Beyond Text-as-Data: Integrating Structural Topic Modeling and Critical Discourse Analysis for Contextualized Transformations Research

Socio-ecological transformations are long-term reconfigurations of socio-technical and socio-economic provisioning systems, contested through struggles over pathways, timelines, and distributive consequences. Because these struggles are fought through public sense-making as well as policy and investment, transformation research increasingly turns to large text corpora. Yet a dilemma persists: computational text analysis scales but often remains descriptive, while qualitative, interpretative methods explain power and meaning in context but is typically limited to small samples. This paper develops a critical realist mixed- methods strategy that integrates Structural Topic Modeling (STM) and Critical Discourse Analysis (CDA). Drawing on critical methodological pluralism, STM is treated as an extensive mapping device that identifies candidate demi-regularities without reifying them as causal laws. CDA then supplies the intensive interpretation needed to situate these patterns, reconstruct their argumentative and ideological work, and support retroductive reasoning about generative mechanisms in open, stratified systems. Thus, by integrating those methods, we enhance both the explanatory power and critical depth of transformation research.