Foucauldian Discourse Analysis — Power, Knowledge, and the Construction of Truth

Both draw on Foucault's ideas, but CDA (associated with Norman Fairclough) foregrounds linguistic analysis — clause structure, transitivity, modality — and typically works at the level of specific texts. FDA operates at a more macro, historical level, examining the epistemic rules (discursive formations) that make statements possible across institutions and time. CDA tends to link discourse to social structures empirically; FDA tends to treat discourse itself as constitutive of those structures. The two approaches can be combined but should not be conflated.

FDA is primarily a documentary and textual method. Policy documents, institutional reports, media texts, legislative records, and historical archives are classic data sources. Interview transcripts can also form part of the corpus, but they are treated as sites where discourses are reproduced — not as windows onto individual experience. If accessing subjective experience is the goal, phenomenology or interpretive phenomenological analysis is more appropriate.

There is no fixed rule. The corpus should be purposive and strategically diverse — capturing the range of institutional sites and historical moments in which the discourse is produced. Adequacy is reached when additional texts are reproducing the same discursive regularities (analogous to saturation in grounded theory). Transparency about selection rationale and corpus boundaries is more important than a large volume of documents.

Yes, with appropriate scope management. A well-bounded corpus (e.g., national policy documents from a defined period, or a set of media articles on a specific issue) is achievable at master's level. The key requirement is solid theoretical grounding in Foucault's concepts before analysis begins. Reading Mills (2004) alongside primary Foucault (especially The Archaeology of Knowledge) and a methods text such as Hook (2001) is recommended before designing the study.

Reflexivity in FDA means acknowledging that you, the researcher, are also a subject produced by discourse — your conceptual categories, your sense of what is 'obvious' or 'problematic' in the data, and your interpretive choices are all shaped by the discourses you inhabit. This is typically addressed through a positionality statement, a research journal, and transparent discussion of how interpretive decisions were made. The goal is not to eliminate the analyst's perspective but to make it visible and accountable.