Thematic Analysis in 6 Steps: Braun & Clarke’s Method Explained

What Is Thematic Analysis?

If you have just finished your interviews and now face a pile of transcripts and no idea what to do with them, this guide is for you. Learning how to do thematic analysis is one of the most useful skills a qualitative researcher can pick up, because it turns messy, human talk into a clear, defensible argument — and it does so without demanding the statistical training of quantitative work or the theory-building machinery of grounded theory.

Thematic analysis is a method for identifying, analysing, and reporting patterns — called themes — within qualitative data. The version almost every student uses comes from Virginia Braun and Victoria Clarke, who set out their approach in a 2006 paper, “Using thematic analysis in psychology” (Qualitative Research in Psychology, 3(2), 77–101). Their framework, often called reflexive thematic analysis, is deliberately flexible: it works with interviews, focus groups, open-ended survey answers, diaries, and even social media posts.

A theme is not just a topic that comes up a lot. It is a pattern of shared meaning, organised around a central idea, that helps answer your research question. “Students talked about stress” is a topic. “Stress is experienced as a private failure, not a normal part of research” is a theme — it says something interpretive, and it can be argued with evidence from your data.

Before You Start

Before touching a single transcript, get three things in order. First, a clear research question — thematic analysis is guided by what you want to find out, and without a question your coding will drift. Second, your data in a workable form: interviews transcribed verbatim, including pauses and laughter if they carry meaning. Third, a way to organise your work. You can code with highlighter pens and sticky notes, in a Word document with comments, or in software such as NVivo. None of these is “more correct” than the others; choose the tool you will actually use consistently.

One warning up front: thematic analysis is not a quick counting exercise. The six steps below describe a recursive process, not a conveyor belt. Expect to move back and forth between phases as your understanding of the data deepens. That is normal — in fact, it is the point.

The 6 Steps of Thematic Analysis

Step 1: Familiarise Yourself with the Data

Read every transcript from start to finish at least once, ideally twice, before you code anything. Take notes as you go: what surprises you, what repeats, what contradicts what you expected. This is active reading, not skimming. If you conducted the interviews yourself, this is also the moment to notice how your own presence shaped the conversation — reflexive thematic analysis treats the researcher as part of the process, not as a neutral observer.

A practical tip: keep a short “first impressions” memo, one page per transcript, listing the two or three ideas that struck you most. You will come back to these memos in Step 4.

Step 2: Generate Initial Codes

Coding means labelling small segments of data with short phrases that capture what is interesting about them. A code is a building block, not a conclusion. Work systematically through the whole dataset, giving equal attention to every transcript — including the boring-looking ones, which often contain the most surprising material.

For example, a student saying “I stare at a blank page for hours before I write a single sentence” might be coded “blank-page paralysis”. Another saying “My supervisor’s comments make me rewrite everything” could be coded “feedback overwhelm”. Keep codes close to the data at this stage. You might generate 80 or 100 codes from twelve interviews, and that is fine. Most will later be merged or dropped.

Code for meaning, not just for words. If two people describe the same experience in different words, they deserve the same code. And code generously: it is easy to discard a code later, but hard to recover an idea you never labelled.

Step 3: Generate Themes

Now start sorting. Look for codes that cluster together around a shared idea and gather them into candidate themes. This is the most creative part of the process. Physically grouping codes — on a table, a whiteboard, or a mind map — helps you see connections you would miss in a list.

Some codes will form a theme quickly. Others will not fit anywhere yet; keep them in a “miscellaneous” pile rather than forcing them in. A theme needs enough data to be worth arguing about, but there is no magic number of codes or extracts. What matters is whether the theme captures something meaningful in relation to your research question.

At the end of this step you should have a handful of candidate themes, each with its codes and supporting extracts gathered underneath. Sketch a thematic map — a simple diagram showing how the themes relate to one another — even if it is rough. It will guide the next step.

Step 4: Review the Themes

This is the quality-control step, and skipping it is the most common reason examiners criticise thematic analyses. Reviewing happens at two levels.

First, check each theme against its coded extracts. Do all the extracts under a theme actually fit together, or are you looking at two different ideas wearing one label? If a theme is incoherent, split it, rename it, or move the misfits elsewhere. Second, check the themes against the whole dataset. Re-read your transcripts and ask: do these themes tell the story of the data, or have I missed something important? This is where your first-impressions memos from Step 1 earn their keep.

Be honest here. A theme that only fits two extracts and overlaps heavily with another theme should probably be merged or demoted to a sub-theme. A strong analysis usually ends with three to five main themes, each distinct and each clearly supported.

Step 5: Define and Name the Themes

For each theme, write a short definition: in two or three sentences, what is this theme about, what is its central idea, and what does it say in answer to your research question? If you cannot define it clearly, the theme is not ready yet.

Then give each theme a concise, informative name. Names should capture the essence of the theme, not just label its topic. “The supervisor relationship” is a label. “The supervisor as gatekeeper: feedback that opens or closes the writing” is a name that tells the reader what the theme argues. Short, punchy names are fine too, as long as they carry meaning.

Step 6: Produce the Report

The final step is writing up. A good thematic analysis report is not a list of themes with quotes pasted underneath; it is an analytic narrative. Each theme gets its own section in which you build an argument, using data extracts as evidence — and you must do more than present the quote. Explain what the extract shows, why it matters, and how it connects to your research question and to the wider literature.

Choose extracts that are vivid and representative, not just the ones that agree with you. Including a dissenting voice strengthens your analysis: it shows you tested your themes against the whole dataset. End the results section by pulling the themes together — what, taken as a whole, do they say about your research question?

Worked Example: From Codes to Themes

Imagine you interviewed twelve master’s students about their experience of writing a thesis. This is an illustrative example, not real data — but it shows how the process works in practice.

After coding, you have extracts like these: “I stare at a blank page for hours” (code: blank-page paralysis); “I only write the night before a supervision meeting” (code: deadline-driven writing); “My supervisor’s comments make me rewrite everything” (code: feedback overwhelm); “I feel like everyone else finds this easy” (code: comparing myself to others); “I rewrite the same paragraph ten times” (code: perfectionist rewriting).

Sorting these codes, two candidate themes emerge. The first, “Writing as an emotional struggle”, gathers blank-page paralysis, perfectionist rewriting, and comparing myself to others — codes that all point to writing being experienced as something painful and private. The second, “Deadlines and supervision drive the writing”, gathers deadline-driven writing and feedback overwhelm — codes showing that progress depends on external pressure rather than internal routine.

At Step 4 you test them. The first theme holds: all its extracts share the idea that difficulty is felt as a personal failing. The second theme needs work — “feedback overwhelm” does not quite fit with “deadline-driven writing”, because one is about being judged and the other about being pushed. You split it: “The supervisor as judge” and “Deadlines as the real supervisor”. Three themes, each distinct, each telling part of the story.

Common Mistakes to Avoid

1. Counting instead of analysing. A theme is not a theme because ten people mentioned it. Frequency can be a clue, but a theme earns its place by capturing something meaningful, not by winning a vote.

2. Using your interview questions as themes. If your themes are “Experiences of supervision”, “Challenges of writing”, and “Suggestions for improvement”, you have simply repeated your interview schedule. Themes should cut across your questions.

3. Themes that are only topic summaries. “Students talked about stress” describes the data; it does not analyse it. Push every theme until it makes a claim.

4. Quotes without interpretation. A long extract followed by “As this participant shows…” is not analysis. Tell the reader what the extract reveals and why it matters.

5. Skipping the review phase. Themes that overlap, contradict each other, or rest on a single extract are the hallmark of an analysis that jumped from Step 3 to Step 6. Reviewing is where a mediocre analysis becomes a good one.

Thematic Analysis vs Other Methods

If your goal is to report what your participants’ experiences mean, and you want a method your examiners will recognise and trust, thematic analysis is usually the right choice. If you need to build a theory, look at grounded theory instead.

Frequently Asked Questions

How long does thematic analysis take? For a small study of ten to fifteen interviews, expect several weeks of part-time work. Familiarisation and coding are the slowest phases; rushing them produces thin themes. Build the time into your thesis plan rather than treating analysis as a weekend job.

How many interviews do I need? There is no fixed number, but most student projects using thematic analysis work with eight to twenty interviews. What matters is whether your data are rich enough to answer your question. Stop collecting when new interviews stop adding new ideas — a point researchers call saturation.

Do I need software like NVivo? No. Software helps you organise large datasets and search across codes, but many excellent thematic analyses have been done with printed transcripts, highlighters, and a spreadsheet. Software organises your thinking; it does not do the thinking for you.

Can I use thematic analysis in my dissertation? Yes — it is one of the most widely accepted qualitative methods in student research, across psychology, education, health, and the social sciences. Just name the approach you are following (Braun & Clarke’s reflexive thematic analysis is the standard reference), describe your six steps transparently, and show your working.

What is the difference between a code and a theme? A code is a label for a small piece of data — a phrase, a sentence, an idea. A theme is a broader pattern of meaning built from many codes. Think of codes as bricks and themes as the walls: you need many of the first to build one of the second.

Key Takeaways

Thematic analysis turns interview talk into an argument, and Braun & Clarke’s six steps give that process a structure examiners recognise: familiarise, code, generate themes, review them, define them, and write them up. The quality of your analysis lives in Steps 4 and 5 — the reviewing and defining that most students rush. Take your time there, keep your themes distinct and interpretive, and let your extracts do the arguing, not just the illustrating.

Reference: Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. For the updated guidance, see Braun, V., & Clarke, V. (2019). Reflecting on reflexive thematic analysis. Qualitative Research in Sport, Exercise and Health, 11(4), 589–597.