You have 14 interview transcripts and 180 pages of text; how do you turn them into findings an examiner can trust? Thematic analysis is the usual answer, and Braun and Clarke’s 2006 paper is the usual citation, but citing the six phases is not the same as doing them. Examiners can tell: “themes” that are just the interview questions restated, quotations stacked without interpretation, a sentence claiming that themes “emerged”, and no trace of how 400 codes became five themes.
This article is part of our guide How to Write a Master’s Thesis by the Mastermind PhD team and relates to our Literature Review support service. Sources last checked: November 2026.
This guide takes the six phases one at a time with a worked coding example, explains what the authors’ later work on reflexive thematic analysis changes, shows how to evidence trustworthiness, compares the software options, and ends with templates for the methodology and findings chapters.
What is thematic analysis? Thematic analysis is a method for identifying, analyzing and interpreting patterns of meaning (themes) across a qualitative dataset such as interview transcripts, focus groups or documents. In the version developed by Braun and Clarke (2006, 2019, 2022), it proceeds through six phases, from familiarization with the data to writing up, and in its reflexive form treats coding as an interpretive act by a researcher who records and reflects on their own role, rather than as a procedure to be made “reliable” between coders.
Which version of thematic analysis are you doing?
Braun and Clarke (2019, 2021) distinguish three families, and your methodology chapter must say which one you follow, because they make different assumptions and demand different quality checks.
| Family | What a theme is | Coding | Quality evidence | Typical use |
|---|---|---|---|---|
| Coding reliability TA | A topic decided largely in advance; themes often map onto interview questions | Codebook fixed early; two or more coders; inter-coder agreement (Cohen’s kappa) | Agreement statistics | Applied, team-based or mixed methods projects with a positivist orientation |
| Codebook TA (framework analysis, template analysis) | Partly predetermined, partly developed from the data | Structured codebook or template that evolves | Audit trail, documented codebook versions | Health services research, policy evaluation |
| Reflexive TA | A pattern of shared meaning organized around a central concept, developed through analysis | Open, organic, evolving; usually one researcher; no codebook required | Reflexivity journal, worked examples, transparency about decisions | Most master’s theses in the social sciences, education and health |
Most student projects fit reflexive TA. If you choose it, do not then report Cohen’s kappa; the authors regard inter-coder reliability as incompatible with the approach, and examiners who know the literature will ask why you mixed the two.
What are the six phases?
The phase names below give the 2006 label first and the 2022 wording in brackets, because both appear in the literature. Time shares are typical for a master’s thesis with 10 to 15 interviews.
| Phase | What you do | Output | Time share |
|---|---|---|---|
| 1. Familiarizing yourself with the data (familiarization) | Transcribe or check transcripts, read the whole dataset twice, note first impressions | Transcripts plus a page of notes per interview | 20% |
| 2. Generating initial codes (coding) | Label every segment relevant to the research question, systematically across all data | A code list with extracts attached (often 150 to 400 codes) | 30% |
| 3. Searching for themes (generating initial themes) | Cluster codes into candidate themes around shared meaning; draw a first thematic map | 6 to 10 candidate themes with their codes | 15% |
| 4. Reviewing themes (developing and reviewing themes) | Check each theme against its extracts, then against the whole dataset; merge, split, discard | 3 to 6 themes that work; a revised map | 15% |
| 5. Defining and naming themes (refining, defining and naming themes) | Write a definition of each theme and its central concept; choose names that say something | One paragraph per theme plus final names | 10% |
| 6. Producing the report (writing up) | Weave extracts into an analytic narrative connected to the research question and the literature | The findings chapter | 10% |
The phases are recursive: coding often restarts after a theme is reviewed, and writing exposes themes that do not hold together.
📱 Get support with your literature review — We agree the databases and inclusion criteria with you; you receive the review matrix with a Turnitin report.
How do you code? A worked example
Coding means attaching a short label to every segment of data that says something relevant to the research question. Codes can be semantic (staying close to what the participant said) or latent (naming an underlying idea). The example below comes from a study of moral distress among ICU nurses.
| Data extract (participant N07) | Code(s) | Type |
|---|---|---|
| “We knew he wasn’t going to make it, but the family wanted everything done, so we kept going for eleven more days.” | Continuing treatment judged futile; family wishes override clinical judgment | Semantic |
| “Nobody asked the nurses. We just carry out the plan.” | Exclusion from decisions; nurses as executors, not participants | Latent |
| “You go home and you can’t switch it off. I had a glass of wine every night that month.” | Carrying distress home; coping through alcohol | Semantic |
| “If you say something, you’re the difficult one.” | Speaking up is penalized; silence as self-protection | Latent |
Code the whole dataset, not just the vivid passages. Keep a code list with a one-line definition for each code; when a new extract fits an existing code, attach it, and when it does not, create a new code. Expect to collapse near-duplicates at the end of phase 2.
How do you move from codes to themes?
A theme is not a topic. “Communication” is a topic that any interview about hospital work would produce; a theme is a claim about patterned meaning, such as “nurses protect themselves by staying silent because speaking up is punished”. Braun and Clarke (2019) call the first kind a topic summary and warn that examiners increasingly recognize it as a sign of shallow analysis.
To cluster, print the code list, group codes that express a shared meaning, and write a sentence that states what the group is about. From the example above, the codes “exclusion from decisions”, “nurses as executors”, “speaking up is penalized” and “silence as self-protection” can form a candidate theme: “Powerless at the bedside: distress comes from executing decisions nurses did not shape and cannot challenge.” The codes about futile treatment and about carrying distress home belong to other candidate themes. Draw the map: themes as boxes, codes as branches, with arrows for the relationships you want to argue later.
How do you review, define and name themes?
Review at two levels. First, read every extract under a theme and ask whether they really share the central meaning; if not, move the extracts or split the theme. Second, reread the full dataset with the candidate themes in hand and ask whether the set tells a coherent story about the research question and whether anything important is missing. Themes built on one participant’s words, or that only restate an interview question, usually fail here.
Then write a definition. A useful format: one sentence for the central concept, two or three sentences on scope (what belongs and what does not), and one sentence on how the theme relates to the others. Name the theme with a phrase that carries meaning, often borrowing participants’ words: “You’re the difficult one: the cost of speaking up” works harder than “Theme 3: Communication”.
How do you demonstrate trustworthiness?
Reflexive TA does not use inter-coder agreement, so you need other evidence. Nowell and colleagues (2017) map Lincoln and Guba’s (1985) criteria onto concrete TA practices; the ones below fit a master’s timeline.
| Criterion | What it asks | What you can show |
|---|---|---|
| Credibility | Do the findings fit the data? | Prolonged engagement with transcripts; extracts for every claim; peer debriefing with your supervisor |
| Transferability | Can readers judge relevance to other settings? | Thick description of participants, setting and recruitment |
| Dependability | Is the process traceable? | Audit trail: dated code lists, thematic maps at each phase, memos recording decisions |
| Confirmability | Are interpretations grounded rather than imposed? | Reflexivity journal; a positionality statement in the methodology chapter |
Keep every version of the code list and the thematic map, dated. Three pages of dated maps in an appendix answer the question “how did 380 codes become five themes” better than any paragraph.
Which software should you use?
| Tool | Strengths | Limits | Fit for a master’s thesis |
|---|---|---|---|
| NVivo | Mature coding, queries, visual maps; widely taught | License cost; learning curve | Good if your university provides it |
| ATLAS.ti or MAXQDA | Similar features; strong for mixed data; student licenses | Cost; another interface to learn | Good |
| Taguette or QualCoder | Free and open source; basic coding and export | Few queries; limited visualization | Enough for 10 to 15 interviews |
| Word and Excel | No cost; comments as codes, a spreadsheet as the code list | Manual; error-prone beyond about 10 transcripts | Workable for small datasets |
Software stores and retrieves; it does not analyze. Whichever tool you use, the methodology chapter should say that coding and theme development were done by you, with the software used for data management.
How do you report thematic analysis in the thesis?
In the methodology chapter, a template: “Data were analyzed using reflexive thematic analysis (Braun & Clarke, 2006, 2022). After transcription, the researcher read each transcript twice and kept familiarization notes. Coding was inductive and predominantly semantic, producing 312 codes across 14 transcripts, managed in NVivo 14. Codes were clustered into eight candidate themes, reviewed against the coded extracts and the full dataset, and consolidated into four themes. Dated code lists and thematic maps are provided in Appendix D, and a reflexivity journal was kept throughout.”
In the findings chapter, present each theme with a heading, a one-paragraph definition, an analytic narrative that interprets the pattern, and two to four extracts labeled with pseudonyms or codes (N07, nurse, public hospital). Keep extracts short and let your commentary carry the argument. Report how many participants contributed to a theme only if that number is meaningful; reflexive TA treats frequency as a weak indicator of importance.
Which mistakes do examiners flag?
- Themes that are the interview questions restated (“Experiences of workload”, “Views on management”).
- Eight or ten themes, each with two sub-themes, for twelve interviews: too many for depth.
- Pages of quotations with a sentence of commentary between them.
- Claims that themes “emerged from the data”, which hides the researcher’s decisions.
- Reporting “data saturation” as the stopping rule while using reflexive TA, which the authors argue against.
- Counting participants per theme and treating the count as evidence of importance.
- Analyzing before all transcripts are complete and never returning to re-code the early ones.
📱 Ask on WhatsApp now — 5 years of experience across all disciplines and 14 academic services. Two-stage payment (80/20); revisions are free until supervisor approval, within 30% of the chapter.
Frequently asked questions
How many themes should a master’s thesis have?
Usually three to six. Fewer than three suggests under-analysis; more than six usually means some are topics rather than themes. Depth per theme matters more than the count, and each theme should answer part of the research question.
Do I need a second coder or inter-coder reliability?
Not in reflexive TA, where the researcher’s interpretation is part of the method and quality is shown through transparency and reflexivity. If your program or supervisor requires agreement statistics, you are doing coding reliability TA and should describe it as such.
Can I do thematic analysis with only 8 or 10 interviews?
Yes. Rich, well-conducted interviews with 8 to 15 participants are typical for a master’s thesis. Justify the number by the scope of the question and the depth of the data, not by a saturation claim.
Is it acceptable to use AI tools to code my transcripts?
Only if your university permits it and you disclose it. Automated suggestions can speed up data management, but the interpretive work must be yours, and any tool that uploads identifiable transcripts to an external server may breach your ethics approval and data protection rules.
What is the difference between thematic analysis and qualitative content analysis?
Content analysis typically produces categories that are counted or described, often with a predefined coding frame. Thematic analysis in the reflexive form develops interpretive themes around shared meaning and does not rely on frequency. They overlap in practice; choose the one whose logic you can defend.
Sources
- Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa
- Braun, V., & Clarke, V. (2019). Reflecting on reflexive thematic analysis. Qualitative Research in Sport, Exercise and Health, 11(4), 589–597. https://doi.org/10.1080/2159676X.2019.1628806
- Braun, V., & Clarke, V. (2021). One size fits all? What counts as quality practice in (reflexive) thematic analysis? Qualitative Research in Psychology, 18(3), 328–352. https://doi.org/10.1080/14780887.2020.1769238
- Braun, V., & Clarke, V. (2022). Thematic analysis: A practical guide. SAGE.
- Nowell, L. S., Norris, J. M., White, D. E., & Moules, N. J. (2017). Thematic analysis: Striving to meet the trustworthiness criteria. International Journal of Qualitative Methods, 16(1), 1–13. https://doi.org/10.1177/1609406917733847
- Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. SAGE.
- Braun, V., & Clarke, V. Thematic analysis [website]. The University of Auckland. https://www.thematicanalysis.net/ (accessed October 2026)
Sources last checked: November 2026. This article is general educational guidance and does not replace your university regulations or your supervisor’s instructions. How we prepare our content.