Thematic Analysis: Braun & Clarke’s Six Phases, Step by Step

What is thematic analysis?

Thematic analysis is a method for identifying, organising and interpreting patterns of meaning (themes) across a qualitative dataset, typically interview transcripts, focus group discussions, open survey answers or documents. Braun and Clarke’s 2006 paper in *Qualitative Research in Psychology* turned it from a loosely described practice into a clearly specified method, and it has since become one of the most frequently used approaches in psychology, health, education and the social sciences.

Its main strength is theoretical flexibility. Unlike grounded theory or interpretative phenomenological analysis (IPA), TA does not come with a built-in epistemology. You can use it from a realist position (reporting participants’ experiences as they describe them) or a constructionist one (analysing how meanings are produced in talk). That flexibility is also its main risk: you have to state your theoretical assumptions explicitly in your methodology chapter instead of leaving them implicit.

Throughout this article we use one running example. You are writing a master’s dissertation on how first-generation students develop a sense of belonging at university. You have conducted 12 semi-structured interviews of 45 to 60 minutes each and transcribed them verbatim.

Reflexive, codebook or coding reliability TA: which one are you doing?

Braun and Clarke (2022) stress that “thematic analysis” is a family of methods, not one single procedure. Before you start, decide which variant you are using, because it determines how you code, whether you work alone and whether reliability statistics make sense.

The six phases of reflexive thematic analysis

The 2006 paper named the phases familiarising yourself with the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the report. In their 2022 book, Braun and Clarke renamed several phases to make clear that themes are generated by you, not found. The logic is the same: you move from closeness to the data, through codes, to a coherent set of themes and a written analysis. You will move back and forth between phases many times.

Phases 1 and 2: familiarisation and coding in practice

Familiarisation means deep immersion. Read each transcript at least twice, ideally once while listening to the audio. Note what strikes you: recurring expressions, contradictions, emotional moments, what is conspicuously absent. These notes are not yet codes; they are your first analytic observations and part of your audit trail.

Coding in reflexive TA is systematic but open. You work through the entire dataset and attach short labels to segments that are relevant to your research question. Codes can be semantic (close to what the participant explicitly said) or latent (interpreting underlying assumptions or ideas). Code generously at first: a segment can carry several codes, and codes will change as your understanding deepens. Go through the whole dataset at least twice.

• Semantic code: “asks tutors for help only by email”
• Latent code: “belonging as something that must be earned through performance”
• Code the whole dataset with equal attention, not only the vivid interviews
• Keep a reflexive journal: what do you bring to the topic (for example your own first-generation background)?

Phases 3 to 5: from codes to themes

A theme captures a pattern of shared meaning united by a central organising concept. This is the most common point of failure in student work: many “themes” are really topic summaries (also called domain summaries) that collect everything participants said about a subject, for example “Experiences with lecturers”. Such a heading tells the reader what you asked about, not what you found.

To generate initial themes, cluster codes that seem to share a meaning, sketch a thematic map and ask what story each cluster tells. In reviewing, check two levels: does each theme fit its coded extracts, and does the set of themes fit the dataset as a whole? Themes may be split, merged or abandoned. In defining and naming, write a short paragraph per theme stating its scope, its core idea and how it relates to the other themes. Good names are informative and often use a short participant quote.

Phase 6: writing up and reporting quality

In reflexive TA, writing is part of the analysis, not something that happens afterwards. Your findings chapter should present each theme with an analytic narrative that makes an argument, supported by well-chosen data extracts. Extracts can be used illustratively (to show what you mean) or analytically (to comment on the specific wording). Avoid simply stringing quotes together.

In the methods section, describe your variant of TA, your theoretical assumptions (for example critical realist, experiential orientation), whether your analysis was inductive or deductive, whether you coded at a semantic or latent level, and how you practised reflexivity. Braun and Clarke advise against claiming that themes “emerged”, against counting how many participants mentioned a theme as evidence of its importance, and against citing data saturation as a stopping criterion, because in reflexive TA the meaning is constructed rather than exhausted.

• State that you used reflexive TA (Braun & Clarke, 2006, 2022) and why it fits your question
• Describe how coding developed over the six phases, with an example of a code and a theme
• Include a thematic map or a table of themes and subthemes
• Reflect on your positionality and how it shaped the analysis
• Departments sometimes require a specific procedure; follow your university’s guidelines (you can upload them to Professor Digital so the feedback uses them)

Common mistakes and how to avoid them

Most problems examiners flag in thematic analysis are not about software or technique but about conceptual clarity. Use the list below as a self-check before you submit. Our free coding tool at /codieren lets you build parent and child codes, attach memos to coded segments and export coded extracts, which helps you show your analytic process in the appendix.

• Mixing variants: claiming reflexive TA but reporting a kappa value
• Themes that are topic summaries or simply mirror the interview questions
• Too many themes (more than six or seven usually signals underdeveloped analysis)
• Quotes without interpretation, or interpretation without evidence
• No statement of theoretical assumptions or researcher positionality
• Using “themes emerged” language that hides your active role