The three stages of building and testing mid-level theories in a realist RCT: a theoretical and methodological...

Randomised controlled trials (RCTs) of social interventions are often criticised as failing to open the 'black box' whereby they only address questions about 'what works' without explaining the underlying processes of implementation and mechanisms of action, and how these vary by contextual characteristics of person and place. Realist RCTs are proposed as an approach to evaluation science that addresses these gaps while preserving the strengths of RCTs in providing evidence with strong internal validity in estimating effects.

In the context of growing interest in designing and conducting realist trials, there is an urgent need to offer a worked example to provide guidance on how such an approach might be practically taken forward. The aim of this paper is to outline a three-staged theoretical and methodological process of undertaking a realist RCT using the example of the evaluation of a whole-school restorative intervention aiming to reduce aggression and bullying in English secondary schools.

First, informed by the findings of our initial pilot trial and sociological theory, we elaborate our theory of change and specific a priori hypotheses about how intervention mechanisms interact with context to produce outcomes. Second, we describe how we will use emerging findings from the integral process evaluation within the RCT to refine, and add to, these a priori hypotheses before the collection of quantitative, follow-up data. Third, we will test our hypotheses using a combination of process and outcome data via quantitative analyses of effect mediation (examining mechanisms) and moderation (examining contextual contingencies). The results are then used to refine and further develop the theory of change.

The aim of the realist RCT approach is thus not merely to assess whether the intervention is effective or not, but to develop empirically informed mid-range theory through a three-stage process. There are important implications for those involved with reporting and reviewing RCTs, including the use of new, iterative protocols.

Current Controlled Trials ISRCTN10751359 (Registered 11 March 2014).