Braun & Clarke thematic analysis: the complete guide
Almost every qualitative methods section that mentions thematic analysis cites the same source: Virginia Braun and Victoria Clarke's 2006 paper. It gave the method a clear, teachable shape — six phases you can follow and, just as importantly, defend to an examiner. This guide walks through those phases, then covers how Braun and Clarke later reframed their own approach as reflexive thematic analysis, and the errors that cost marks.
Finish your thematic analysis on time — with total ease, without sacrificing academic rigour. The thematic analysis tool runs the six phases across your own studies or transcripts — free to start.
What Braun & Clarke thematic analysis is
Thematic analysis is a method for spotting, organising, and interpreting patterns of meaning across a qualitative dataset — interview transcripts, open survey responses, focus groups, or documents. Braun and Clarke's contribution was not to invent it but to spell it out: a step-by-step process a student could actually carry out, with named phases and clear decisions at each one.
That clarity is why the 2006 paper became one of the most cited articles in the social sciences, and why later teaching pieces such as Ahmed et al. (2025) still build directly on their framework.
Two things set it apart from neighbouring methods. It is not tied to a single theory, so you can use it whether you lean realist or constructionist. And it works at two levels: semantic, staying close to what people explicitly said, or latent, reading for the assumptions underneath. You decide which, and you say so in your write-up.
Still deciding whether this is the right method for your question at all? Braun and Clarke wrote a paper on exactly that choice, which we cover in reflexive TA vs. other qualitative methods.
The six phases (steps) of thematic analysis
The phases are numbered, but they are not a one-way conveyor belt. You will loop back — a theme that falls apart in phase four sends you back to recode in phase two. Braun and Clarke are explicit that the process is recursive. Naeem and colleagues give a helpful worked account of moving through the steps toward a finished model in their step-by-step process (2023), if you want a second walkthrough alongside this one.
1. Familiarise yourself with the data
Read everything, then read it again. Before you code a single line you need to know your material well enough to recall where things were said. Most researchers transcribe their own interviews for exactly this reason: the typing forces attention. Jot early notes as you go, but hold off on formal coding until you have been through the whole dataset at least once.
2. Generate initial codes
Work through the data and label anything that looks relevant to your question. A code is a short tag for a feature of the data — “fear of judgement”, “workarounds for missing tools”. Code generously; you can always drop codes later, but you cannot cluster what you never tagged. Keep the extract that each code sits on, because you will need those quotes to defend your themes. If typing out that first pass by hand is the part slowing you down, the tool on this site will run it for you: paste your transcripts or survey text and it hands back codes with the exact quote behind each one, ready for you to check and adjust.
3. Search for themes
Now shift from codes to patterns. Sort your codes into candidate themes — broader ideas that pull several codes together and say something about your question. This is where sticky notes, tables, or a whiteboard earn their keep, or, if you coded your data with the tool above, it has already clustered the codes into candidate groups for you, so this phase starts from a sorted draft instead of a blank table. A theme is not a topic word like “support”; it is a claim, such as “support only helps when it arrives before the deadline”.
4. Review the themes
Test each candidate theme twice. First, check the coded extracts inside it hang together. Second, check the themes make sense against the full dataset. Themes that overlap get merged; themes with only one thin quote behind them get cut or folded elsewhere. You should end with a set that is distinct, each earning its place.
5. Define and name the themes
Write a short definition of what each theme is and is not. If you cannot describe a theme in a sentence or two, it is probably doing too much and needs splitting. Names should be sharp and readable — an examiner skimming your contents page should grasp the analysis from the theme names alone.
6. Produce the report
Write the analysis so that the argument, the themes, and the evidence read as one piece. Each theme gets its definition, the quotes that support it, and your interpretation of what it means for the research question. The write-up is analysis, not a data dump: quotes illustrate the point you are making, they do not replace it.
From 2006 to reflexive thematic analysis
Since 2006 Braun and Clarke have refined how they talk about the method, and the current name for their version is reflexive thematic analysis. The phases are much the same. What changed is the thinking behind them. They now reject the idea that themes “emerge” from data on their own, as if waiting to be found. Themes are produced by a researcher who brings a perspective, and that perspective is a resource, not a contaminant to be scrubbed out.
The practical upshot: in reflexive thematic analysis you are not chasing a single “correct” coding that a second coder would replicate. You are building a considered interpretation and being honest about your part in it — the subject of our guide on writing reflexivity in thematic analysis. That is a different goal from a codebook approach, where a fixed set of codes is applied for consistency. If your project needs the codebook style instead, the difference between letting codes surface and applying a fixed scheme is the subject of our guide on inductive vs. deductive coding.
Mistakes that cost marks
Three problems show up again and again in student thematic analyses, and examiners spot them fast.
- Topic summaries dressed as themes. “Communication” is a bucket, not a theme. A theme makes a point: “communication broke down whenever roles were unclear”. If your theme name is a single noun, it is probably still a topic.
- Quotes doing the analysis. A string of quotes under a heading is data, not interpretation. Say what the quote shows and why it matters for your question.
- Claiming themes “emerged”. In reflexive thematic analysis this phrasing signals you have misread the method. You constructed the themes; write as though you did.
Running the analysis without losing weeks
Finish your thematic analysis on time — with total ease, without sacrificing academic rigour. The thematic analysis tool on this site codes your findings and clusters them into candidate themes with verbatim quotes, so you start phase three with a structured draft instead of a blank page. For a walk through what to paste and how to read the output, see how to use the tool, or read our guide on synthesising findings across studies.
Frequently asked questions
How many phases are in Braun and Clarke's thematic analysis?
Six: familiarising yourself with the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the report. Braun and Clarke set them out in their 2006 paper, and the phases are the same in the later reflexive version — what changed is the thinking behind them, not the steps.
What's the difference between thematic analysis and reflexive thematic analysis?
Reflexive thematic analysis is Braun and Clarke's own update to their 2006 method, not a different method. The six phases stay the same; what changes is that themes are treated as built by the researcher's judgement rather than as patterns that “emerge” from the data on their own.
How many themes should a thematic analysis have?
There is no fixed number. Braun and Clarke's guidance is that themes should be distinct and each earn its place against the full dataset, which in practice usually lands somewhere between three and six for a typical student project. A study with twelve thin themes almost always has several that should be merged.
Is thematic analysis inductive or deductive?
It can run either way. Coding data-up with no predefined code list is inductive; coding against a fixed scheme you set before reading the data is deductive. Reflexive thematic analysis is usually run inductively, but nothing about the six phases requires it — see our guide on inductive vs. deductive coding for how to justify whichever you pick.
References
- Braun, V. and Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), pp.77–101. doi:10.1191/1478088706qp063oa
- Naeem, M., Ozuem, W., Howell, K. and Ranfagni, S. (2023). A Step-by-Step Process of Thematic Analysis to Develop a Conceptual Model in Qualitative Research. International Journal of Qualitative Methods, 22. doi:10.1177/16094069231205789
- Ahmed, S.K., Mohammed, R.A., Nashwan, A.J., Ibrahim, R.H., Abdalla, A.Q., M. Ameen, B.M. and Khdhir, R.M. (2025). Using thematic analysis in qualitative research. Journal of Medicine, Surgery, and Public Health, 6, p.100198. doi:10.1016/j.glmedi.2025.100198