In this paper I provide a preliminary sketch of the types of logics of evaluation in the third sector. I begin by tracing the ideals that are evident in three well-articulated yet quite different third sector evaluation practices: the logical framework, most significant change stories, and social return on investment. Drawing on this analysis, I then tentatively outline three logics of evaluation: a scientific evaluation logic (systematic observation, observable and measurable evidence, objective and robust experimental procedures), a bureaucratic evaluation logic (complex, step-by-step procedures, analysis of intended objectives) and a learning evaluation logic (openness to change, wide range of perspectives, lay rather than professional expertise).These logics draw attention to differing conceptions of knowledge and expertise and their resource implications, and have important consequences for the professional status of the practitioners, consultants, and policy makers that contribute to and/or are involved in evaluations in third sector organizations.