What's in a mechanism? Development of a key concept in realist evaluation

Background: The idea that underlying, generative mechanisms give rise to causal regularities has become a guiding principle across many social and natural science disciplines. A specific form of this enquiry, realist evaluation is gaining momentum in the evaluation of complex social interventions. It focuses on 'what works, how, in which conditions and for whom' using context, mechanism and outcome configurations as opposed to asking whether an intervention 'works'. Realist evaluation can be difficult to codify and requires considerable researcher reflection and creativity. As such there is often confusion when operationalising the method in practice. This article aims to clarify and further develop the concept of mechanism in realist evaluation and in doing so aid the learning of those operationalising the methodology. Discussion: Using a social science illustration, we argue that disaggregating the concept of mechanism into its constituent parts helps to understand the difference between the resources offered by the intervention and the ways in which this changes the reasoning of participants. This in turn helps to distinguish between a context and mechanism. The notion of mechanisms 'firing' in social science research is explored, with discussions surrounding how this may stifle researchers' realist thinking. We underline the importance of conceptualising mechanisms as operating on a continuum, rather than as an 'on/off' switch. Summary: The discussions in this article will hopefully progress and operationalise realist methods. This development is likely to occur due to the infancy of the methodology and its recent increased profile and use in social science research. The arguments we present have been tested and are explained throughout the article using a social science illustration, evidencing their usability and value.