Realist evaluation is becoming increasingly popular as an evaluation methodology. Its main objective is to uncover the mechanisms that lead to observed outcomes following an intervention and the contextual conditions that enabled this. The focus is on explaining why, for whom and in what circumstances an intervention works. It is a theory-driven approach and is explicitly method neutral, meaning that both quantitative and qualitative data collection methods can be used to unearth the underlying mechanisms that cause the intervention outcomes. In this review, we aim to map the methods used in realist evaluation studies, to draw lessons from the findings and to reflect on ways forward. We found that qualitative methods and interviews specifically are most commonly used in realist evaluations; that theory is often absent behind the methods and sampling techniques used; and that more innovative methods remain underexplored. We conclude the review by proposing four ways forward: (1) developing realist surveys, (2) exploring the relevance of innovative methods, (3) increasing the attention paid to sampling procedures and (4) strengthening the theory-driven nature of method. We believe that these four action points can strengthen the practice of realist evaluation and its outcomes.

Keywords: Data collection methods; Interview; Observation; Realist evaluation; Surveys; Theory-driven evaluation.

Copyright © 2022 Elsevier Ltd. All rights reserved.