Article révisé par les pairs
Résumé : Central to realist evaluation is its focus on underlying generative mechanisms that cause outcomes in particular contexts, often presented as context–mechanism–outcome configurations. The first author brought together 22 realist evaluation experts in a Delphi technique-inspired exercise to explore the potential of creating a database from context–mechanism–outcome configurations identified in the empirical literature, to enhance learning about what context–mechanism–outcome configurations are and to further the cumulation of knowledge. The exercise, and the comments it evoked, showed why developing such a database is not straightforward and the views on its desirability and utility are mixed. The main reasons are the lack of common concepts to organize the database and the problem of the triple hermeneutic, different uses and non-uses of the context–mechanism–outcome configuration heuristic, and different perceptions of the quality of a context–mechanism–outcome configuration. The discussion points raised in this exercise offer valuable insights into realist reasoning and highlight interesting avenues for further research and debate.

Documents en relation