Reflexive Thematic Analysis — Braun & Clarke

Reflexive TA treats coding as an interpretive act carried out by a single analyst (or small team working collaboratively), with themes constructed fresh from each dataset. Codebook TA pre-specifies a shared code dictionary, often uses inter-rater reliability checks, and aims for replicability across analysts. The two approaches rest on different epistemological premises: reflexive TA is broadly constructivist and foregrounds the analyst's subjectivity; codebook TA tends toward a (post-)positivist ideal of researcher-independent findings.

There is no fixed minimum in reflexive TA. Braun and Clarke argue that sample size should be determined by the research question, data richness, and the depth of analysis sought — not by a universal number. Smaller samples of 6–15 in-depth interviews are common in rich, experiential studies; larger open-ended survey datasets may involve hundreds of responses. The guiding question is whether the dataset provides sufficient breadth and depth to address the research question.

Software is optional. NVivo, ATLAS.ti, or Dedoose can assist with organising, searching, and retrieving coded data, but they do not perform the interpretive work. The intellectual core of RTA — deciding what is meaningful, constructing codes, building themes, and writing reflexive memos — cannot be delegated to software. Many researchers do RTA with word-processing documents, spreadsheets, or printed transcripts and coloured pens.

A good theme makes a clear interpretive claim about a pattern of shared meaning in the dataset that is relevant to the research question. It is not a simple label for a topic people discussed. It should be internally coherent (the data extracts within it belong together), distinct from other themes, and analytically significant — it says something interesting about the data beyond summarising it.

Both are qualitative approaches using iterative coding, but they differ in goal and procedure. Reflexive TA aims to identify patterns of meaning across a dataset in relation to a research question; it does not aim to build a substantive theory. Grounded theory aims to generate a formal theory grounded in data through theoretical sampling, constant comparison, and theoretical saturation. Grounded theory also typically involves concurrent data collection and analysis, whereas RTA usually analyses a complete dataset.