Using Models in Realist Research

Roundtable Purpose of the session: Aim: To explore the use of models in research and evaluation that uses a realist approach. Background: Realist researchers and evaluators often make use of models of various kinds: for example causal loop diagrams, flow diagrams and ad-hoc illustrations. Typically, these models aim to represent what is happening that we can't see directly and provide a scaffold for building and / or connecting programme theories or highlighting patterns of emergence. Such models, however sophisticated, often depict the association of events but don't explain the connections. This may indicate a limitation of current models to adequately model generative causation. This roundtable will enable participants to look in detail at some current examples of practice, bring their own ideas and discuss some fundamental questions relating to the use of models in realist research. Outline of session: 0-20 mins. The roundtable will be facilitated by four academics with varied experience of realist research and philosophy. We will present three examples of models used in realist research. Each academic will outline the context and development of the model, explain its intention and highlight any perceived limitations. 20-50 mins. Participants will be welcomed to share any models they have used or find particularly useful. We will provide flipchart paper and markers for people to illustrate their ideas. We will facilitate a discussion framed by the following questions. • What is distinctive about a realist model, if anything? • What are models useful for, and what are they less useful for? • What is the relationship between a model and a theory in the realist theory-led approach? 50-60 mins. We will then pose, in simple terms, challenges to using models in realist research. This will address issues of: successionist and generative causation and the basis of models, and reductive/non-reductive explanations of the social world. 60-80 mins. We will lead a further discussion which tackles the question… • How can the realist generative picture of causation be modelled? We will have prepared positions to support this part of the discussion if required. 80-90 mins. The chair will summarise key learning points and future directions.