How far does the gender gap extend?

Abstract

The importance of women on the bench and the influence of gender on judicial decision making has garnered much scholarly attention. We examine the voting behavior of male and female justices in 718 Fourth Amendment search and seizure votes cast on state Supreme Courts between 1980 and 2000. We find that women justices, controlling for institutional, political, and legal constraints, are more likely to rule in favor of the criminal defendant than their male brethren in cases decided after 1991 but not before. We also find women justices serving with female colleagues are more inclined to render liberal votes. We conclude the influence of gender may be evident in a wider variety of cases than those dealing with women's lives but that this influence is dependent upon the existence of a critical mass of women on the state court benches.

Acknowledgements

We appreciate the efforts of Rowan Rowzanski at Colorado State University and Alex Bernstein at San Diego State University, who helped gather the data reported in this paper.

Notes

1 Based on Gilligan's (1982) “different voice” thesis and MacKinnon's (1987) extension, many authors attempt to define a “feminist perspective.” Most agree that in women's rights cases (e.g., abortion, sexual harassment) the perspective dictates women judges are more likely to rule for women's legal claims. Beyond such cases, expectations are less clear though some themes emerge (see Pruitt, 1994 for a review). Marcus et al. (1985) maintain women justices will try to enact legal changes to better reflect a concern for community, Sherry (1986) asserts they are more likely to emphasize connection, context and responsibility over rights and Resnik (1988) posits they handle cases with greater care and compassion.

2 Davis et al. (1993) consider 1981–1990 U.S. Courts of Appeals’ search and seizure cases. While they find gender to be insignificant, this need not refute the arguments here because their study is of cases from the 1980s when few women occupied the bench and because federal judges do not face retention.

3 For additional analysis, see Sunstein, Schkade, and Ellman (2004). The authors find the likelihood of liberal decisions on three-judge panels when all three justices are liberal is higher than when only two justices are liberal. Thus, even though the majority of the court is liberal in both instances, the addition of a third like-minded judge results in increased liberal court outcomes.

4 Law school socialization may also help explain the value of using a tipping point. Guinier, Bartow, and Stachel (1994) and Torrey, Ries, and Spiliopoulos (1998) argue that the law school adversarial environment strongly discourages female students from speaking out and expressing different viewpoints. The authors also note that as the number of women law students has increased, female students have become more assertive. This law school socialization process also might influence the way these female law students learn to interact with male colleagues, including eventual interactions between male and female judges.

5 Women held 10 state supreme court seats in 1980, 39 in 1992 (Allen & Wall, 1993: 157) and 87 in January 2001 (Kay & Sparrow, 2001: 7). By 2000, nine states each had three women justices sitting on seven-member panels and four states each had two women on panels of five (Kay & Sparrow, 2001: 7).

6 We examine 175 randomly selected 4th Amendment cases with an explicit privacy claim using “Fourth Amendment and Privacy” search parameters in the Westlaw data base. Per-curium decisions are excluded, leaving 148 usable cases. While this process yields relatively few cases per state/per decade, the number of data points within each type of retention method, decade, type of legal variable and judge's gender is sufficient to allow for confidence in the results. The number of cases per state or per judge is less relevant here because we are not concerned with how justices behave per state but rather per “type” (elected or appointed) of state and per “type” (male or female) of judge. Moreover, the results presented here are consistent with other, recent gender studies (see McCall, 2005) as well as with studies on partisanship cohorts (Sunstein et al., 2004). Still, additional cases would have bolstered our analysis.

7 Dramatic gains occurred in other criminal justice areas. The number of sworn, women police officers nearly tripled in U.S. cities during the 1980s (McCall, 2004). In 1980 women accounted for 33.5% of all law school students, compared to less than 4% in the early 1960s (see Torrey et al., 1998: 267 footnote 1).

8 We use the definition of “liberal” and “conservative” advanced by Gottschall (1983: 168) who defines, “attitudinal liberalism as a relative tendency to vote in favor of the legal claims of the criminally accused and prisoners in criminal and prisoner's rights cases and in favor of the legal claims of women and racial minorities in sex and race discrimination cases respectively.”

9 Reports are similar in medicine where female doctors spend more time with patients and are more willing to work in impoverished areas (see Rhode, 1993: 1551 footnote 14).

10 See Asch (1951) for the importance of a ‘like mind’ to a male's ability to resist the conforming pressures of the majority, and McCall's (2004: 377–378) review of behavioral and attitudinal changes among women officers when joined by a substantial number of other female officers.

11 Others (Traut and Emmert, 1998) apply Segal and Spaeth's (1993) attitudinal model to state judges.

12 Brace et al. (2000) create a measure of judicial ideology for each state supreme court justice by weighting Berry et al.'s (1998) citizen ideology scores by the justices’ partisan affiliations. Using the citizen ideology scores, the authors “generate partisan weights for these scores by using logit to predict the partisan affiliations for the justices as a function of the initial ideology score, computing probabilities and pseudo residuals, and then multiplying the pseudo residual by ideology and adding the product to the ideology measure” (Bonneau & Hall, 2001: 28). As such, the PAJID scores consider judge partisanship in their construction. A PAJID score of 50 indicates a moderate justice in California as well as in Texas. We use the PAJID scores because they control for citizen ideology and generally perform better than partisanship (Brace et al., 2000). Nonetheless, we also ran the models using judicial party identification (coded Democrat = 1, Republican = 0) instead. Party was insignificant and there were no real differences in the significance or relative magnitude of the other variables. For a complete description of the derivation of the PAJID scores, see Brace et al. (2000).