Drawing on a qualitative study of an educational reform and its external evaluation, I describe how a well-intentioned but poorly conceptualized evaluation helped perpetuate asymmetries in the generation and use of evaluation findings. This article explores this project's failure to clarify evaluator roles, identify intended users and expected subjects, and generate formal agreements so that the evaluation could meet the information needs of its diverse stakeholder groups. I offer suggestions for minimizing evaluators' contributions to the creation or maintenance of these kinds of asymmetries and, when these prove to be of limited avail, when to opt out of an evaluation project.