Using Model-Based Clustering to Improve Qualitative Inquiry: Computer-Aided Qualitative Data Analysis, Latent Class Analysis, and Interpretive Transparency

A combination of computer-aided qualitative data analysis (CAQDAS) and latent class analysis (LCA) can substantially augment the qualitative analysis of textual data sources used in third-sector studies. This article explains how to employ both techniques iteratively to capture often implicit ideas and meaning-making by third-sector leaders, donors, and other stakeholders. CAQDAS facilitates the coding, organization, and quantification of qualitative data, effectively creating parallel qualitative and quantitative data structures. LCA facilities the discovery of latent concepts, document classification, and the identification of exemplary qualitative evidence to aid interpretation. For third-sector research, CAQDAS and LCA are particularly promising because diverse stakeholders usually do not share homogenous views about core issues such as organizational effectiveness, collaboration, impact measurement, or philanthropic approaches, for example. The procedure explained here provides a rigorous method for discovering and understanding diversity in perspectives and is especially useful in medium-n research settings common to third-sector scholarship.