The limits of realist evaluation: Surfacing and exploring assumptions in assessing the best value performance regime

This article builds on an earlier experiment in applying Realist Evaluation (RE) techniques to a set of Best Value Reviews (BVRs) undertaken in a single English local authority. That experiment used a range of assumptions regarding context (C), mechanisms (M) and outcomes (O) that restricted the possible pathways in the resulting CMO causal loop. They were 'heroic' in nature and left largely implicit. The article subjects those assumptions to rigorous criticism. Five hypotheses are tested concerning the nature of the context within which the BVRs occurred; the potentially skewed nature of the review mechanisms chosen; the impact of process outcomes and goals-setting problems on BVR outcomes; the scope for strong linkages to be formed between context, mechanism and outcomes, such that deterministic effects ensue; and the need to define the boundaries of the evaluand (the BVR). Realist evaluations have typically focused on individual services or programmes. It is contended that RE methods need to be adapted to address cumulative impacts on policy and organizational culture that are inherently political in nature.