Dispute System Design and Bias in Dispute Resolution

This article examines the role of mediator race and gender in perceptions of procedural justice as measure of accountability and representative bureaucracy in a national mediation program for complaints of employment discrimination at a large federal organization, the United States Postal Service. Mediation represents a forum of accountability in which employees may hold an employer accountable for violating federal law prohibiting forms of employment discrimination, in this case, race discrimination, sex discrimination, and sexual harassment. Representative bureaucracy theory suggests passive or symbolic representation when the demographics of public officials should mirror those of the public they serve. Some research suggests active representation when race or gender of a public official match those of a member of the public. During the period 1997-99, mediation exit surveys collected information about the nature of an employee’s claim. Using the nature of the claim as a proxy variable for a claimant’s race or gender, researchers examine complainants’ perceptions of mediation when the nature of the claim matches the demographics (race or gen-der) of the mediator. In this exploratory research, analyses show no statistically significant difference in satisfaction with the fairness of mediation process or mediator based on race of the mediator in race discrimination claims. However, there is statistically significantly lower satisfaction with mediation outcome based on mediator race as African American when it matches the nature of the claim of race discrimination. In all analyses of sex discrimination and sexual harassment claims, there are no statistically significant differences in satisfaction with the mediation process, mediator, or outcome when the mediator is female, although findings are borderline as to sexual harassment claims. These findings suggest the need for further research on accountability and bureaucratic representation in the diversity of mediator rosters. In addition, researchers need to control for the dispute system design and context within which mediation occurs