|
Article Excerpt
CONTENTS
INTRODUCTION I. PAST EMPIRICAL FINDINGS A. Direct Effect of Gender on Judging B. Indirect Effect of Gender on Judging II. DATA A. Overview of the Data Set B. Dependent and Independent Variables C. Control Variables III. REGRESSION RESULTS A. Gender Differences in Voting Behavior B. Impact of Female Judges IV. POSSIBLE EXPLANATIONS FOR THE EFFECT OF GENDER ON PANEL DECISIONMAKING A. Deliberation B. Deference C. Logrolling D. Moderation CONCLUSION APPENDIX A wise old man and a wise old woman reach the same conclusion.... (1) [Female judges] bring an individual and collective perspective to [their] work that cannot be achieved in a system which reflects the experience of only a part of the people whose lives it [a]ffects. (2)
INTRODUCTION
This Note provides data to illuminate whether and how the presence of female judges on three-judge federal appellate panels affects collegial decisionmaking in a subset of gender-coded cases--those involving claimants alleging sexual harassment or sex discrimination in violation of Title VII of the Civil Rights Act of 1964. (3) An empirical analysis of 556 federal appellate cases decided in 1999, 2000, and 2001 reveals that judges' gender mattered to case outcomes. Though plaintiffs lost in the vast majority of cases, they were twice as likely to prevail when a female judge was on the bench.
This Note has three goals. First, it contributes to the literature on the role of gender in individual judicial decisionmaking. I show that for at least two types of cases--Title VII sex discrimination and sexual harassment--a significant correlation existed between gender and individual federal appellate judges' decisions. In my data set, female judges were significantly more likely than male judges to find for plaintiffs. Second, the Note begins to illuminate the impact of gender on panel decisionmaking, by showing that the presence of a female judge significantly increased the probability that a male judge supported the plaintiff in the cases analyzed. This analysis reaffirms the importance of collegiality in appellate courts, (4) thus distinguishing the Note from past literature, which focused almost exclusively on male/female differences. Third, the Note proposes several possible explanations for how the presence of a female judge might increase the likelihood that a male judge will support the plaintiff in gender-coded cases.
Part I reviews past empirical findings on the direct and indirect effects of gender on judging. Part II describes my data, and Part III reports the findings of regression analyses. Part IV proposes four possible mechanisms for the indirect effect of gender on collegial decisionmaking that I observed: deliberation; male judges' deference to female judges; logrolling, or strategic bargaining; and moderation of male judges' anti-plaintiff preferences. The Conclusion argues that the effects I observed should inform future debates about gender diversification of the federal appellate bench.
I. PAST EMPIRICAL FINDINGS
Some scholars expect that increasing the number of female judges will make courts more receptive to the arguments of claimants in gender-coded cases like the Title VII cases analyzed in this Note. (5) One form of this argument goes so far as to expect female judges to "seize decision-making opportunities to liberate other women" in deciding cases. (6) Thus far, the validity of arguments that gender affects case outcomes is uncertain because the literature has produced inconsistent empirical findings. In this Part, I detail past research on the direct and indirect effects of gender on judicial decisionmaking.
A. Direct Effect of Gender on Judging
Previous empirical studies examining whether male and female judges decide cases differently have produced conflicting results. Two studies of state supreme courts showed that female justices were more likely than their male counterparts to support plaintiffs in sex discrimination cases. (7) However, in federal district courts, researchers have not found evidence that judges' gender affected decisions in civil rights or discrimination cases. (8) In one study, Jennifer Segal found that female Clinton appointees to federal district courts were actually less likely than male Clinton appointees to rule for female sex discrimination plaintiffs. (9) While some studies provided evidence that female federal appellate judges were more likely to side with plaintiffs, particularly in discrimination cases, (10) other studies produced opposite findings (11) or revealed no significant gender differences. (12)
Assuming that gender differences do exist, there are several possible explanations for the weak and inconsistent findings of previous studies. First, much of the earlier work, of necessity, focused on a relatively small pool of female judges, possibly resulting in sample sizes that were too small to register differences. (13) Second, because female judges were once novelties on courts, selection bias may have resulted in the first female judges being more like male judges in their substantive views. (14) Additionally, the pressure of being "tokens" may have encouraged early female judges to conform their views to those of their male colleagues regardless of their personal preferences. (15) Third, past studies examined a broad range of issue areas--many where gender was arguably less salient than in the Title VII sexual harassment and sex discrimination cases examined in this Note. (16) Fourth, many past studies contained "inadequate controls for factors other than gender" that might have affected decisionmaking, including ideology and other background variables such as past careers or experience. (17) Finally, gender differences in voting behavior might be muted on collegial courts if the presence of a female judge affects male judges' decisions.
B. Indirect Effect of Gender on Judging
Past research on the indirect effect of judges' gender on federal appellate case outcomes also produced inconsistent results. In a study of employment discrimination cases decided in the federal courts of appeals between 1981 and 1996, Nancy Crowe found that the presence of a female judge on a panel did not have a statistically significant effect on the decisions of the male judges with whom she sat. (18) Because Crowe focused exclusively on nonunanimous decisions, however, she may have preordained her result. By definition, nonunanimous cases involve disagreement, and thus we are less likely to find male judges agreeing with their female colleagues in such cases. Furthermore, nonunanimous cases are rare, (19) and judges' decisions in such cases may not be indicative of their general voting patterns.
Two other studies provided some evidence that the presence of female judges on federal appellate panels affected case outcomes. Analyzing a sample of cases decided between 1977 and 1996, Tajuana Massie, Susan Johnson, and Sara Gubala discovered that the presence of one or more female judges increased the probability that male judges took pro-plaintiff positions in criminal procedure and civil rights cases. (20) Similarly, Sean Farhang and Gregory Wawro found that in employment discrimination cases decided in 1998 and 1999, male judges on panels with female judges were more likely to favor plaintiffs than their colleagues on all-male panels. (21) However, in both studies, the magnitude of female judges' impact (as measured by the change in the probability that a male judge found for the plaintiff) was relatively low--likely because of the broad range of case types included in each study.
Furthermore, empirical flaws in both of the above studies reduce the validity of their results. First, the studies failed to account for individual characteristics other than gender, ideology, and race that could affect judges' decisions. (22) As a result, they may have overestimated the effect of gender. Second, both studies used flawed measures of colleague ideology, thus potentially underestimating the impact of a judge's colleagues on his or her decisions. Farhang and Wawro included a single variable representing the average of the Poole common space scores (23) of the appointing president for the judge's two panel colleagues. (24) This meant that a conservative and a liberal appointee canceled each other out in the model and thus likely resulted in biased measures of the other coefficients. Massie and her coauthors did not directly measure the ideology of the panel; instead, they calculated the median common space scores for the appointing president of each active judge in the circuit and applied that figure to every panel from that circuit. (25) Third, the homogeneity and thus the comparability of the cases examined are questionable. Farhang and Wawro analyzed cases from a two-year period marked by significant changes in Supreme Court precedent. (26) Massie and her coauthors used a twenty-year period that saw even greater changes in Supreme Court doctrine (27) as well as the adoption of the Civil Rights Act of 1991, (28) which significantly strengthened the protections of Title VII. Because Massie and her coauthors did not control for significant doctrinal change and because the number of female federal appellate judges increased during their sample period, what looks like a greater propensity of panels with female members to rule for plaintiffs could be a pro-plaintiff doctrinal shift.
This Note improves on past research designs by using a limited time frame with no significant changes in Supreme Court precedent or federal statutes, to increase the homogeneity of the cases in the sample; (29) by including variables representing a judge's past careers, age, and federal appellate experience; and by coding separate variables to represent the ideology of each panel colleague. Further, this Note may be more relevant than some of the prior studies because I focus on a narrow set of gender-coded cases, where scholars anticipating direct and indirect effects of gender argue that the presence of female judges is essential to producing fair outcomes. (30)
II. DATA
A. Overview of the Data Set
My data set includes published and unpublished decisions in all sexual harassment and sex discrimination cases decided by the federal courts of appeals between 1999 and 2001 (31) where the plaintiff's cause of action fell under Title VII of the Civil Rights Act of 1964. (32) I generated the set via keyword searches on the Westlaw electronic database. (33) Cases are included from all eleven numbered circuits and the District of Columbia Circuit. The sample includes 556 total cases, consisting of 1666 decisions (i.e., votes) rendered by individual judges. (34) The sexual harassment data set contains 1091 decisions--188 from female judges and 903 from male judges. The sex discrimination data set includes 773 decisions--127 from female judges and 646 from male judges. (35) The data set is limited to cases involving appellate review on the merits of the claims. (36)
The judges in the data set include 54 females (38 Democratic appointees and 16 Republican appointees) and 273 males (116 Democratic appointees and 157 Republican appointees). These sample sizes are large enough to merit confidence in the statistical power of the results. Of the judges, 107 are district judges sitting by designation (of whom 7 are female), and 3 are judges on the U.S. Court of International Trade (of whom 1 is female).
As Table 1 indicates, the data show that in an overwhelming number of Title VII sex discrimination and sexual harassment cases--nearly 75% (416 of 556)--the plaintiffs lost. In 38% (158 of 416) of these cases where plaintiffs lost, female judges were on the panel. But female judges were also on the panel in 62% (87 of 140) of the cases in which plaintiffs won. The data indicate that the presence of a female judge significantly increased the probability that the plaintiff would prevail. Further, the dissent rate on mixed-gender panels was low (approximately 6%), indicating that male judges decided with their female colleagues rather than against them. (37)
These results contrast with the results of a sample of 367 Title VII race discrimination cases I assembled from the years analyzed. In those cases, there were no statistically significant differences between the outcomes of all-male panels and of panels with at least one female judge.
By looking more closely at individual judges' rulings in the data set, it is apparent that gender and judicial ideology (as measured by party of the appointing president) significantly affected the results. As Table 2 indicates, female judges ruled for plaintiffs more often than did male judges, and judges appointed by Democratic presidents found for plaintiffs more often than did Republican appointees. (38) Additionally, Republican-appointed females supported plaintiffs at about the same rate as Democrat-appointed males--29% and 30%, respectively. (39)
To better understand the effect of gender on individual judges' decisions and to control for other factors that may influence a judge's decision, most notably judicial ideology, I conducted probit regression analyses. (40)
For these analyses, I separated the sexual harassment and sex discrimination cases. (41) Although sexual harassment and sex discrimination cases both fall under Title VII and are conceptually similar, scholars have shown that judges treat the two types of cases differently. Even though the Supreme Court has never held that sexual harassment requires conduct of a sexual nature, studies have shown that plaintiffs are more likely to win when they allege sexualized behavior. (42) By contrast, to prevail in sex discrimination cases plaintiffs must present sufficient evidence for judges to infer that sexism is the reason for gender disparities. (43)
Before analyzing the data, I had two conflicting hypotheses regarding whether male/female differences would be greater in sexual harassment or in sex discrimination cases. On the one hand, judges' emphasis on evidence of sexualized behavior in sexual harassment cases might result in less significant gender differences in those cases than in sex discrimination cases, which require judges to infer discrimination. On the other hand, research showing that males and females in the general population differ in their perceptions of what constitutes sexual harassment (44) leads to the opposite expectation--more significant gender differences in sexual harassment cases. Surprisingly, the data indicate that individual-level gender differences were greater in sexual harassment...
|
