Realist Evaluation

Developed by Ray Pawson and Nick Tilley in the 1990s, realist evaluation is a theory-driven approach for evaluating policies, programmes, projects, and interventions. It investigates complex causal relationships using Context-Mechanism-Outcome (CMO) configurations to answer: “What works, for whom, in what circumstances, and why?” Rooted in scientific realism, it assumes outcomes depend on contextual and human factors. Rather than offering universal judgements of effectiveness, it refines programme and middle-range theories. Methodologically neutral, it often employs mixed methods to test and revise hypotheses. Closely related approaches include realist synthesis and realist research, which share the same ontological and epistemological foundations. Emerging from theory-driven evaluation and the principle of generative causation, realist evaluation treats programmes as theories to be tested. Its distinct data collection strategies, such as the “learner–teacher cycle” in realist qualitative interviews, facilitate iterative theory development. The entry traces its origins, philosophy, and applications, highlighting its strength in evaluating complex interventions.