1 See, e.g., Harold Dwight Lasswell, Power and Personality ch. 4 (New York: W. W. Norton, 1948); Glendon Schubert, ed., Judicial Behavior: A Reader in Theory and Research (Chicago: Rand McNally, 1964); Goldman, Sheldon, Voting Behavior on the United States Courts of Appeals Revisited, 69 Am. Pol. Sci. Rev. 491 (1975); Adamany, David W., The Party Variable in Judges' Voting: Conceptual Notes and a Case Study, 63 Am. Pol. Sci. Rev. 57 (1969); Stuart S. Nagel, Comparing Elected and Appointed Judicial Systems (Beverly Hills, Cal.: Sage Publications, 1973); Grossman, Joel B., Social Backgrounds and Judicial Decision-making, 79 Haw. L. Rev. 1551 (1966); Jaros, Dean & Canon, Bradley C., Dissent on State Supreme Courts: The Differential Significance of the Characteristics of Judges, 15 Midwest J. Pol. Sci. 322 (1971); Richard Johnston, Supreme Court Voting Behavior: A Comparison of the Warren and Burger Courts, in Robert L. Peabody, ed., Cases in American Politics (New York: Praeger, 1976); Gibson, James L., From Simplicity to Complexity: The Development of Theory in the Study of Judicial Behavior, 5 Pol. Behav. 7 (1983).Google Scholar

2 But see Baum, Lawrence, Judicial Specialization, Litigant Influence, and Substantive Policy: The Court of Customs and Patent Appeals, 11 Law & Soc'y Rev. 823 (1977); Vines, Kenneth N., Federal District Judges and Race Relations Cases in the South, 26 J. Pol. 338 (1964); Frank J. Sorauf, The Wall of Separation: The Constitutional Politics of Church and State (Princeton, N.J.: Princeton University Press, 1976); Tate, C. Neal, Personal Attribute Models of the Voting Behavior of U.S. Supreme Court Justices: Liberalism in Civil Liberties and Economic Decisions, 1946–1978, 75 Am. Pol. Sci. Rev. 355 (1981).Google Scholar

3 See, however, Canon, Bradley C., Characteristics and Career Patterns of State Supreme Court Justices, 45 St. Gov't 34 (1972); id., The Impact of Formal Selection Processes on the Characteristics of Judges–-Reconsidered, 6 Law & Soc'y Rev. 579 (1972); Henry Robert Glick, Supreme Courts in State Politics: An Investigation of the Judicial Role (New York: Basic Books, 1971); Kenneth N. Vines, The Judicial Role in the American States: An Exploration, in Joel B. Grossman & Joseph Tanenhaus, eds., Frontiers of Judicial Research (New York:Google Scholar

4 See Kagan, Robert A., Cartwright, Bliss, Friedman, Lawrence M., & Wheeler, Stanton, The Business of State Supreme Courts, 1870–1970, 30 Stan. L. Rev. 121 (1977); id., The Evolution of State Supreme Courts, 76 Mich. L. Rev. 961 (1978); Friedman, Lawrence M. et al., State Supreme Courts: A Century of Style and Citation, 33 Stan. L. Rev. 773 (1981);Note, Courting Reversal: The Supervisory Role of State Supreme Courts, 87 Yale L.J. 1191 (1978).Google Scholar

5 Thus unpublished opinions were excluded. We also excluded short per curiam or memorandum opinions–-those less than one printed page. The total sample (5,904) is smaller than the original target figure (18 cases × 21 sample years × 16 courts = 6,048 cases) because two sample states, South Dakota and Idaho, were not admitted to the union until 1889 and 1890 and therefore provided no cases for the 1870–85 sample years.Google Scholar

6 The Dictionary of American Judges, published in 1955, was useful, but covered only judges sitting in that year.Google Scholar

7 The term judges, when used in the rest of this article, will mean state supreme court judges unless stated otherwise.Google Scholar

8 See Meador, Daniel J., German Appellate Judges: Career Patterns and American-English Comparisons, 67 Judicature 16 (June-July 1983); id., English Appellate Judges from an American Perspective, 66 Geo. L. J. 1349 (1978); John Philip Dawson, The Oracles of the Law (Ann Arbor: University of Michigan Law School, 1968); Damaska, Mirjan, Structures of Authority and Comparative Criminal Procedure, 84 Yale L.J. 480, 509, 522 (1975); Henry W. Ehrmann, Comparative Legal Cultures 70–77 (Englewood Cliffs, N.J.: Prentice-Hall, 1976); Henry Abraham, The Judicial Process (3d ed. New York: Oxford University Press, 1975).Google Scholar

9 See, e.g., Richard E. Ellis, The Jeffersonian Crisis: Courts and Politics in the Young Republic (New York: Oxford University Press, 1971); Lawrence M. Friedman, A History of American Law pt. 3, ch. 2 (New York: Simon & Schuster, 1973).Google Scholar

10 Bruce A. Ackerman, Private Property and the Constitution 187 (New Haven, Conn.: Yale University Press, 1977).Google Scholar

11 See data presented in Abraham, supra note 8, at ch. 2.Google Scholar

12 These were Burton, Clark, Douglas, Fortas, Frankfurter, Goldberg, Jackson, Reed, White, and Warren himself.Google Scholar

13 Of the nine judges on the present Court, however, only Powell, Rehnquist, and White were wholly without judicial experience when appointed. See John Cound, The Justices of the Supreme Court, app. A in William B. Lockhart, Yale Kamisar, & Jesse H. Choper, The American Constitution (5th ed. St. Paul, Minn.: West Publishing Co., 1981).Google Scholar

14 Cf. former Senator Roman Hruska, who asked, during the controversy over President Nixon's nomination of Harold Cars well, whether mediocre people are not entitled to some representation on the court.Google Scholar

15 In contrast, appointments to the U.S. Supreme Court have shown a decline in prior judicial experience. Over half (56%) of the justices appointed between 1870 and 1900 had over five years of prior judicial experience, but only 41% of the justices appointed between 1900 and 1930 did, and as noted above, only one in four (24%) of the justices appointed since 1930 have had extensive judicial apprenticeships. See Abraham, supra note 8.Google Scholar

16 There has been some decrease in the number of judges who skipped the lower courts entirely, from 46.2% in 1900–1920 to 38.7% in 1950–70, but that minor “trend” is even less sharp when we correct for the disproportionate impact of New Jersey judges on the 1900–1920 figure (New Jersey judges had less prior judicial experience on the average then other states' judges in the 1900–1945 period, and there were many more of them; New Jersey's Court of Errors and Appeals then had 16 members). When we take the average of all judges in each state and then average the results from the 16 states, New Jersey's impact is moderated: the average proportion of judges with no prior judicial experience is then 43.4% (1900-1920), 43.9% (1925–45), and 39.4% (1950–70). Moreover, there has been a slight increase in the number of judges who served only a brief lower court term of one or two years. The percentage of judges with two years or less prior judicial experience (including no years) was 48.7% (1900-1920), 49.8% (1925–45), and 44.2% (1950–70). Correcting for New Jersey, the comparable figures are 46%, 47.8%, and 45%.Google Scholar

17 This reflected a more general trend. Many states added IACs in the 1960s and gave their courts discretion to select which cases to review. See Fair, Daryl R., State Intermediate Appellate Courts: An Introduction, 24 W. Pol. Q. 415 (1971); and Kagan, et al., The Evolution of State Supreme Courts, supra note 4. The eight states in our sample that had not created IACs by 1970 were Maine, Rhode Island, West Virginia, Minnesota, Kansas, Idaho, Nevada, and South Dakota.Google Scholar

18 In the entire 1900–1970 period, there was only one female SSC judge in our 16-state sample–-Susie Sharp of North Carolina.Google Scholar

19 Of judges with law enforcement experience, 46.7% were in county prosecutor offices, 39.5% in state attorney general offices, and 12.5% in U.S. attorney offices. These figures include those who served as assistant district attorneys and assistant attorneys general. The clearest historical trend is an increase in the number of judges with experience as state attorneys general relative to local district attorneys. Canon, Characteristics and Career Patterns, supra note 3, examined the backgrounds of SSC judges in all 50 states in the 1961–68 period. He states that as many as 51 % of judges had held positions in prosecutors' and county attorneys' offices, acknowledging that it is often not clear from the job title whether a position was prosecutorial in nature (at 39).Google Scholar

20 Of judges who held political office, 76.4% were state legislators, 16.8% were local executives (mayors, county executive officers), and 3 % were United States congressmen. The changes in this distribution over time are not striking.Google Scholar

21 An average of 13 % of the judges in all three time periods held both prosecutorial and other political office prior to their judicial careers, and we have included these judges in both categories. This is why the percentages in table 2 would exceed 100 if they were totaled.Google Scholar

22 It is quite possible that because of the ideal of the “neutral” judge, SSC judges tend to underreport affiliations with political party organizations in their statements for official biographies.Google Scholar

23 The percentage of judges who did not hold prosecutorial or other significant political office but who were political party officials has been relatively stable over time. But the percentage of those who served as counsel to public agencies or on school boards jumped from 15.6% of all judges in the 1900–1945 period to 26.1% of 1950–70 judges.Google Scholar

24 See, e.g., Crow, John E., Subterranean Politics: A Judge Is Chosen, 12 J. Pub. L. 275 (1963).Google Scholar

25 The one marked change in any indicator of political involvement is a decrease–-from 86% in 1900–1920 to 70.5% in 1950–70–-in the proportion of judges whose self-reported biographical statements disclose some political party identification.Google Scholar

26 It is inevitably arbitrary to choose any number of years of experience as dividing “experienced” from “less experienced” judges, or to say that a certain amount of judicial experience so outweighs prior career experiences that we can characterize a person's primary career background as “judge.” But most people would agree, we think, that a demanding professional job like judging is not quickly mastered and that at least five years of experience are required before the professional has developed fully his or her own perspective on the job. See Lenore Alpert, Learning About Trial Judging: The Socialization of State Trial Judges, in James A. Cramer, ed., Courts and Judges (New York: Sage Publications, 1981).Google Scholar

27 See Jerold S. Auerbach, Unequal Justice: Lawyers and Social Change in Modem America 28–29 (New York: Oxford University Press, 1976); Robert Stevens, Two Cheers for 1870: The American Law School, in Donald Fleming & Bernard Bailyn, eds., Law in American History (Boston: Little, Brown & Co., 1971).Google Scholar

28 See Auerbach, supra note 27, at 95–97. In 1905, Auerbach reports, one-third of all law students attended night law schools.Google Scholar

29 Id.Google Scholar

30 Stevens, supra note 27, at 493–511.Google Scholar

31 For judges sitting in 1900–1920, the average age on first reaching the court was 50. An “average” judge elected or appointed in 1920 was 30 in 1900.Google Scholar

32 Auerbach, supra note 27, at 95.Google Scholar

33 In the early 1920s, the Association of American Law Schools (AALS) and the American Bar Association (ABA) standards called for at least two years of college as a prerequisite of law school. Yet in 1927, the ABA standards were met by only half the law schools in the country, and only about one-third of all law students were in AALS-approved schools. Stevens, supra note 27, at 494–96.Google Scholar

34 The education figures for the 1950–70 judges match those for the bar as a whole. According to a 1970 survey, 73.3% of all lawyers in the 16 states had college degrees, and 93% had law school degrees. The figures for lawyers in all 50 states were identical, confirming the representativeness of our sample of 16. See Bette H. Sikes et al., The 1971 Lawyer Statistical Report 8 table 4 (Chicago: American Bar Foundation, 1972).Google Scholar

35 The full list is Cornell, Columbia, Duke, Harvard, Stanford, University of Chicago, University of Michigan, University of Pennsylvania, University of Virginia, Yale.Google Scholar

36 The full list of rank 2 law schools is Georgetown, George Washington, Hastings, Indiana University, New York University, Northwestern, Rutgers, State University of New York at Buffalo, University of California at Berkeley, UCLA, University of Illinois, University of Iowa, University of Minnesota, University of North Carolina, University of Southern California, University of Texas, University of Wisconsin, Vanderbilt.Google Scholar

37 Canon, Characteristics and Career Patterns, supra note 3, presents data on the judges sitting on all 50 courts in the period 1961–68 and gives the percentage who attended ten prestigious private law schools. It is the same as our list of rank 1 schools, except that Vanderbilt is substituted for Michigan (a state university). Canon's data show that 16% of the judges sitting in the 1960s attended these prestigious law schools (id. at 35). If we exclude the Michigan Supreme Court, with its numerous Michigan law graduates, from our calculations, 24 % of the judges on the other 15 courts in the 1950–70 period attended rank 1 schools, a considerably higher figure than Canon's. One reason is that northeastern states are slightly overrepresented in our sample: in 1950–70, 50% of the Maine court's members attended rank 1 schools, 40% of New Jersey's, and 36% of Rhode Island's. Canon notes a similar preponderance of prestigious private law school backgrounds for judges in the Northeast (id. at 35).Google Scholar

38 One could argue that in the last few decades, American law schools, regardless of prestige, have become more alike and that diversity has declined. Nevertheless, the sheer number of schools (with different admissions criteria, tuition levels, and recruitment patterns) has remained large. Access to judgeships has not been dependent on graduating from one of a few elite schools.Google Scholar

39 Whereas 78.3% of judges sitting in our 16 states during the 1950–70 period were college graduates, studies in California, New Jersey, Ohio, and Tennessee (in 1957), Iowa (1967), and New Mexico (in 1981) reveal that an average of 53% of state legislators in those specific years were college graduates. In 1979, 16% of the United States population over 25 had graduated from college. See Fred R. Harris & Paul L. Hain, America's Legislative Process: Congress and the States 57 (Glenview, Ill.: Scott, Foresman & Co., 1983).Google Scholar

40 For 1900–1920 judges, 14.7% had published; for 1925–45, 11.3%.Google Scholar

41 About 7% of the 1900–1920 and the 1925–45 judges taught law before reaching the court. About 4% in all three time periods taught law during or after their SSC tenures.Google Scholar

42 See generally, James Willard Hurst, The Growth of American Law: The Law Makers chs. 12–13 (Boston: Little, Brown & Co., 1950). On the early two-man firm, see Daniel H. Calhoun, Professional Lives in America: Structure and Aspiration, 1750–1850 (Cambridge: Harvard University Press, 1965). Auerbach, supra note 27, provides a polemical account of the growth of urban law firms and contrasts them with the “traditional” legal profession.Google Scholar

43 An American Bar Foundation survey showed that by 1948, 50% of lawyers in private practice worked in cities of 200,000 or more. In 1970,62% practiced in cities of 100,000 or more. Sikes et al., supra note 34, at 14–19 table 7.Google Scholar

44 Id. The American Bar Foundation survey shows that in 1949,68.7% of private attorneys were sole practitioners. By 1970, however, that number had declined to 50%; half of all private practitioners worked in firms. With the growth in numbers of lawyers employed directly by government, corporations, and other organizations, the proportion of all lawyers who were sole practitioners in 1970 was 36.6%.Google Scholar

45 See, e.g., Ladinsky, Jack, The Impact of Social Backgrounds of Lawyers on Law Practice and the Law, 16 J. Legal Educ. 127 (1963) (study of Detroit-area lawyers practicing in 1960). But see Nelson, Robert L., The Changing Structure of Opportunity: Recruitment and Careers in Large Law Firms, 1983 A.B.F. Res. J. 109.Google Scholar

46 See Jerome E. Carlin, Lawyers on Their Own (New Brunswick, N.J.: Rutgers University Press, 1962), and id., Lawyers' Ethics: A Survey of the New York City Bar (New York: Russell Sage Foundation, 1966); Erwin Smigel, The Wall Street Lawyer (New York: Free Press of Glencoe, 1964).Google Scholar

47 See Ladinsky, supra note 45, and Carlin, Lawyers on Their Own, supra note 46. These are, of course, statistical generalizations. Many sole practitioners have had lucrative and sophisticated commercial or trial practices and strong educational backgrounds.Google Scholar

48 See Wood, Arthur L. & Wardwell, Walter I., The Lawyer and Community Leadership, 9 J. Legal Educ. 162 (1956). Firm lawyers, on the other hand, appear from Wood and Wardwell's study to be more likely than sole practitioners to be involved in pro bono work with nonpolitical organizations.Google Scholar

49 Id. See also Landon, Donald D., Lawyers and Localities: The Interaction of Community Context and Professionalism, 1982 A.B.F. Res. J. 459.Google Scholar

50 We defined urban or city lawyer on a relative basis to account for differences between states and changes over time in the size of cities. City meant (1) one of the three largest cities in the state at each sample year; or (2) any city over 100,000 in population, if there were more than three; or (3) the state capital (if it was not one of the three largest cities). Our goal was to distinguish among types of legal practice within each state.Google Scholar

51 We suspect that most of the large category of judges in table 7 whose former practice is unknown were small-town lawyers, because on measures such as type of law school attended, legal scholarship, and movement from private practice into public office (all discussed later), they resembled known former small-town lawyers much more closely than either urban firm lawyers or urban solo lawyers. Canon, Characteristics and Career Patterns, supra note 3, using different (more stringent) measures of urban practice, indicates even more former small-town lawyers on the courts in the 1960s. His data cover judges in all 50 states in the 1961–68 period. In his study, an urban lawyer was defined as one who practiced in a Standard Metropolitan Statistical Area over 100,000 according to the 1960 census. By comparison, our study included as urban a lawyer who practiced in, for example, Boise, Idaho–-not a city by Canon's definition, but both the biggest city in the state and the state capital. By Canon's definition, only 30.6% of judges in 1961–68 had practiced in a city (id. at 37); hence he concludes that former urban lawyers were strongly underrepresented. In our definition, 54.6% of those we have information about (and 45.5% of all judges in our 16 states) in the 1960–70 period were former urban lawyers.Google Scholar

52 The results are rather consistent over time. Of judges who had practiced in city firms, 57.2% of those sitting in 1900–1920 had graduated from rank 1 or 2 schools, as had 60% of those sitting in 1925–45, and 73.3% of those in 1950–70. The corresponding figures for former urban solo lawyers were 37%, 49%, and 45.5%, and for former small-town lawyers, 22.6%, 32.2%, and 48.8% in 1950–70. Judges who had been city firm lawyers were also much less likely than their colleagues to have skipped law school entirely in favor of a law office clerkship. At least half of 1900–1920 judges had clerked, but only 14% of those who had practiced in urban firms had obtained their legal education that way.Google Scholar

53 Judges who graduated from elite law schools (ranks 1 and 2) were more likely to publish legal articles (21.6% had done so) than judges who graduated from rank 3 or 4 schools or had clerked (11.2%). The elite-school graduates were more likely to have taught law, however, only among 1900–1920 judges (19.7% compared to 3.8% for other judges); this difference disappeared for 1925–45 judges and 1950–70 judges. The correlation between having published and having taught law, incidentally, was rather low–-Pearson's R was .04 for 1900–1920 judges, .16 for 1925–45 judges, and .35 for 1950–70 judges.Google Scholar

54 Canon's study of judges in all 50 states in 1961–68 shows an average age at first accession to the court of 55.8. Canon, Characteristics and Career Patterns, supra note 3, at 37.Google Scholar

55 Canon's data, id., show an average retirement age of 72.Google Scholar

56 Canon's data (1961-68, all 50 courts), id., show an average age of 61.7.Google Scholar

57 In our 16 states, the longevity record was held by Judge W. A. Johnston, who sat on the Kansas Supreme Court from 1884 to 1935. Charles W. Parker was on New Jersey's Court of Errors and Appeals for 40 years (1907-47). Grafton Green of Tennessee (1910-47) and Lucien D. Gardner of Alabama (1914-51) served for 37 years, and John C. Anderson of Alabama (1904-40) for 36.Google Scholar

58 This average was calculated by computing the mean years of experience for each panel of judges on each court at five-year intervals. For example, the average time since first accession to the court for the three Idaho Supreme Court judges sitting in 1900 was 7.7 years; for those sitting in 1905,7.3 years; and so on through 1970, a process repeated for all 16 SSCs. These 240 means (15 time points × 16 states) were then averaged.Google Scholar

59 We constructed this “turnover” measure by computing the percentage of “new judges” on each court at five-year intervals; a new judge was one who had not been on the court at the previous sample year (five years earlier). In 1900–1920, 38.5% of the judges were such new judges (averaging the 16 states); in 1925–45, the figure was 39.5%, and in 1950–70, 35.3%.Google Scholar

60 Pitney was appointed to the United States Supreme Court in 1912. Brennan was an associate justice of the New Jersey Supreme Court for four years before being appointed to the Court in 1956.Google Scholar

61 There have been, of course, a few well-known appointments of judges from other states, such as Holmes of Massachusetts and Cardozo of New York. But of the other appointments to the Court in this century, only four others had been SSC judges. Three were appointed by Taft in 1909 and 1910: Joseph Lamar of Georgia; Willis Van Devanter, who had been on the Wyoming Supreme Court for one year; and Horace Lurton, who had been on the Tennessee Supreme Court before moving to the federal circuit court of appeals in 1893. Cound, supra note 13. Sandra O'Connor, appointed by President Reagan in 1981, had sat on the Arizona Court of Appeals for two years but not on the Arizona Supreme Court.Google Scholar

62 Fourteen percent had been both prosecutors and elected political officials.Google Scholar

63 Similarly, very few (16.7%) of the Michigan Supreme Court judges in 1900–1920 came to the bench from a career entirely in private practice (the 16-state mean then was 40.6%). Yet well over half (57.9%) of its judges in 1950–70 did (compared to a national average of 38.3%).Google Scholar

64 These states began the century with partisan elections. By 1966, Michigan, South Dakota, Minnesota, and Nevada had switched to nonpartisan elections, although partisanship remained strong in Michigan. In 1958 Kansas adopted a version of the Missouri plan. (In the Missouri plan the judge is first appointed, via a process strongly influenced by the bar association, then, after a term in office, ratified or rejected by the electorate.) Illinois, meanwhile, retained partisan elections for the initial selection of its judges, but by 1966 had instituted “merit” elections (i.e., a Missouri-type plan) for determining their subsequent retention or rejection. Winters, Glenn R., Selection of Judges–-An Historical Introduction, 44 Tex. L. Rev. 1081, 1087 (1966).Google Scholar

65 See Canon, The Impact of Formal Selection Processes, supra note 3.Google Scholar

66 California Supreme Court judges have since 1934 been appointed by the governor, subject to approval by a commission consisting of the attorney general, the chief justice, and the senior presiding courts of appeal judge, and then subject to retention or rejection by the electorate at the next gubernatorial election.Google Scholar

67 California IAC judges, too, typically had served a substantial trial court apprenticeship. See Berg, Larry L. et al., The Consequences of Judicial Reform: A Comparative Analysis of the California and Iowa Appellate Systems, 28 W. Pol. Sci. Q. 263 (1975); Thomas Y. Davies, Organizational Behavior and the Distribution of Case Outcomes in a California Court of Appeal 36–39 (Ph.D. diss., Northwestern University, 1980).Google Scholar

68 E.g., in states with a high proportion of judges who had held elective political office, two (Rhode Island, Alabama) had a higher than average percentage of judges with five or more years of lower court experience; two (North Carolina, Tennessee) were average; and two (New Jersey, Maine) ranked low in prior judicial experience. The same spread in prior judicial experience occurred in states that tended to select judges with prior careers as public prosecutors. Even when we examine the data separately in three time periods, there is no significant relationship between the predominant prior career among a state's judges and the extent of their prior judicial experience.Google Scholar

69 The figures in table 10 are averages for all three time periods.Google Scholar

70 The proportion of judicially inexperienced former politicians on a court rarely exceeded 35% of the court in any state for a substantial period. In 1900–1920, New Jersey (41.5%) and Nevada (36.4%) exceeded that level. In 1925–45, the leaders were South Dakota (50%), North Carolina (46.2%), and Maine and Kansas (both 37.5%). In 1950–70, only North Carolina (36.8%) was in this category.Google Scholar

71 The proportion of judges with less than five years of lower court experience and backgrounds in law enforcement exceeded 35% in the following cases for 1900–1920 Kansas (53.8%), Nevada (45.5%), South Dakota (37.5%); for 1925–45: Kansas (56.3%), South Dakota (41.7%); for 1950–70: South Dakota (40%). Kansas (37.5%), Nevada (36.4%), Idaho (35.7%).Google Scholar

72 The 35% level for this category was exceeded by North Carolina (40%). Maine (38.9%). and New Jersey (36.8%) in 1900–1920; by New Jersey alone (35.5%) in 1925–45, and by Michigan (36.8%) in 1950–70.Google Scholar

73 As noted in table 7, information on prior practice was lacking for 42.6% of 1900–1920 judges, 27.7% of 1925–45 judges, and 21.7% of 1950–70 judges.Google Scholar

74 The fourth was tiny Rhode Island, surrounded by states with rank 1 law schools.Google Scholar

75 Kagan et al., The Evolution of State Supreme Courts, supra note 4, and Friedrnan et al., supra note 4.Google Scholar

76 The Michigan Supreme Court issued nonunanimous opinions in 44.4% of the cases we sampled in 1960,1965, and 1970, compared to a 16-state mean of 16.5%. Friedman et al., supra note 4, at table 5. For the dedicated reader of these statistics, we should note that in the table just cited in Friedman et al., the figures for Tennessee and West Virginia were inadvertently reversed.Google Scholar

77 Idaho's dissent rate in 1940–70 was 22.3% of sampled cases, compared to a small-state mean, excluding Idaho, of 8.4%.Google Scholar

78 In 1940–70 the Idaho Supreme Court cited nine or more prior cases in 69.8% of its sampled opinions compared to a 16-state mean of 54% and a small-state mean, excluding Idaho, of 42%.Google Scholar

79 In 1940–70, in all the sampled cases in which constitutional issues were raised, the Idaho court failed to support the claim of unconstitutionality, compared to a 16-state mean of 19.4% and a mean of 30% in the high-discretion courts and in Michigan and Minnesota.Google Scholar

80 See, e.g., Preble Stolz, Judging Judges: The Investigation of Rose Bird and the California Supreme Court (New York: Free Press, 1981).Google Scholar

81 The Illinois Supreme Court, 55.6% of whose judges in 1950–70 were former prosecutors, ranked fifteenth in decisions for criminal defendants in that period, deciding in their favor in only 17.2% of sampled cases, compared to a 16-state mean of 30.5%. Nevada, fourth in percentage of former prosecutors, was thirteenth in decisions in favor of criminal defendants (20%).Google Scholar

82 South Dakota, which ranked first in number of former prosecutors on the court in 1950–70, supported the criminal defendant in 35% of sampled cases, and Idaho did in 30.8%.Google Scholar

83 The correlation was also reduced by the fact that the Maine Supreme Court, which ranked sixteenth in former law enforcement officials, had no corresponding propensity to rule in favor of criminal defendants, ranking fourteenth in that measure. And Kansas, ranking eighth in number of former prosecutors (about average) was sixteenth in deciding for criminal defendants, ruling in their favor in fewer than 10% of sampled criminal cases in 1950–70.Google Scholar

84 In every decade from 1870 to 1960, sampled cases from the Rhode Island court revealed a nonunanimous opinion rate of less than 4%. In 1960, 1965, and 1970, however, that figure rose to 11.1%.Google Scholar

85 In West Virginia, the high dissent-rate ranking, despite high political party consensus, might be attributed to court structure: the West Virginia court, like that in New Jersey and California, had full case-selecting discretion, enabling and encouraging it to select the most difficult and controversial appeals. The hypothesis is that court structure outweighs political party consensus/dissensus in stimulating dissent. Thus, too, Rhode Island, despite party dissensus, could have a low dissent rate because it lacked case-selecting discretion and had a high proportion of routine, noncontroversial cases. But consider the Tennessee Supreme Court, which had a partial IAC buffer and some case-selecting discretion yet ranked last in nonunanimous opinions (only 3.3% for 1950–70). Tennessee's judges were all Democrats. Did party consensus outweigh court structure? We are less inclined to explain Tennessee's low dissent rate in terms of party consensus than in terms of a century-old “judicial tradition” of nondissent that survived changes in court structure. This is similar to the case of Rhode Island, where a low dissent rate has persisted despite political party division on the court.Google Scholar

86 Another “outlier” that reduced the correlation was the Minnesota court, which ranked fourth in unconstitutionality rulings (second, if we eliminate courts with high case-selecting discretion), but was almost entirely (86%) Republican, 1950–70.Google Scholar

87 In 1950–70, Kansas, Nevada, Alabama, and Michigan courts, which ranked lowest in proportion of judges with five years of lower court experience, ranked sixth, fifteenth, twelfth, and thirteenth in reversal rates. But Illinois and Rhode Island, second and third in prior judicial experience, also had below-average reversal rates. The courts with high case-selecting discretion (California, New Jersey, West Virginia, and Tennessee) had high reversal rates regardless of level of prior judicial experience.Google Scholar

88 Idaho, for example, with a 45% average turnover rate for judges in each five-year period 1950–70, ranked fourth of the courts in dissent. When we set aside courts with a strong tradition of nondissent (Maine, Rhode Island, Alabama, and Tennessee) and those with case-selecting discretion (which seems to encourage dissent) (California, New Jersey, and West Virginia) in order to isolate more clearly the influence of turnover, the relationship remains only very weakly in the direction of high turnover-lower dissent.Google Scholar

89 In our 1950–70 sample, California Supreme Court opinions averaged 6.2 pages; New Jersey, 6.8 pages; and West Virginia, 6.7 pages–-qualifying them as the three wordiest of the 16 courts; the Tennessee Supreme Court's opinions averaged only 4.4 pages, shortest of the 16 states. Almost 81% of the New Jersey court's opinions contained nine or more citations to court cases, as did 74% in California and 71 % in West Virginia; the Tennessee Supreme Court cited nine or more cases in only 36.7% of its opinions–-fifteenth of the 16 courts.Google Scholar

90 In a 1940–70 sample of cases, 19.6% of New Jersey and 17.5% of California Supreme Court opinions cited law review articles (putting them first and second among the 16 states in that regard), whereas only 3.4% of Tennessee and 3% of West Virginia Supreme Court cases did so (eleventh and thirteenth among the 16 courts).Google Scholar

91 On average, each California Supreme Court opinion in the 1940–70 period was cited 4 times by out-of-state courts and each New Jersey opinion, 3 times, compared with a 16-state mean of 1.5 times. The average Tennessee opinion was cited 1.2 times by out-of-state courts, and the average West Virginia opinion, 0.7. Kagan et al., The Evolution of State Supreme Courts, supra note 4, at 992 table 5. See also Friedman et al., supra note 4, at 806, and Gregory A. Caldeira, On the Reputation of State Supreme Courts, 5 Pol. Behav. 83 (1983).Google Scholar

92 The one sharp difference is that the Rhode Island court, as compared with Maine's, ruled far more often in favor of criminal defendants in the 1950–70 period. Rhode Island's court was evenly divided between Democrats and Republicans, while Maine's was predominantly Republican.Google Scholar

93 See, e.g., Philip, E. Johnson, The Accidental Decision and How It Happens, 65 Calif. L. Rev. 231 (1977); John B. Oakley & Robert S. Thompson, Law Clerks and the Judicial Process (Berkeley: University of California Press, 1980).Google Scholar

94 See, e.g., Arthur T. Vanderbilt 11, Changing Law: A Biography of Arthur T. Vanderbilt (New Brunswick, N.J.: Rutgers University Press, 1976).Google Scholar

95 See, e.g., Robert A. Kagan, What If Abe Fortas Had Been More Discreet?in Nelson Polsby, ed., What If? Explorations in Social Science Fiction (Lexington, Mass.: Lewis Publishing, 1983).Google Scholar

96 See Gibson, supra note 1.Google Scholar

97 For sophisticated examples, at the trial-judge level, see Caldeira, supra note 91, and Gibson, James L., Environmental Constraints on the Behavior of Judges: A Representational Model of Judicial Decision Making, 14 Law & Soc'y Rev. 343 (1980). For a valuable study examining the effect of lawyers' briefs on SSC decisions, see Thomas B. Marvell, Appellate Courts and Lawyers (Westport, Conn.: Greenwood Press, 1978).Google Scholar