Does the Judge Matter? Exploiting Random Assignment on a Court of Last Resort to Assess Judge and Case Selection Effects

Journal of Empirical Legal Studies

Volumn 9, Issue 2, 2012, Pages 246-290

  Eisenberg, Theodore ^a   Fisher, Talia ^b   Rosen Zvi, Issi ^b  

^a MIT Sloan School of Management   (United States)

^b TEL AVIV UNIVERSITY   (Israel)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 84867373668     PISSN: 17401453     EISSN: 17401461     Source Type: Journal    
DOI: 10.1111/j.1740-1461.2012.01253.x     Document Type: Article

References (114)

1. For example, Orley Ashenfelter, Theodore Eisenberg & Stewart J. Schwab, Politics and the Judiciary: The Influence of Judicial Background on Case Outcomes, 24 J. Legal Stud. 257 (1995); Matthew Hall, Randomness Reconsidered: Modeling Random Judicial Assignment in the U.S. Courts of Appeals, 7 J. Empirical Legal Stud. 574 (2010).
2. At the trial-court level, most studies ignore the vast bulk of civil case outcomes, which consist of settlements. Since settlement is the dominant outcome for cases in which plaintiffs succeed, ignoring settlements calls into question findings of judge and judicial background effects.
3. Denise M. Keele, Robert W. Malmsheimer, Donald W. Floyd & Lianjun Zhang, An Analysis of Ideological Effects in Published Versus Unpublished Judicial Opinions, 6 J. Empirical Legal Stud. 213, 234-36 (2009).
4. Jonathan P. Kastellec & Jeffrey R. Lax, Case Selection and the Study of Judicial Politics, 5 J. Empirical Legal Stud. 407 (2008).
5. Christina L. Boyd, Lee Epstein & Andrew D. Martin, Untangling the Causal Effects of Sex on Judging, 54 Am. J. Pol. Sci. 389, 394 (2010).
6. Cass R. Sunstein, David Schkade, Lisa Michelle Ellman & Andres Sawicki, Are Judges Political? An Empirical Analysis of the Federal Judiciary (2006).
7. Hall, supra note 1.
8. Ashenfelter etal., supra note 1, at 268-70.
9. Boyd etal., supra note 5.
10. For example, Joshua D. Angrist & Jörn-Steffen Pischke, Mostly Harmless Econometrics: An Empiricist's Companion 63 (2009).
11. Kate Stith & José A. Cabranes, Fear of Judging: Sentencing Guidelines in the Federal Courts 9-11 (1998); James M. Anderson etal., Measuring Interjudge Sentencing Disparity: Before and After the Federal Sentencing Guidelines, 42 J.L. & Econ. 271 (1999).
12. James M. Anderson etal., Measuring Interjudge Sentencing Disparity: Before and After the Federal Sentencing Guidelines, 42 J.L. & Econ. 271 (1999).
13. A. Abigail Payne, Does Inter Judge Disparity Really Matter? An Analysis of the Effects of Sentencing Reforms in Three Federal District Courts, 17 Int'l Rev. L. & Econ. 337 (1997) (using data from 1980 to 1991 for select types of cases in three federal district courts); Joel Waldfogel, Aggregate Inter-Judge Disparity in Federal Sentencing: Evidence from Three Districts, 4 Fed. Sentencing Rep. 151 (1991) (using data from three different district courts from 1984 to 1987); Joel Waldfogel, Does Inter-Judge Disparity Justify Empirically Based Sentencing Guidelines?, 18 Int'l Rev. L. & Econ. 293 (1998).
14. Joel Waldfogel, Aggregate Inter-Judge Disparity in Federal Sentencing: Evidence from Three Districts, 4 Fed. Sentencing Rep. 151 (1991) (using data from three different district courts from 1984 to 1987)
15. Joel Waldfogel, Does Inter-Judge Disparity Justify Empirically Based Sentencing Guidelines?, 18 Int'l Rev. L. & Econ. 293 (1998).
16. Ryan W. Scott, Inter-Judge Sentencing Disparity After Booker: A First Look, 63 Stan. L. Rev. 1 (2010). There is evidence of congressional overreaction to perceived downward departures by judges. Max Schanzenbach, Have Federal Judges Changed Their Sentencing Practices? The Shaky Empirical Foundations of the Feeney Amendment, 2 J. Empirical Legal Stud. 1 (2005); Scott, supra. Compare Beth A. Freeborn & Monica E. Hartmann, Judicial Discretion and Sentencing Behavior: Did the Feeney Amendment Rein in District Judges?, 7 J. Empirical Legal Stud. 355 (2010).
17. Max Schanzenbach, Have Federal Judges Changed Their Sentencing Practices? The Shaky Empirical Foundations of the Feeney Amendment, 2 J. Empirical Legal Stud. 1 (2005)
18. Scott, supra. Compare Beth A. Freeborn & Monica E. Hartmann, Judicial Discretion and Sentencing Behavior: Did the Feeney Amendment Rein in District Judges?, 7 J. Empirical Legal Stud. 355 (2010).
19. For example, Timothy R. Johnson, James F. Spriggs II & Paul J. Wahlbeck, Passing and Strategic Voting on the U.S. Supreme Court, 39 Law & Soc'y Rev. 349 (2005).
20. Yoram Shachar & Miron Gross, Acceptance and Rejection of Appeals to the Supreme Court: Quantitative Analyses, 13 Legal Stud. 329 (1996) (in Hebrew)
21. Keren Weinshall-Margel, Attitudinal and Neo-Institutional Models of Supreme Court Decision Making: An Empirical and Comparative Perspective from Israel, 8 J. Empirical Legal Stud. 556 (2011).
22. Stephen J. Choi, Mitu Gulati, Mirya Holman & Eric A. Posner, Judging Women, 8 J. Empirical Legal Stud. 504 (2011). Boyd, Epstein, and Martin and Greiner and Rubin, echoing earlier work, point out some difficulty in asserting that any inherent individual characteristic, such as sex or race, causes behavior such as judicial voting patterns. Boyd etal., supra note 5, at 397; D. James Greiner & Donald B. Rubin, Causal Effects of Perceived Immutable Characteristics, 93 Rev. Econ. Stat. 775 (2011). Nevertheless, they recognize that it can be interesting to compare people with such characteristics to see if there are differences in outcome. Boyd etal., supra note 5, at 397 (quoting Donald Rubin). In the Israeli context, a study performed by Bogoch and Don-Yechiya examined the effect of the presence of a female justice on the panel on the tendency to accept or reject appeals as well as its impact on sentencing severity. Bryna Bogoch & Rachelle Don-Yechiya, The Gender of Justice: Bias Against Women in Israeli Courts 133-36 (1999).
23. D. James Greiner & Donald B. Rubin, Causal Effects of Perceived Immutable Characteristics, 93 Rev. Econ. Stat. 775 (2011). Nevertheless, they recognize that it can be interesting to compare people with such characteristics to see if there are differences in outcome. Boyd etal., supra note 5, at 397 (quoting Donald Rubin). In the Israeli context, a study performed by Bogoch and Don-Yechiya examined the effect of the presence of a female justice on the panel on the tendency to accept or reject appeals as well as its impact on sentencing severity. Bryna Bogoch & Rachelle Don-Yechiya, The Gender of Justice: Bias Against Women in Israeli Courts 133-36 (1999).
24. Bryna Bogoch & Rachelle Don-Yechiya, The Gender of Justice: Bias Against Women in Israeli Courts 133-36 (1999).
25. Boyd etal., supra note 5.
26. For example, Ashenfelter etal., supra note 1, at 262 n.12.
27. Boyd etal., supra note 5, at 392 (footnote omitted). One reason for varying results is that judge and judge characteristic effects may exist at one court level but not at another due to the selection of cases for appeal. While the mass of cases may be simple enough to preclude judge characteristics from influencing their outcome, cases appealed likely tend to be closer cases in which the legal outcome is less clear. In such close cases, there is more room for judicial characteristics to exhibit an influence. Ashenfelter etal., supra note 1, at 281.
28. Weinshall-Margel, supra note 16.
29. Justices are sometimes appointed to the ISC on a temporary basis. Some of the justices appointed on a temporary basis eventually receive permanent status. Others are directly appointed on a permanent basis.
30. Eli Salzberger, Acting Justices in the Supreme Court and Judicial Independence-Theoretical Analysis and Empirical Findings, 19 Bar Ilan L. Rev. 541 (2003). See also Benyamin Blum, To Concur or Not to Concur: That is the Question: Theoretical and Practical Questions Regarding the Judicial Independence of Judges Appointed Temporarily to the Israeli Supreme Court, Thesis, Master of the Science of Law, Stanford University (May 2006).
31. See also Benyamin Blum, To Concur or Not to Concur: That is the Question: Theoretical and Practical Questions Regarding the Judicial Independence of Judges Appointed Temporarily to the Israeli Supreme Court, Thesis, Master of the Science of Law, Stanford University (May 2006).
32. Miron Gross & Yoram Shachar, How Are Supreme Court Panels Selected?-A Quantitative Analysis, 29 Hebrew U. L. Rev. 567 (1999).
33. Sunstein etal., supra note 6, at 18.
34. Keele etal., supra note 3. For a study that verified its sample with respect to publication bias, see Richard L. Revesz, Ideology, Collegiality, and the D.C. Circuit: A Reply to Chief Judge Harry T. Edwards, 85 Va. L. Rev. 805 (1999).
35. see Richard L. Revesz, Ideology, Collegiality, and the D.C. Circuit: A Reply to Chief Judge Harry T. Edwards, 85 Va. L. Rev. 805 (1999).
36. Sunstein etal., supra note 6, at 82-83.
37. Ashenfelter etal., supra note 1; Boyd etal. supra note 5; Hall, supra note 1.
38. Boyd etal. supra note 5; Hall, supra note 1.
39. An analogous situation exists with respect to legislative votes. For example, in a study of senatorial voting on U.S. Supreme Court nominees, the votes of senators on a single nominee are not independent; they are influenced by the nominee's qualifications. Shipan notes this issue in the context of Senate voting and adjusts the standard errors to account for this nonindependence.
40. Charles R. Shipan, Partisanship, Ideology, and Senate Voting on Supreme Court Nominees, 5 J. Empirical Legal Stud. 55, 62-63 (2008). When one is interested not in the nominal senator's votes but in a characteristic of the senator, such as political party, the voter's nonindependence across cases of legislative votes needs to be additionally accounted for and not just the nominee-level nonindependence.
41. Anna Harvey & Barry Friedman, Ducking Trouble: Congressionally Induced Selection Bias in the Supreme Court's Agenda, 71 J. Pol. 574, 574-76 (2009); Kastellec & Lax, supra note 4, at 408 ("[T]he Court's selection process raises the potential for selection bias in the inferences we draw from its cases."); Theodore Eisenberg, Talia Fisher & Issi Rosen-Zvi, Case Selection and Dissent in Courts of Last Resort: An Empirical Study of the Israel Supreme Court in Empirical Studies of Judicial Systems 2011 (Yun-chien Chang ed. 2012) (forthcoming).
42. Kastellec & Lax, supra note 4, at 408 ("[T]he Court's selection process raises the potential for selection bias in the inferences we draw from its cases."); Theodore Eisenberg, Talia Fisher & Issi Rosen-Zvi, Case Selection and Dissent in Courts of Last Resort: An Empirical Study of the Israel Supreme Court in Empirical Studies of Judicial Systems 2011 (Yun-chien Chang ed. 2012) (forthcoming).
43. Theodore Eisenberg, Talia Fisher & Issi Rosen-Zvi, Case Selection and Dissent in Courts of Last Resort: An Empirical Study of the Israel Supreme Court in Empirical Studies of Judicial Systems 2011 (Yun-chien Chang ed. 2012) (forthcoming).
44. Richard L. Revesz, Environmental Regulation, Ideology, and the D.C. Circuit, 83 Va. L. Rev. 1717 (1997); Sunstein etal., supra note 6, at 22-23.
45. Sunstein etal., supra note 6, at 22-23.
46. The Israeli judiciary description is based on the descriptions in Eisenberg etal., supra note 31
47. Theodore Eisenberg, Talia Fisher & Issi Rosen-Zvi, Israel's Supreme Court Appellate Jurisdiction: An Empirical Study, 96 Cornell L. Rev. 693, 700-04 (2011).
48. See generally Courts Law (Consolidated Version), 5744-1984, 38 LSI 271 (1983-1984).
49. Courts Law (Consolidated Version) ch. 2, art. 3.
50. Ordinances of Courts (Establishment of the Central District Court), 2007, KT 6585, 824. In results not reported below, we assessed whether screening outcomes in the ISC differed based on the district court being reviewed. We found no significant differences.
51. Courts Law (Consolidated Version) ch. 2, art. 2.
52. Basic Law: Judicature §20.
53. Courts Law (Consolidated Version) §§26, 30.
54. Eisenberg etal., Cornell, supra note 33.
55. See Courts Law (Consolidated Version) §41(a).
56. Israeli law, which does not differentiate between appeals of acquittals and convictions, grants the prosecution the authority to appeal a defendant's acquittal. For discussion of this authority, see Eisenberg etal., Cornell, supra note 33, at 783 n.52.
57. see Eisenberg etal., Cornell, supra note 33, at 783 n.52.
58. Criminal Procedure Law (Consolidated Version), 5742-1982, 36 LSI 35, §§146-147 (1981-1982).
59. CA 103/82 Chenion Haifa v. Matzat Or 36(3) PD 123 [1982].
60. See DC 4927/92 State of Israel v. Ben Yehuda (unpublished opinion).
61. CrimA 1245/93 Shtarkman v. State of Israel 47(2) PD 177 [1993].
62. DC 3251/91 Yishai v. State of Israel, PD 45(5) 441 (1991). Our prior work calls into question adherence to the Chenion Haifa standards. Eisenberg etal., Cornell, supra note 33.
63. Criminal Procedure Rules, 5734-1974, §44(7).
64. Criminal Procedure Law (Consolidated Version), 5742-1982, 36 LSI 35, §205 (1981-1982).
65. Eisenberg etal., supra note 31.
66. Eisenberg etal., Cornell, supra note 33
67. Eisenberg etal., supra note 31.
68. The website does not include cases decided in camera, but since those cases are an insubstantial fraction of the cases decided by the Court, the omission does not materially affect the analysis here.
69. See Courts Law (Consolidated Version), 5744-1984, 38 LSI 271, sect;70(a) (1983-1984).
70. Preliminary results indicated that the coding of case categories in our original data differed from the case categories that the ISC uses internally to assign discretionary jurisdiction cases to individual screening justices. In preparing Eisenberg etal., supra note 31, which focused on screening decisions by individual justices, we reviewed every discretionary case in the original data to reclassify the cases into categories more relevant to screening by individual justices.
71. We had no desire to identify individual justices and, in prior work, did not do so when it was unnecessary to presenting the results of interest about the ISC. Eisenberg etal., supra note 31. In this article, institutional details about justices, such as when they were appointed, opinion assignments, and who were the presiding justices, made identification unavoidable to achieve coherent presentation of results. We take no position on whether individual justices' voting patterns are normatively desirable.
72. Eisenberg etal., supra note 31.
73. The workload estimate starts with a measure of workload attributable to the cases justices sat on. Table5 shows that cases vary substantially in the workload they impose (as measured by time to resolution) depending on whether a case is civil or criminal in type, and on whether a case's jurisdiction in the ISC is mandatory or discretionary. We initially estimate workload by summing, for each justice, the time to resolution for all the cases on which he or she sat. We then make two adjustments. Since we undersampled civil cases relative to criminal cases, this initial sum is adjusted for the civil cases not in our sample. Justices sat on such cases but receive no "credit" for that work in our initial summation because we did not observe these cases. We therefore multiply civil case times to resolution by 24/17 to reflect our sample design, which includes 24 months of criminal cases and 17 months of civil cases. We then adjust for six justices not serving for the full two years of our study. Justice Barak left the Court in mid-September 2006 and, as Table2 shows, five justices joined the Court during our study period. Simply summing these justices' cases' times to resolution would distort workload's role in their case assignments because they could not have accumulated workload credit while not on the Court. We adjust their workloads by the inverse of the fraction of the 24 months of our study that they served on the Court. This in effect treats them as having a workload that reflects their time on the Court. This ex post workload calculation can only be a proxy for the workload's effect on case assignment, which likely occurs in real time based on justices' estimated pending workloads at the time of each case filing.
74. For example, Files Opened by the Police, by Type of Offence, Statistical Abstract of Israel 2011 (tbl. 11.3), available at .
75. For example, Ashenfelter etal., supra note 1, at 259.
76. Sometimes, the assignment to a treatment is known to be random for some subgroup. Dan A. Black, Jeffrey A. Smith, Mark C. Berger & Brett J. Noel, Is the Threat of Reemployment Services More Effective Than the Services Themselves? Evidence from Random Assignment in the UI System, 93 Am. Econ. Rev. 1313 (2003)
77. Angrist & Pischke, supra note 10, at 63. We assume that, conditional on case specialization, seniority, and workload, case assignment in the ISC is random. Our analysis is therefore in the spirit of models in which the subgroup is known to be randomly assigned, but we have more to account for with respect to the assignment process because it is based on multiple factors.
78. Eisenberg etal. supra note 31.
79. A trivial number of criminal case panels involved more than three justices.
80. Eisenberg etal. supra note 31.
81. For example, Administrative Office of the U.S. Courts, Judicial Business of the United States Courts: Annual Report of the Director 26 (2010) (for purposes of caseload analysis, "[t]he Federal Judiciary has employed techniques for assigning weights to cases since 1946.").
82. These figures are based on the cases that reported data allowing for the time computation.
83. This equivalence varied from 63 percent in sentence-only cases to 47 percent in other mandatory jurisdiction criminal cases.
84. As noted above, we excluded from our analysis interlocutory appeals, which arise largely in discretionary jurisdiction civil cases. To assess the importance of this exclusion, we analyzed 79 cases retrospectively identified as interlocutory appeals. Eisenberg etal., Cornell, supra note 33, at 708 n.75. Justice Grunis did the vast majority of the screening in these cases, 90.8 percent. This may help account for his relatively low share of discretionary civil cases shown in Figure1B. The workload imposed by these cases appears to be substantially lower than that in other discretionary civil cases. We had information on the time to disposition for 77 of the cases and they had a mean time of 203 days with a median time of 129 days.
85. Eisenberg etal., supra note 31.
86. Ashenfelter etal., supra note 1, at 268-70.
87. These expected null effects are analogous to those suggested in testing for randomness in a study of mandatory appeals. Hall, supra note 1, at 577-78. He focused on the explanatory variable of interest in that study, political party.
88. The distribution of civil case characteristics, with their more complex party structures than criminal cases, are more difficult to assess. A civil case appellate outcome can favor confounding combinations of litigants-individuals (men and women) or entities, governments or private parties. The litigant categories can appear as plaintiffs or defendants, a further possible confounding factor. Eisenbeg etal., Cornell, supra note 33.
89. Eisenberg etal., supra note 31; Kastellec & Lax, supra note 4.
90. Justice Barak's low share of Arab appellant cases is likely partly attributable to his retirement from the ISC in September 2006. The proportion of mandatory criminal Arab defendant cases was higher in 2007 (45.3 percent) than in 2006 (42.8 percent), though not significantly so. This also may help explain his low proportion of Arab appellant cases in Table11, which focuses on presiding justices.
91. Another significant difference with respect to Arab defendants exists for seniority, also likely related to Justice Barak's retirement and the low-seniority acting justices, combined with an increasing portion of Arab defendant cases. Justices who heard Arab defendant cases had on average 3.9 years of seniority on the Court. Justices who heard Jewish defendant cases had on average 4.3 years of seniority on the Court, a significant difference (p=0.007). We control for seniority in our analysis of case outcomes as it is a factor acknowledged to be a basis for nonrandom case assignment. As Table10 shows, female defendant cases comprised 1.5 percent of female justice cases and 2.3 percent of male justice cases, a near-significant difference. Since, as shown below, female defendants tend to fare better on appeal than male defendants, this pattern of case assignment should bias our results against finding a pro-defendant female justice effect, suggesting that results may be a bit stronger than reported.
92. But the difference was nearly significant (p=0.053) and was largely attributable to Justice Levy being the presiding justice in a relatively large share of cases involving "preparation and conspiracy to commit a crime."
93. Our results provide new insights, but also confirm some results reported by a 1996 Shachar-Gross study with respect to previous compositions of the ISC. In a study of 7, 147 ISC cases representing 40 percent of all cases published in the years 1948-1994, it was found that while all justices exhibited pro-government tendencies, significant variation existed among the individual justices. Shachar & Gross, supra note 15, at 351-55.
94. The Bogoch and Don Yechiya study did not find a significant gender effect on appeal outcome. Their study, however, encompassed a relatively small number of cases and the number of female justices in the relevant time period (1988-1992) was much smaller (only two justices). Bogoch & Don-Yechiya, supra note 17, at 133-36.
95. This finding is interesting in light of the Salzberger study, which found that judicial independence (defined as rulings against the government) correlates with an acting justice's chances of obtaining a permanent position. See Salzberger, supra note 23.
96. Theodore Eisenberg, Appeal Rates and Outcomes in Tried and Nontried Cases: Further Exploration of Anti-Plaintiff Appellate Outcomes, 1 J. Empirical Legal Stud. 659, 677-82 (2004)
97. Theodore Eisenberg & Henry S. Farber, Why Do Plaintiffs Lose Appeals? Biased Trial Courts, Litigious Losers, or Low Trial Win Rates? Princeton Univ., Industrial Relations Section, Working Paper #567 (July 2011).
98. Two previous studies attempted to assess the independence of the individual justices by examining the tendency of acting justices to dissent before and after attaining tenure. Both studies found no significant correlation between the acting justice status and the tendency to dissent. See Salzberger, supra note 23; Blum, supra note 23, at 73.
99. This significance level is based on a logistic regression model that controls for sentence-only cases, panel seniority, workload, and individual seniority in mandatory criminal cases. A simple nonparametric test for trend yields p=0.019.
100. In additional models not reported here, we used a different set of dummy variables for offense categories. In those models, the dummy variables represent the first major offense subject, as described in Table4, referred to in the Court's opinion, without substitution in cases in which the second or third offense referred to was murder, rape, or offenses against another's freedom.
101. Hall, supra note 1, at 578.
102. Eisenberg etal., supra note 31.
103. Eisenberg etal., Cornell, supra note 33.
104. Eisenberg etal., supra note 31.
105. About 2.4 percent of discretionary jurisdiction criminal case votes are excluded because they involved votes that we did not characterize as favoring the defendant or the state.
106. Eisenberg etal., supra note 31.
107. We have explored other models, using the seniority of the presiding justice as the measure of seniority for a case. Significant justice effects remain but standard errors deteriorate as expected.
108. Theodore Eisenberg, Paula L. Hannaford-Agor, Valerie P. Hans, Nicole L. Waters, G. Thomas Munsterman, Stewart J. Schwab & Martin T. Wells, Judge-Jury Agreement in Criminal Cases: A Partial Replication of Kalven and Zeisel's The American Jury, 2 J. Empirical Legal Stud. 171 (2005) (variables included judge, jury, and lawyer views of the strength of the evidence). Ex post assessment of the merits is sometimes possible in medical malpractice cases due to the availability of documented information about the case and expert reviewers. But even here the merits of behavior can be debated.
109. Tom Baker, Reconsidering the Harvard Medical Practice Study Conclusions About the Validity of Medical Malpractice Claims, 33 J. L., Med. & Ethics 501 (2005); Philip G. Peters, Jr., What We Know About Malpractice Settlements, 92 Iowa L. Rev. 1783 (2007).
110. Philip G. Peters, Jr., What We Know About Malpractice Settlements, 92 Iowa L. Rev. 1783 (2007).
111. Ashenfelter etal., supra note 1; Brian T. Fitzpatrick, Judicial Study of Class Action Settlements and Their Fee Awards, 7 J. Empirical Legal Stud. 811 (2010); Gregory C. Sisk etal., Charting the Influences on the Judicial Mind: An Empirical Study of Judicial Reasoning, 73 N.Y.U.L. Rev. 1337 (1998); Laura Beth Nielsen, Robert L. Nelson & Ryon Lancaster, Individual Justice or Collective Legal Mobilization? Employment Discrimination Litigation in the Post Civil Rights United States, 7 J. Empirical Legal Stud. 175, 188 (2010).
112. Brian T. Fitzpatrick, Judicial Study of Class Action Settlements and Their Fee Awards, 7 J. Empirical Legal Stud. 811 (2010)
113. Gregory C. Sisk etal., Charting the Influences on the Judicial Mind: An Empirical Study of Judicial Reasoning, 73 N.Y.U.L. Rev. 1337 (1998)
114. Laura Beth Nielsen, Robert L. Nelson & Ryon Lancaster, Individual Justice or Collective Legal Mobilization? Employment Discrimination Litigation in the Post Civil Rights United States, 7 J. Empirical Legal Stud. 175, 188 (2010).