As a result of recent federal legislation, states are required to provide mediation as a dispute resolution option to parents and school districts involved in special education disputes. Although states are mandated to provide mediation services, the law gives little guidance as to how states should select a mediation model, select mediators, or measure mediator performance and program success. Many states rely on participant satisfaction surveys and statistics on the number of agreements reached to demonstrate program effectiveness and quality. This paper examines what such survey data measures and whether it is a reliable indicator of the procedural and substantive fairness of the mediation process. Specifically, this article first examines the historical context of Pennsylvania's efforts to create the Pennsylvania Special Education Mediation Service. Next, the article looks at two sources of program data: one quantitative and the other qualitative. The former consists of approximately two thousand participant post-mediation questionnaires collected over three years, from 1997-2000. Participants answered questions about mediator performance, fairness of the process, for of PaSEMS office, and suitability of surroundings. Parties indicated almost unanimous satisfaction with process and mediator. 82% indicated they were satisfied or very satisfied w/ the results. Comments on questionnaires were also evaluated, most of which reflected procedural concerns. The latter source consists of the performance evaluations of eight special education mediators. Mediators scored best on managing the startup, expressing empathy nonverbally, and managing personalities. Mediators scored least well on moving the parties toward an improved relationship. Although more data is needed, Pennsylvania's experience suggests the need to examine the relationships that exist between mediation agencies and the institutions in which they are administratively housed, the development and rigorous implementation of relevant mediator evaluation instruments, stricter standards for mediator training, and a way of relating measures of program quality to stakeholder goals and program design.