Realist Approach to Qualitative Data Analysis

Abstract

Background A realist approach has gained popularity in evaluation research, particularly in understanding causal explanations of how a program works (or not), the circumstances, and the observed outcomes. In qualitative inquiry, the approach has contributed to better theoretically based explanations regarding causal interactions. Objective The aim of this study was to discuss how we conducted a realist-informed data analysis to explore the causal interactions within qualitative data. Methods We demonstrated a four-step realist approach of retroductive theorizing in qualitative data analysis using a concrete example from our empirical research rooted in the critical realism philosophical stance. These steps include (a) category identification, (b) elaboration of context-mechanism-outcome configuration, (c) demi-regularities identification, and (d) generative mechanism refinement. Results The four-step qualitative realist data analysis underpins the causal interactions of important factors and reveals the underlying mechanisms. The steps produce comprehensive causal explanations that can be used by related parties—especially when making complex decisions that may affect wide communities. Discussion The core process of realist data analysis is retroductive theorizing. The four-step qualitative realist data analysis facilitates this theorizing by allowing the researcher to identify (a) patterns, (b) fluctuation of patterns, (c) mechanisms from collected data, and (d) to confirm proposed mechanisms.