A case study approach taken with three years of data on a hospital's trauma program indicate that, although service user data is important, its political importance far outweighs evaluation value and utility, especially for hard-to-access samples. Implications of this challenge to traditional assumptions for program evaluation are discussed. (SLD)