference is not surprising. 3Harold J. Spaeth, An Introduction to Supreme Court Decision Making (San Francisco: Chandler Publishing Company, 1972), p. 66. 4Harold J. Spaeth, "A Theory and Methodology for the Explanation and Prediction of Supreme Court Decisions, " an unpublished manuscript. 5The tentative nature of this conclusion stems from the fact that no measure of salience presently exists for the analysis of judicial decision making. Although it is tempting to draw the inference that in comparing issue areas the issue area with the greatest percentage of "voice" is the most salient issue area for the Court as a whole, such an inference must be tenuous at best due to the necessary interpersonal comparisons of utility that this implies and the obvious circularity that such a measure of salience would entail. 6This dichotomization into Liberal and Conservative sub- groupings is based upon Harold Spaeth' 8 analysis of the Court during this time period. From computer analysis of the voting 130 131 behaviors of the individual justices, Spaeth determined each justice' 3 response to the Freedom, Equality, and New Dealism issue area. Developing a typology based upon negative or positive weighting of each issue area on the part of the individual justice, Spaeth characterized each of the justices as follows: Responses Justices Description Freedom Equality New Dealism Douglas + + + Liberal Warren + + + Liberal Goldberg + + + Liberal Fortas + + + Liberal Brennan + + + Liberal Marshall + + + Liberal Black + - + Populist White 0 0 0 Moderate Stewart 0 0 0 Moderate Clark - - + New Dealer Whittaker - - - Conservative Frankfurter - - - Conservative Harlan - - - Conservative Note: A "+" indicates support for the issue area in question, a "-" nonsupport, and a "0" neutrality. Source (with adaptations): Harold J. Spaeth, op. cit. , supra Foot- note 3, p. 68. For purposes of dichotomization Black is included in the Liberal category despite his non-support for Equality, and White and Stewart are classified as Conservative even though their response to each value is neutral, and Clark is included in the Conservative category despite his positive support for New Dealism. This dichotomy is consistent with conventional categorization. Marshall' s behaviors are exluded from this analysis due to his infrequent participation during the time period under consideration. For further discussion 132 of the typology and classification scheme see both Spaeth' s works cited in footnotes 3 and 4. 7This could also be the result of a lack of group cohesion or an indication the Liberals do not function as much as a subgroup as the Conservatives. r—- ;-'~+—+-.4-r~e~y CHAPTER V SUMMARY AND CONCLUSION The purpose of this dissertation has been to formulate and apply a theory of intracourt influence on the United States Supreme Court. As discussed in Chapter I, scholars have specu- lated about the nature of this process but with little concerted effort to formulate viable theory. Attempts to verify the specula- tions proferred by these scholars have been hindered by this lack of theory, as well as empirical constraints such as the inability to directly observe the Court in conference and by limitations inherent in inferences drawn from the public papers of former justices. By focusing upon empirically observable phenomena this disserta- tion has attempted to overcome these limitations, while at the same time providing insights into the intracourt influence process. The theory itself is a simple one. The basic assumption is that Supreme Court Justices are rational political actors and that each justice is motivated in his decision making by the desire to have Court -made public policy approximate as closely as possible 133 134 his personal policy preferences. Since the rule structures governing collegial decision making on the Supreme Court severely circum- scribe the extent to which all justices are able to have their personal policy preferences reflected in the Court-made public policy across all issues, the theory assumes that each justice attempts to influ- ence the decisions of his colleagues regarding public policy. From these basic assumptions the theory posits specific observable (overt) behaviors which are held to be influence strategies employ- able by the individual justice, and specifies the circumstances, frequency, and intensity of their use. The observable behaviors that are considered influence behaviors are analogous to those suggested by A. O. Hirschman. Hirschman contends that in the economic setting, an individual consumer has at his disposal two alternative means to influence a producer with regard to the quality of the product he produces. These are exit and/or "voice. " Hirschman specifies the relation- ship between these two influence strategies; when and under what circumstances the consumer will use either option, separately or jointly. Hirschman' 8 basic contention is that the use of these two influence strategies is directly related to the extent of quality deterioration perceived by the consumer with regard to the par- ticular product of his concern. 135 Though not completely isomorphic to the Court, Hirschman' 8 basic paradigm was considered sufficiently analogous to be appro- priately applied in the context of the Supreme Court influence process. On the Court, ”exit" is the act of leaving the majority. Since the justices in the majority determine what is to be Court-made public policy, the act of leaving the majority indicates that an individual justice considers that Court -made public policy unsatisfactory. "Voice" is the act of expressing one' s dissatisfaction directly through separate opinion writing. The use of either exit or "voice, " or both, was held to be a function of the degree of difference between Court-made public policy and personal policy preference. It was hypothesized that the frequency and intensity (extremity) of "voice" employed by a justice as a member of the majority would increase as the difference between Court -made public policy and personal policy preference increased up to the point where exit occurred. When this was tested, it was found that "voice" did, indeed, increase approaching exit (both in terms of frequency and intensity) in the aggregate and when controlling for issue area, time period, and subgroupings within time period. It was also hypothesized that "voice" after exit would be greater than "voice" before exit. When tested, this, too, was supported by the data; "voice" after exit was substantially greater than "voice" before 136 exit. For after exit "voice, " it was hypothesized that, if the difference between Court -made public policy and personal policy preference increased, the frequency and intensity of "voice" would decrease. When this hypothesis was tested it was found that it was unsupported in the aggregate and when controlling for issue area, time period, and subgroupings. This latter failure of the theory to account for "voice" behaviors after exit places the confidence one might have in the theory in doubt. This doubt may be minimized, however, if a satisfactory alternative explanation can be suggested. As conceptualized in the present formulation of the theory, "voice" behavior after exit is integrally a part of the intracourt influence process; that is, that justices, once having left the majority, are still concerned enough with the particular public policy created to continue to attempt to influence their colleagues. This, however, appears to be an erroneous assumption in light of the fact that numerous commentators maintain that the function of the dissenting opinion is an effort on the part of the dissenter to shift the arena of combat. 2 Having failed to successfully influence his immediate colleagues, the dissenter "voices" his personal policy preferences in order to influence future judges (an appeal to the "brooding spirit of the law"), 3 to influence other decision-making 9"“ —-'*--r--1m’ 137 bodies to alter the Court -made public policy through non -judicial means, 4 or to undermine or weaken the effect of the majority policy position in its application in the lower courts or in its enforcement by the executive. 5 In other words, in the context of this theory, a justice, having exhausted the effectiveness of both covert and i overt influence strategies as a member of the majority, and having " decided that all attempts have failed to alter the Court -made public policy of his immediate colleagues, exits directly from the majority bargaining process and also exits from the intracourt influence process. The justice, when exiting and ”voicing, " shifts his influence attempts from the immediate collegial setting to other decision -making arenas; that is, he exits from the intracourt influence processlo an extracourt influence process. Since the norms of the Court clearly deny personal inter- action between a justice and other political actors in other decision -making bodies with regard to matters of Court -related business, a justice is unable to employ covert influence strategies in order to accomplish his goal, and, in turn, must rely on "voice" (the dissenting opinion) as the sole vehicle for extracourt influence attempts. This explains the substantial difference between "voice" after exit and "voice" before exit. To effectuate extracourt influence a justice must rely on "voice. " 138 The absence of a positive or negative relationship between "voice" after exit and increasing differences between Court-made public policy and personal policy preference is also explainable in light of this reformulation. In the intracourt influence process held to function before exit, a justice is able to determine with relative certainty the effects of his covert and overt influence strategies. Whether "voice" is effective in altering the Court-made public policy in subsequent cases or whether "voice" employed at a certain stage is likely to effectuate policy change in future cases is readily apparent. Being integrally a part of the bargaining process provides the justice with high levels of information which allow reasonably sound judgments about the effects of various strategic choices. This readily apparent feedback as a result of high information levels, however, is not likely to be available to the dissenting justice. Whether his appeals to future justices or to other decision -making bodies will be successful or not must be left within the province of conjecture. Lacking face -to -face, personal contact with those he attempts to influence places extensive constraints upon the justices' abilities to evaluate the relative probabilities of the success or failure of his strategic choice. Thus, "voice" remains at a rela- tively high level consistent across all differences between F~'BJ~x—m§hmm 139 Court -made public policy and personal policy preference in order to insure that the dissenter' s position is heard and to increase the probabilities that the influence desired will be secured. With this reformulation in mind, one can have a reasonable degree of confidence in the theory as a satisfactory explanation of the intracourt influence process on the United States Supreme Court. Besides the necessary addition of an alternative explanation for after exit "voice, " a number of problems must be pointed out before accepting the theory and its application as valid. Most notable among these is the substantive methodological problem of accepting as interval level measurement the ostensibly ordinal measurement provided by cumulative scaling. Since the methodological questions raised by this and the justifications for its acceptance have been discussed at length in Chapter 11, further discussion is not needed here. Another problem which may bring the degree of confidence in the theory into question relates to two facets of the data set used to test the hypotheses. First, the sample of category scales used to test the hypotheses is but a small subsection, albeit a cross- section, of the total universe of category scales during the time period under consideration. It might be contended that the results obtained were more a function of the non —random sample of category 140 scales than an indication of the true nature of the intracourt influence process. This, indeed, may be the case. However, rather than reject the results of this study out of hand on this basis, such a contention merely underscores the need for future research which would consider the universe of highly refined category scales rather than just the subset used here. A second problem which is apparent when considering the ‘J~).M1.A .4. Ann-‘4“ . I I data set used is the fact that only a limited time period was con- sidered. Analysis of other time periods may reveal that the explanatory power of this theory is applicable only to the Warren Court from 1958-1969. Again, though possible, the results obtained here need not be discounted solely in light of its limited time per- spective. Future research considering other time periods under the leadership of other Chief Justices is clearly necessary. Besides extending analysis to a broader range of category scales and to different time periods, further research is needed to ascertain the consequences of intracourt influence. As discussed briefly in Chapter II, it would be of interest to ascertain the rela- tive effectiveness of the available influence strategies in effectuating change in Court —made public policy over time. Not only would research in this regard be valuable in the intracourt context, but it would also be of importance to study the effects of the extracourt 141 influence process discussed in the alternative explanation of after exit ”voice. " If such research is undertaken, major steps toward a fuller understanding of the influence process on the Supreme Court will have been made. I0.- “I“..J .- ' n ' 0m ' 115413404. ~:- FOOTNOTES 1A. O. Hirschman, Exit, Voice, and Loyalty (Cambridge: B” Harvard University Press, 1970). 2Walter Murphy, Elements of Judicial Strategy (Chicago: The University of Chicago Press, 1964), p. 60. " Y‘t“?“‘“.fi 71.2; 5 ‘- 3Chief Justice Hughes contended that a "dissent in a court of last resort is an appeal to the brooding spirit of the law, to the intelligence of a future day when a later decision may possibly correct the error to which the dissenting judge believes the court to have been betrayed. " (C. Evans Hughes, The Supreme Court of the United States [New York: Garden City Books, 1928], p. 68.) Justice Cardozo similarly holds that a dissenting opinion is an appeal to history, particularly to future judges (Benjamin Cardozo, "Law and Literature," 14 Yale Review 699, 1925, pp. 715-716). 4" a dissent can become an a l ' . . . ppea to contemporar1es -—to members of Congress, to the President and executive officials, to lower court judges, to the bar or to other interest groups, or to the public at large--to change the decision of the majority. " Murphy, op. cit. , p. 60. 5This is consistent with Justice White' s view that "the only purpose which a dissent can accomplish is to weaken the effect of the opinion of the majority and thus engender want of confidence in the conclusions of courts of last resort. " (Pollock v. Farmers Loan and Trust Company, 157 U. S. 429-608.) See also S. Sidney Ulmer, "Dissent Behavior and the Social Background of Supreme Court Justices, " 32 Journal of Politics, August, 1970, p. 581; and Joel Grossman, "Dissenting Blocs on the Warren Court: A Study in Judicial Role Behavior," 30 Journal of Politics, November, 1968, pp. 1068-1090. 6This notion of "exiting to" something else is consistent with Hirschman' s conceptualization of exit in the market setting. 142 143 For Hirschman a consumer choosing the exit option does so only after considering the probable utility to be gained from the purchase of another product. Thus exit is a dual process of deciding to cease purchase of one product and the decision to purchase another product of comparable quality. In the context of the Court, then, the decision to exit is based on the justices' evaluation of two probabilities: 1) whether remaining with the majority and "voicing" will effectuate the desired change in Court-made public policy, and 2) whether exit to an extracourt influence process will result in the desired changes in Court -made public policy either by future justices or by non -judicial means. 1"?“ 115.7 "m Iranian? BIBLIOGRAPHY W’T‘I“‘f‘ m . ' 51 BIBLIOGRAPHY Baker, Bela O.; Hardyck, Curtis D.; and Petrinovich, Lewis F. "Weak Measurements vs. Strong Statistics." 26 Educa- tional and Psychological Measurement, Summer, 1966, p. 291-309. Becker, Theodore. Political Behavior and Modern Jurisprudence. Chicago: Rand McNally, 1964. Blalock, Hubert. Causal Inferences in Nonexperimental Research. Chapel Hill: University of North Carolina Press, 1964. Cardozo, Benjamin. "Law and Literature." 14 Yale Review, 1925, pp. 699-716. Danelski, David. A Supreme Court Justice Is Appointed. New York: Random House, 1964. Downs, Anthony. An Economic Theory of Democracy. New York: Harper and Row, 1957. Friedman, Milton. Essays in Positive Economics. Chicago: University of Chicago Press, 1953. Goldman, Sheldon. "Behavioral Approaches to Judicial Decision Making: Toward a Theory of Judicial Voting Behavior. " 2 Jurimetrics Journal, March, 1971, pp. 142 -164. Grossman, Joel. "A Model for Judicial Policy Analysis. " In Grossman, Joel, and Tannenhaus, Joseph, eds. Frontiers of Judicial Research. New York: John Wiley and Sons, Inc., 1969. Pp. 405-460. "Dissenting Blocs on the Warren Court: A Study of Judicial Role Behavior. " 30 Journal of Politics, November, 1968, pp. 1068 -1090. 144 "1“. *5“. rattan-*1... 7‘3, 145 , and Tannenhaus, Joseph. Frontiers of Judicial Research. New York: John Wiley and Sons, Inc. , 1969. Hays, William. Statistics for Psychologists. New York: Holt, Rinehart, and Winston, 1963. Hirschman, A. 0. Exit, Voice, and Loyalty. Cambridge: Harvard University Press, 1970. Howard, J. Woodford. "On the Fluidity of Judicial Choice. " 62 American Political Science Review, March, 1968, pp. 43- 56. Hughes, C. Evans. The Supreme Court of the United States. New York: Garden City Books, 1928. Jahnige, Thomas, and Goldman, Sheldon, eds. The Federal Judicial System. New York: Holt, Rinehart, and Winston, Inc., 1968. Labovitz, Sanford. "The Assignment of Numbers to Rank Order Categories. " 35 American Sociological Review, June, 1970, pp. 515-524. Luce, Duncan R. , and Raiffa, Howard. Games and Decisions. New York: John Wiley and Sons, 1957. Murphy, Walter, and Pritchett, Herman. Courts, Judges, and Politics. New York: Random House, 1961. . Elements of Judicial Strategy. Chicago: The Uni- versity of Chicago Press, 1964. Nagel, Stuart. ”Political Party Affiliation and Judges' Decisions. " 55 American Political Science Review, December, 1961, pp. 843 -850. Olson, Mancur. "The Relationship Between Economics and Other Social Sciences. " In Lipset, Seymour Martin, ed. Politics and the Social Sciences. New York: Oxford University Press, 1969. Pp. 137-162. WMM- Wu,- 1 I 146 Pollack v. Farmers Loan and Trust Company, 158 US 429 (1894). Rohde, David W. "Policy Goals and Opinion Coalitions in the Supreme Court. " 26 Midwest Journal of Political Science, May, 1972, pp. 208-224. Strategy and Ideology: The Assignment of Majority Opinions in the United States Supreme Court. Unpublished Ph.D. Dissertation, University of Rochester, 1971. Schlesinger, Joseph. Ambition and Politics: Political Careers in the United States. Chicago: Rand McNally, 1966. Schmidhauser, John. "Judicial Behavior and the Sectional Crisis of 1837-1860. " 23 Journal of Politics, November, 1961, pp. 615-640. Schubert, Glendon. "Ideologies and Attitudes, Academic and Judicial. " 29 Journal of Politics, February, 1967, pp. 3- 40. Judicial Policy Making. Chicago: Scott, Foresman, and Company, 1965. Quantitative Analysis of Judicial Behavior. Glencoe: The Free Press, 1959. . The Judicial Mind. Evanston: Northwestern University Press, 1965. Senders, V. L. Measurement and Statistics. London: Oxford University Press, 1958. Siegel, Sidney. Nonparametric Statistics. New York: McGraw- Hill, 1956. "Social Science Approaches to the Judicial Process. " 79 Harvard Law Review, June, 1966, pp. 1551-1628. Spaeth, Harold J. "A Theory and Methodology for the Explanation and Prediction of Supreme Court Decisions. " An unpub- lished manuscript. 147 An Introduction to Supreme Court Decision Making. San Francisco: Chandler Publishing Company, 1972. . "Unidimensionality and Item Invariance in Judicial Scaling." 10 Behavioral Science, July, 1965, pp. 290- 303. Stevens, S. S. "Mathematics, Measurement, and Psychophysics." In Stevens, S. S. , ed. Handbook of Experimental Psy- chology. New York: John Wiley and Sons, Inc. , 1951. Tannenhaus, Joseph. "The Cumulative Scaling of Judicial Decisions. " 79 Harvard Law Review, June, 1966, pp. 1583- 1594. 1HNV'I‘A— I .l Tufte, Edward R. "Improving Data Analysis in Political Science. " 21 World Politics, July, 1969, pp. 641-654. Ulmer, S. Sidney. ”Dissent and the Social Backgrounds of Supreme Court Justices. " 32 Journal of Politics, August, 1970, pp. 580-598. "Sub ~group Formation in the United States Supreme Court. " In Jahnige, Thomas, and Goldman, Sheldon, eds. The Federal Judicial System. New York: Holt, Rinehart, and Winston, Inc., 1968. Pp. 222-229. "Supreme Court Behavior and Civil Rights. " 13 Western Political Quarterly, June, 1960, pp. 288-311. Vines, Kenneth. "The Judicial Role in the American States. " In Grossman, Joel, and Tannenhaus, Joseph. Frontiers of Judicial Research. New York: John Wiley and Sons, Inc. , 1969. Pp. 461-488. Wells, Richard, and Grossman, Joel. "The Concept of Judicial Policy Making. " In Jahnige, Thomas, and Goldman, Sheldon, eds. The Federal Judicial System. New York: Holt, Rinehart, and Winston, Inc., 1968. Pp. 294-304. APPENDIX CATEGORY SCALES uni-- ' ' W_ T»; .6%-1- "“ .‘nA-x'b I APPENDIX CATEGORY SCALES This Appendix presents the 22 category scales employed in this application of the theory of intracourt influence. For each scale, the scale title is presented followed by a list of the cases used in the order in which they appear in the scale. 1: Contempt of Court Re Green 8 L. Ed. 2d. 198 1959 Holt v. Virginia 14 L. Ed. 2d. 290 1965 Shillitani v. U.S. 16 L. Ed. 2d. 622 1966 Re McConnell 8 L. Ed. 2d. 434 1962 Bloom v. Illinois 20 L. Ed. 20. 522 1968 Harriss v. U.S. 15 L. Ed. 2d. 240 1965 Levine v. U.S. 4 L. Ed. 2d. 989 1960 Brown v. U.S. 3 L. Ed. 2d. 609 1959 U.S. v. Barnett 12 L. Ed. 2d. 23 1964 Frank v. ULS. 23 L. Ed. 2d. 162 1969 Cheff v. Schnackenberg 16 L. Ed. 2d. 629 1966 Unger v. Sarafite 11 L. Ed. 2d. 921 1964 2: First Amendment NAACP v. Alabama 12 L. Ed. 2d. 325 1964 Bond v. Floyd 17 L. Ed. 2d. 235 1966 Red Lion Broadcasting v. FCC 23 L. Ed. 2d. 371 1969 Brandenburg v. Ohio 23 L. Ed. 2d. 430 1969 Bates v. Little Rock 4 L. Ed. 2d. 480 1960 Carrol v. Princess Anne 21 L. Ed. 2d. 325 1968 Mills v. Alabama 16 L. Ed. 2d. 484 1966 Freedman v. maryland 13 L. Ed. 2d. 649 1965 Louisiana v. NAACP 6 L. Ed. 2d. 301 1961 Talley v. California 4 L. Ed. 2d. 559 1960 UMW v. Illinois State Bar 19 L. Ed. 2d. 426 1967 148 ‘ _‘h‘xfit—X‘T ._ . 5?. imam.” ' 149 First Amendment (cont'd) Wood v. Georgia 8' Brotherhood of Railroad Train- men v. Virginia Dombrowski v. Pfister Shelton v. Tucker Re Sawyer NAACP v. Bulton Gibson v. Florida Committee Times Film Corp. v. Chicago Lathrop v. Donohue Legislative Investigation Raley v. Ohio Flaxer v. U.S. Scull v. Virginia Slagle v. Ohio Dombrowski v. Gojack v. U.S. De Gregory v. Attorney General Silber v. Deutch v. U.S. Price v. U.S. Yellin v. U.S. Slagle v. Ohio Morgan v. Ohio Uphaus v. Wyman Barenblatt v. U.S. McPhail v. U.S. Wilkinson v. U.S. Braden v. U.S. Wheeldin v. Wheeler Eastland U.S. Security Risks Cramp v. Board of Public Instruction Schneider v. Smith Greene v. McElroy Baggett v. Bullitt Greene v. U.S. Whitehill v. Elkins Vitarelli v. Seaton Elfbrandt v. Russell Keyishian v. Board of Regents Fleming v. Nestor Re Anastaplo Konisberg v. State Bar Nelson v. Los Angeles Cafeteria Workers v. McElroy Nostrand v. Little Taylor v. McElroy 12 14 ONWQOWU'I _8—8—3 —8 Ommmwwwmommm mmmmwww —L KN-F-m-FQQ-bQONWkDJNWKO-Q 1 1 t" o HHHHHL—‘HH O O O O O O O HHHHHHHHHHH?‘ HHHHHHL“ HHHHHHHHHHBHHEt‘H Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. 2d. 569 2d. 89 2d. 22 2d. 231 2d.1473 2d. 405 2d. 929 2d. 403 2d.1191 2d.1344 2d. 183 2d. 865 2d. 277 2d. 577 2d. 870 2d. 292 798 963 240 2d. 778 2d. 277 2d.1344 2d.1090 2d.1115 2d. 136 2d. 633 2d. 623 2d. 605 2d. 2d. 2d. 2d. 285 2d. 799 2d.1377 2d. 377 2d. 576 2d. 228 2d.1012 2d. 321 2d. 629 2d.1435 2d. 105 Zd. 494 2d.123O 2d. 892 2d.1528 (1962) 1964 1965 1960 1959 1963 1963 1961 1961 1959 1958 1959 1961 1967 1966 1966 1962 1961 1962 1963 1961 1959 1959 1959 1960 1961 1961 1963 (1961) 1968 1958 1964 1964 1967 1959 1966 1967 1960 1961 1961 1960 1961 1960 1959 150 Federal Internal Security Legislation U.S. v. Laub Travis v. U.S. Albertson v. SACB Communist Party v. Catherwood ‘ Lamount v. Postmaster General Lincoln Brigade v. SACB American Committee v. SACB Noto v. U.S. Travis v. U.S. Aptheker v. Secretary of State U.S. v. Robel U.S. v. Brown Communist Party v. SACB Scales v. U.S. Killian v. U.S. Zemel v. Rusk Dennis v. U.S. Rabinowitz v. Kennedy Obscenity (state) Stanley v. Georgia Smith v. California Kingsley v. Regents Bantam Books v. Sullivan Interstate Circuit v. Redrup v. New York A Quantity of Books v. Kansas A Book v. Attorney General Jacobellis v. Ohio Mishkin v. New York Ginsberg v. New York Sit-ins Wright v. Georgia Barr v. Columbia Peterson v. Greenville Robinson v. Florida Garner v. Louisiana Boynton v. Virginia Shuttlesworth v. Birmingham Lombard v. Louisiana Griffin v. Maryland Bell v. Maryland Barr v. Columbia Bouie v. Columbia Hamm v. Rock Hill Brown v. Louisiana Dallas 17 1 1 Add .54.; dm#¥¥ GU14 HHHHHHEHL—i bt'itflt'it" 20 18 12 16 12 20 H555 .0. Ebb-(HE Htflfit‘t’t" HHHHHb'HHHHHHHH Ed. Ed . Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 526 536 165 919 398 46 39 836 340 892 508 484 625 782 256 179 973 542 205 1512 584 225 515 809 793 195 549 766 525 771 207 206 555 558 754 822 766 894 500 657 151 Protest Shuttlesworth v. Birmingham Shuttlesworth v. Birmingham Gregory v. Chicago Cox v. Louisiana Edward v. South Carolina Tinker v. Des Moines Federal Employees Union v. Logan Valley Plaza Cox v. Louisiana Street v. New York Walker v. Birmingham Adderley v. Florida Cameron v. Johnson U.S. v. O'Brien Reapportionment Baker v. Carr Gray v. Sanders Reynolds v. Sims Davis v. Mann Roman v. Sincock Moore v. Ogilvie Swan v. Adams Fortson v. Toombs Maryland Committee v. Tawes Wesberry v. Sanders W.M.C.A. v. Lomenzo Lucas v. Colorado General Assembly Avery v. Midland Kirkpatrick v. Priesler Wells v. Rockefeller Fortson v. Morris Wright v. Rockefeller Fortson v. Dorsey Burns v. Richardson Sailors v. Kent Board Dusch v. Davis Moody v. Flowers Bugging Silverman v. U.S. Katz v. U.S. Alderman v. U.S. Lee v. Florida Berger v. New York Butemko v. U.S. Lopez v. U.S. Pugach v. Dollinger 15 22 22 13 21 20 13 22 18 20 20 HHL—‘t'it‘t'ib‘ tidbit-1171517" HHHHHHHHHHH HEHHHHHL—‘L—‘t‘ib bet‘bfit‘t“ Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. 2d. 176 2d. 162 2d. 134 2d. 471 2d. 697 2d. 731 2d. 603 2d. 487 2d. 572 2d.1210 26. 149 2d. 182 2d. 672 2d. 663 2d. 821 2d. 506 2d. 609 2d. 620 2d. 1 2d. 501 2d. 527 2d. 595 2d. 981 2d. 568 2d. 632 2d. 45 2d. 519 2d. 535 2d. 512 2d. 401 2d. 376 2d. 650 2d. 656 2d. 643 2d. 734 2d. 576 2d. 176 2d.1166 2d.1040 2d. 176 2d. 462 20. 678 . J ‘41 .r was ”AW ’ .1 152 Bugging (cont'd) Osborn v. U.S. 17 L. Ed. 2d. 394 $1966; Lanza v. New York 8 L. Ed. 2d. 384 1962 Search and Seizure (1958-1961) Marcus v. Search Warrant 6 L. Ed. 2d.1127 1961 Henry v. U.S. 4 L. Ed. 2d. 134 1959 Chapman v. U.S 5 L. Ed. 2d. 828 1961 Mapp v. Ohio 6 L. Ed. 2d.1081 1961 Elkins v. U.S. 4 L. Ed. 2d.1669 1960 Rios v. U.S. 4 L. Ed. 2d.1688 1960 Ohio v. Price 4 L. Ed. 2d.1708 1960 Frank v. Maryland 3 L. Ed. 2d. 877 1959 Abel v. U.S. 4 L. Ed. 2d. 668 1960 Wilson v. Schnettler 5 L. Ed. 2d. 620 1961 Draper v. U.S. 3 L. Ed. 2d. 327 1959 Jones v. U.S. 4 L. Ed. 2d. 697 1960 Search and Seizure (1962-1966) Preston v. U.S. 11 L. Ed. 2d. 771 1964 Stanford v. Texas 13 L. Ed. 2d. 431 1965 One Plymouth Sedan v. Penn. 14 L. Ed. 2d. 170 1965 Aguilar v. Texas 12 L. Ed. 2d. 723 1964 Camara v. Municipal Court 18 L. Ed. 2d. 930 1967 See v. Seattle 18 L. Ed. 2d. 943 1967 Beck v. Ohio 13 L. Ed. 2d. 142 1964 Sun v. U.S. 9 L. Ed. 2d. 441 1963 Ker v. California 10 L. Ed. 2d. 726 1963 Cooper v. California 17 L. Ed. 2d. 730 1967 U.S. v. Ventresca 13 L. Ed. 2d. 684 1965 Warden v. Hayden 18 L. Ed. 2d. 782 1967 Lewis v. U.S. 17 L. Ed. 2d. 312 1966 Search and Seizure (1967-1969) Sabbath V. 0.3. 20 L. Ed. 2d. 828 1968 Sibron v. New York 20 L. Ed. 2d. 917 1968 Davis v. Mississippi 22 L. Ed. 2d. 676 1969 Simmons v. U.S. 19 L. Ed. 2d.1247 1968 Chimel v. California 23 L. Ed. 2d. 685 1969 Mancusi v. De Forte 20 L. Ed. 2d.1154 1968 Bumper v. North Carolina 20 L. Ed. 2d. 797 1968 Harrison v. U.S. 20 L. Ed. 2d.1047 1968 Massachusetts v. Painten 19 L. Ed. 2d. 770 1968 Kaufman v. U.S. 22 L. Ed. 2d. 227 1969 Spinelli v. U.S. 21 L. Ed. 2d. 637 1969 Terry v. Ohio 20 L. Ed. 2d. 889 1968 Peters v. New York 20 L. Ed. 2d. 917 1968 Harris v. U.S. 19 L. Ed. 2d.1067 1968 Frazier v. Cupp 22 L. Ed. 2d. 684 1969 15. 153 Self-incrimination Leary v. U.S. U.S. v. Covington Sanitation Men v. sioner Garderner v. Broderick Murphy v. Waterfront Commission Haynes v. U.S. Marchetti v. U.S Grosso v. U.S. Stoner v. California Chapman v. California Stevens v. Marks Griffin v. California Mathis v. U.S. California v. Orozco v. U.S. Stewart v. U.S. Garrity v. New Jersey Fahy v. Connecticut Miranda v. Arizona Malloy v. Hogan Spevack v. Klein Cohen v. Hurley Hutchinson v. U.S. Schmerber v. California Mills v. Louisiana Piemonte v. U.S. Shotwell Manufacturing Co. v. U.S. Namet v. U.S. U.S. v. Welden U.S. v. Gainey Campbell Painting Corp. v. Reid U.S. v. Blue Commis- Stewart Anti-trust Radiant Burners v. Peoples Gas Co. U.S. v. United Shoe Machinery Corp. FTC v. Universal Bundle FTC v. Brown Shoe Carnation Co. v. Conf. Leh v. Corp. U.S. v. R. Co. Pacific General Petroleum Boston and Maine 10 20 18 16 15 15 13 t"'L'* L—‘HHH HHHUHEHHHF‘HHHHHHHHHHHH tit" (fit-1 Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. 2d. 2d. 57 94 2d.1089 2d.1082 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 678 925 889 906 856 705 724 106 581 694 511 84 562 171 694 655 574 156 157 908 2d.1193 2d.1028 2d. 2d. 2d. 2d. 357 278 152 658 2d.1094 2d. 2d. 2d. 2d. 2d. 2d. 2d. 2d. 510 358 562 749 587 709 134 728 ( é ( ( ( 1961) 1968i 1967 1966 1966) 1965) 1965) j‘-m (An-m 15 154 Anti-trust (cont'd) U.S. v. Ward Baking Co. 11 Continental Ore Co. v. Union Carbide 8 FTC v. Anheuser-Busch, Inc. 4 FTC v. Simplicity Pattern Co. 3 Melrose Distillers v. U.S. 3 FTC v. Jantzen 18 Walker Process Equipment v. Food Machinery and Chemical U.S. v. Wise U.S. v. General Motors 1 U.S. v. RCA Klos v. Broadway-Hale Stores 3 .5 \flmmm Ht" (:1 tit—'t“ t‘1 HHEHHHBF‘L—i tit'it‘ HHHHBHHHU FTC v. Consolidated Foods 14 U.S. v. Schwinn 18 United Shoe v. Hanover Shoe 20 FTC v. Texaco 21 Citizen Publishing Co. v. U.S. 22 FTC v. Mary Carter Paint Co. 15 U.S. v. Sealy 18 U.S. v. Singer Manufacturing Co. 10 U.S. v. Borden Co. 8 Case-Swayne v. Sunkist 19 FTC v. Broch and Co. 7 U.S. v. Parke Davis 4 U.S. v. Loew's Inc. 9 FTC v. Sun Oil Co. 9 FTC v. Colgate-Palmolive 13 FTC v. Borden Co. 16 Utah Pie Co. v. Continental Baking 18 U.S. v. Phosphate Export Assn. 21 FTC v. Meyer 19 Albrecht v. Herald Co. 19 Silver v. New York Stock Exchange 10 Perma Mufflers v. Inter- national Parts 20 U.S. v. National Dairy 9 Minnesota Mining and Manu- facturing v. New Jersey Wood Finishing Co. 14 Atlantic Refining Co. v. FTC 14 U.S. v. Container Corp. 21 t1 o HHHUH O O t“ 0 Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. 2d. 743 2d. 777 2d.1385 2d.1079 2d. 800 2d. 11 2d. 2d. 2d. 247 590 415 2d. 354 2d. 741 2d. 95 2d.1249 2d.1231 2d. 394 2d. 148 2d. 128 2d.1238 823 627 621 353 305 2d. 11 466 904 153 406 544 2d.1222 998 589 982 2d. 561 2d. 2d. 405 445 506 2d. 2d. (1964) E 1962 1960 1959 1959 1967 1965 1962 1966 1959 1959 1965 1967 1968 1968 1969 1965 1967 1963 1962 1967 1962 1960 1962 1963 1965 1966 i (1967) 1 1968 1968 1968 E (1963) 1 1967 1963 1 (1965) E 1965 1969 3 12”” ‘ "ic‘nwzrx 1 . 155 .Anti-trust (cont'd) Perkins v. Standard Oil 23 Denver and Rio Western Rail- road Co. v. Railroad F‘ t1: C1: 0 2d. 599 (1969) Trainmen 18 L. Ed. 2d. 905 $1967} U.S. v. Grimwell Corp. 16 L. Ed. 2d. 778 1966 Fortner Enterprises v. U.S. Steel 22 L. Ed. 2d. 495 1969 FTC v. Broch and Co. 4 L. Ed. 2d.1124 1960 Poller v. CBS 7 L. Ed. 2d. 458 1962 Simpson v. Union Oil 12 L. Ed. 2d. 98 1964 White Motor Co. v. U.S. 9 L. Ed. 2d. 738 1963 Kelley v. Kosuga 3 L. Ed. 2d. 475 1959 Tampa Electric Co. v. Nashville Coal Co. 5 L. Ed. 2d. 580 é1961; U.S. v. Atlantic Refining Co. 3 L. Ed. 2d.1054 1959 Eastern Railroad Conference v. Noerr Motor Freight, Inc. 5 L. Ed. 2d. 464 (1961) Sunkist Growers v. Winkler and Smith 8 L. Ed. 2d. 305 (1962) Mergers U.S. v. Diebold 8 L. Ed. 2d. 176 (1962) Levin v. Mississippi River Fuel Corp. 17 L. Ed. 2d. 834 (1967) U.S. v. 1st City National Bank v 18 L. Ed. 2d. 151 (1967) Maryland and Virginia Mil Producers Assn. v. U.S. 4 L. Ed. 2d. 880 (1960) FTC v. Proctor and Gamble Co. 18 L. Ed. 2d. 303 1967 SEC v. New England Electric 19 L. Ed. 2d.1042 1968 U.S. v. Pabst Brewing Co. 16 L. Ed. 2d. 765 1966 U.S. v. 3rd National Bank 19 L. Ed. 2d.1015 1968 U.S. v. El Paso Gas Co. 12 L. Ed. 2d. 12 1964 Brown Shoe Co. v. U.S. 8 L. 3d. 2d. 510 1962 SEC v. New England Electric 16 L. Ed. 2d. 456 1966 California v. FPC 8 L. Ed. 2d. 54 1962 Edison v. El Paso Gas Co. 17 L. Ed. 2d. 814 1967 Utah Public Service Commis- sion v. El Paso Natural Gas 23 . Ed. 2d. 474 E1969; U.S. v. Van's Grocery Co. 16 . Ed. 2d. 555 1966 U.S. v. Philadelphia National Bank 10 U.S. v. Continental Can Co. 12 0 t3 CL 0 2d. 915 1963 2d. 953 1964 HUHF‘L“ HE 131 Q. C U.S. v. 1st National Bank 12 . Ed. 2d. 1 1964 U.S. v. Alcola 12 . Ed. 2d. 314 1964 FTC v. Dean Foods Co. 16 . Ed. 2d. 802 (1966) 156 Mergers (cont'd) Baltimore and Ohio Rail Co. v. U.S. 18 L. Ed. 2d. 159 (1967) International Boxing Club v. U.S. 3 L. Ed. 2d. 270 {1958; U.S. v. Dupont 6 L. Ed. 2d. 318 1961 Penn Central Merger 19 L. Ed. 2d. 723 1968 Minneapolis and St. Louis Rail Co. v. U.S. 4 L. Ed. 2d. 223 (1959) Gilbertvill Trucking Co. v. U.S. 9 L. Ed. 2d. 177 21962; Penn Central Merger 19 L. Ed. 2d. 723 1968 no.443 Transportation Regulation U.S. v. Seaboard Air Line 4 L. Ed. 2d. 25 (1959) Union Pacific Rail Co. v. U.S. 4 L. Ed. 2d. 776 (1959) American Trucking Assn. V. U.S. 4 L. Ed. 2d.1527 (1959) Porter Co. v. Central Vermont Rail Co. 6 L. Ed. 2d. 284 (1961) New Jersey v. New York, Susquehanna and Western Rail Co. 9 L. Ed. 2d. 541 (1963) ICC v. New York, New Haven and Hartford Rail Co. 10 L. Ed. 2d. 108 (1963) Massachusetts Trustees v. U.S. 12 L. Ed. 2d. 268 (1964) Consolo v. Federal Maritime Commission 16 L. Ed. 2d. 131 (1966) U.S. v. Atcheson, Topeka and Santa Fe Rail Co. 18 L. Ed. 2d. 803 (1967) Federal Maritime Commission v. Svenska 19 L. Ed. 2d.1071 (1968) ICC v. Atlantic Coast Line Rail Co. 16 L. Ed. 2d. 109 (1966) Atcheson, Topeka, and Santa Fe Rail Co. v. Reddish 7 L. Ed. 2d. 147 (1961) ICC v. Frisco Transportation Co. 3 L. Ed. 2d. 172 (1958) Illinois Central Rail Co. v. Norfolk and Western Rail Co. 17 L. Ed. 2d. 162 1966 U.S. v. Drum 7 L. Ed. 2d. 360 1962 CAB v. Deta Air Lines 6 L. Ed. 2d. 869 1960 ICC v. J-T Transport Co. 7 L. Ed. 2d. 147 1961 U.S. v. Atcheson, Topeka and Santa Fe Rail Co. 18 L. Ed. 2d. 847 (1967) '~vaumv .‘ '4' .—l- TL... "- — V 17: 18: 157 Transportation Regulation (cont'd) Hewitt-Robbins v. Eastern Freightways 9 Willis Shaw Frozen Express v. U.S. 12 Red Ball Motor Freight, Inc. v. Shannon 12 T.I.M;E. v. U.S. 3 Arrow Transportation Co. v. Southern Rail Co. 10 Chicago, Milwaukee, St. Paul and Pacific Rail Co. v. U.S. 6 Volkswagon v. Federal Mar- itime Commission 19 American Commercial Lines v. Louisville and Nashville Rail Co. 20 Western Pacific Rail Co. v. U.S. 15 U.S. v. Montgomery 11 Chicago Board of Trade v. U.S. 11 Baltimore and Ohio Rail Co. v. Aberdeen and Rockfish Rail Co. 21 L. t"I t‘11:" t1 t."l O O O O t1 1:1 o o HHF‘H Public Utility Regulation FPC v. Southern California Edison 11 FCC v. Schreiber 14 United Gas Pipe Line Co. v. FPC 17 FPC v. Sunray Dx Oil Co. 20 Texaco v. Superior Court 6 FPC v. Tennessee Gas Transmission Co. 9 Texas Gas Transmission Corp. v. Shell Oil Co. 4 O 3 5 3 9 1 U.S. v. Southwest Cable Co. 2 Atlantic Refining Co. v. Public Service Commission United Gas v. Gallery Properties 1 California v. Lo-Vaca Gathering Co. TVA v. Kentucky Utilities FPC v. Hunt Washington Public Power v. FPC O O O HHHHHHFHHHHHHL“ _s (I) Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. Ed. 2d. 142 2d. 211 2d. 341 2d. 952 2d. 52 2d. 772 2d.1090 2d.1289 2d. 294 2d. 797 2d. 788 2d. 219 2d. 638 2d. 383 2d. 181 2d. 388 2d. 584 2d. 199 2d.1208 2d.1001 2d.1312 2d. 284 2d. 357 2d. 787 2d. 878 2d. 869 (1962) (1964) 113333 (1963) (1961) (1968) (1968) 513223 (1964) (1968) 1965 1966 1968 1961 (1962) 83283 $1964; (1959) (1965) 1965 E19682 1964 (1967) ‘51?“ ‘rfi'fi—‘W .12-Tweaking- -.. _ ., I. .3 -' 158 Public Utility Regulation (cont'd) National Coal Assn. v. Trans- continental Gas Pipe Line 5 FPC v. Union Electric Co. 14 Northern National Gas Co. v. Kansas State Commission Sunray Oil Co. v. FPC Sun Oil Co. v. FPC California v. FPC FPC v. Tuscarora Indian Nation FPC v. Memphis Light Power Reactor Development Co. v. Electrical Workers L. Ed. 2d. 377 11961; L. Ed. 2d. 239 1965 Ed. 2d. 601 1963 . Ed. 2d.1623 1960 Ed. 2d.1639 1960 . Ed. 2d. 357 1963 L. Ed. 2d. 584 E1960? EFL-‘b‘ L. Ed. 2d. 153 1958 L. Ed. 2d. 924 (1961) —8 ON W4> O-h-p-kO Pre-emption , ' Hotel Employees v. Sax Enterprise 3 Radio and T.V. Technicians v. Broadcast Service of t“ o :13 Q. 0 2d. 289 (1958) Mobile, Inc. 13 . Ed. 2d. 789 (1965) Plumber's Union v. Door County 3 . Ed. 2d. 872 1958 Williams v. Lee 3 . Ed. 2d. 251 1959 Liner v. Jafco 11 . Ed. 2d. 347 1964 Laborer's Union v. Curry 9 . Ed. 2d. 514 1963 AVCO v. Aero Lodge 20 . Ed. 2d. 126 1968 Teamster's Union v. Guiver 3 . Ed. 2d. 312 1959 Journeymen v. Borden 1O . Ed. 2d. 638 1963 Iron Workers v. Perko 1O . Ed. 2d. 646 1963 United Mine Workers v. Gibbs 16 2d. 218 1966 Donovan v. Dallas 12 . Ed. 2d. 409 1964 San Diego Unions v. Garmon 3 . Ed. 2d. 775 1959 0 C13 C14 0 2d.1109 1960 2d. 582 1966 De Veau v. Braisted 4 Linn v. Plant Gaurd Workers 15 Burlington Truck Lines b“ t“ tit-1H t4 .t'dt‘t‘b‘rt‘tdt‘lr‘bt‘t‘t“ 1‘.“ L11 LT: 04 DJ 0 0 v. U.S. 9 . Ed. 2d. 207 (1962) Wallis v. Pan American Petroleum 16 . Ed. 2d. 369 1966 Smith v. Evening News Assn. 9 . Ed. 2d. 246 1962 Dowd Box Co. v. Courtney 7 . Ed. 2d. 483 1962 Hanna Mining Co. v. Marine Engineers 15 . Ed. 2d. 254 (1965) Retail Clerks v. Schermerhorn 11 . Ed. 2d. 179 (1963) 20: 21: 22: 159 Taxation of Gifts A oowm-h-sxoxom-ptooowaoxox Tellier Lester Commissioner v. Commissioner v. U.S. v. Kaiser Peurifoy v. Commissioner Commissioner v. Stidger U.S. v. Correll Stanton v. U.S. Rudolph v. U.S. U.S. v. Gilmore U.S. v. Patrick U.S. v. Zacks Commissioner v. Duberstein Commissioner v. Bilder Cammarano v. U.S. Davis v. U.S. A—A _L Comity Turner v. Memphis 7 Willmingham v. Morgan Zwickler v. Koota Georgia v. Rachel England v. Louisiana Board Brady v. Maryland Alleghaney v. Marshada Irvin v. Dowd Greenwood v. Peacock Louisiana Power and Light Co. v. Thibodaux Cleary v. Bolger Harrison v. NAACP United Gas Pipe Line Co. v. Ideal Cement Co. Martin v. Creasy Dresner v. Tallahassee A Avid WOW mom» 4 Judicial Administration Pennsylvania Rail Co. v. U.S. 4 Stern v. South Chester Tube Co. 20 Glover v. St. Louis and Santa Fe Rail Co. 21 Western Union v. Pennsylvania 7 Will v. U.S. Florida Lime and Avacado Growers v. Jacobsen 4 O O O I O O HHHHHHHE‘HHHL—‘bt‘b‘ Hb‘t" Eli—1t" HF‘t‘ibt—‘HL—‘t‘t‘i bibs—1H5 Ed. Ed. Ed. Ed. Ed. 2d. 185 2d. 306 2d.1233 2d. 30 2d. 53 2d. 537 2d.1218 2d. 484 2d. 570 2d. 580 2d. 128 2d.1218 2d. 65 2d. 462 2d. 335 2d. 2d. 2d. 2d. 762 396 444 925 2d. 440 2d. 215 2d.1163 2d. 900 2d. 944 2d.1058 2d. 390 2d.1152 2d. 623 2d.1186 2d. 208 2d.1165 2d. 117 519 139 305 568 2d. 2d. 2d. 2d. 160 Judicial Administration (cont'd) U.S. v. 1st National City Bank 1 U.S. v. American-Foreign Steamship Corp. U.S. v. Mersky Prima Paint v. Flood and 3 Ed. 2d. 365 (1965) 4 4 Conklin 18 15 8 2d.1491 $19603 Ed. 2d. 423 1960 0 Cd Q4 0 0 {33 Q. o 2d.1270 1967 2d. 194 1965 2d. 671 1962 2d. 643 1967 o 131 Q. 0 Swift and Co. v. Wickham Glidden Co. v. Zdanok Moody v. Flowers 18 WW DID-4 o o Kramer v. Carribean Mills 23 . Ed. 2d. 9 1969 U.S. v. King 23 . Ed. 2d. 52 1969 Steelworkers v. Bouligny 15 . Ed. 2d. 217 1965 Idlewild Liquor Corp. v. Epstein 8 . Ed. 2d. 794 11962; Banks v. California 23 . Ed. 2d. 653 1969