Mapping mental models through an improved method for identifying causal structures in qualitative data

Abstract

AbstractQualitative data are commonly used in the development of system dynamics models, but methods for systematically identifying causal structures in qualitative data have not been widely established. This article presents a modified process for identifying causal structures (e.g., feedback loops) that are communicated implicitly or explicitly and utilizes software to make coding, tracking, and model rendering more efficient. This approach draws from existing methods, system dynamics best practice, and qualitative data analysis techniques. Steps of this method are presented along with a description of causal structures for an audience new to system dynamics. The method is applied to a set of interviews describing mental models of clinical practice transformation from an implementation study of screening and treatment for unhealthy alcohol use in primary care. This approach has the potential to increase rigour and transparency in the use of qualitative data for model building and to broaden the user base for causal‐loop diagramming.