As an example of educational research, program evaluation has to justify the claims it makes about the nature and efficacy of the programs it investigates. This paper argues that both Campbellian quantitative experimentation as well as Stakean case study claim more than their foundationalist (empiricist) theories of knowledge permit. As a consequence, their claims to knowledge are no more valid than opinion or mere belief. To provide a theoretically more coherent framework for the just- ification of evaluation knowledge, a non-foundational, material-ist-pragmatist epistemology is introduced which does not produce such dualisms as mind/body, fact/value, and profes- sional/nonprofessional. As a result, this theory of knowledge makes evaluation practically relevant as well.