What thematic analysis is
Thematic analysis is a method for identifying, organizing and interpreting patterns of meaning (themes) across a dataset, usually interview transcripts. It is flexible, widely used in business and management research and suits questions about experience, perception and process. Braun and Clarke's reflexive approach is the most cited version, and it treats analysis as an active, interpretive process rather than a mechanical sorting of data.
Be clear about the difference between a topic and a theme. A topic is a subject people mentioned (onboarding). A theme is an idea about the data with a point to make (new hires judge a bank by how quickly managers include them in real client work, not by the formal induction).
Six phases of analysis
| Phase | What you do | Output |
|---|---|---|
| 1. Familiarize | Transcribe, read and reread; note first impressions | Notes and a feel for the data |
| 2. Generate initial codes | Label meaningful segments across the entire dataset | A list of codes with extracts |
| 3. Search for themes | Group codes into candidate themes | Draft theme map |
| 4. Review themes | Check themes against extracts and the full dataset | Refined themes |
| 5. Define and name themes | Write a definition and scope for each theme | Theme definitions |
| 6. Write up | Weave themes, quotations and analysis into an argument | Findings chapter |
The phases are not strictly linear. You will move back and forth, especially between phases 3 and 4.
Coding: a worked example
A code is a short label for a segment of data that is relevant to your question. Code for meaning, not just for topics, and keep codes close to the participants' language at first.
From extract to code to theme (hypothetical)
| Extract (participant 7, relationship manager) | Initial code | Candidate theme |
|---|---|---|
| "In my first week I sat through eight hours of compliance slides. I only felt like part of the team when my manager took me to a client meeting in week three." | Formal induction felt generic; belonging came from client contact | Belonging through real work |
| "My supervisor checked in every Friday, even if it was just ten minutes. That made it easy to ask silly questions." | Regular short check-ins lower the barrier to asking questions | Accessible supervision |
| "Nobody told me who owned the client files. I spent a month unsure who to ask." | Unclear ownership of tasks and information | Role ambiguity |
Keep a codebook as you go: each code, a short definition, an example extract and when to use it. Code the whole dataset, not just the interesting parts, and revise codes when you find that two overlap or one is too broad. Software such as NVivo, ATLAS.ti or MAXQDA helps with organization, but it does not do the analysis for you; a spreadsheet works for small projects.
Building and testing themes
Cluster related codes into candidate themes, then test each one.
- Does the theme have a central idea? If you cannot state it in a sentence, it is probably a topic.
- Is there enough data? Supported by several participants and extracts, not one voice.
- Are themes distinct? Overlapping themes should be merged or sharpened.
- Do themes answer the question? Drop interesting material that does not.
- Does the whole set tell a coherent story? Check how themes relate to each other.
| Theme | Central idea | Codes it contains | Participants |
|---|---|---|---|
| Belonging through real work | New hires feel part of the team when given real client responsibility early | Client contact; early responsibility; formal induction felt generic | 11 of 16 |
| Accessible supervision | Short, regular contact lowers the cost of asking for help | Weekly check-ins; approachable manager; fear of asking | 13 of 16 |
| Unclear roles and ownership | Ambiguity about who owns what delays integration | File ownership; unclear escalation; mixed messages | 9 of 16 |
Counting participants per theme is optional in qualitative work and is not a measure of importance, but it can show how widespread a pattern is. Do not turn it into statistics.
Writing up the findings
Each theme gets a section with a definition, an analytic paragraph that makes the point and two or three quotations that illustrate it, each followed by interpretation. Quotations show; analysis explains. A findings chapter that is a string of quotations without commentary reads as raw data.
Analytic structure of a theme (hypothetical)
Theme 1: Belonging through real work. Participants described belonging not as a product of induction but of early inclusion in client work. One manager said, "In my first week I sat through eight hours of compliance slides. I only felt like part of the team when my manager took me to a client meeting in week three" (P7). Eleven of the sixteen participants made a similar contrast between formal induction and meaningful tasks. This suggests that onboarding effectiveness depends less on the volume of information delivered than on the speed with which new hires are given a legitimate role, which extends research that treats induction as a single event.
Anonymize quotations with participant codes, remove identifying details and tidy only trivial filler words (and say so). Link the themes back to the literature and the framework in the discussion chapter, as described in our data analysis chapter guide.
Working on this assignment now? Get a price for help with your paper.
Get an instant quoteA codebook entry
A codebook keeps coding consistent over weeks of work and lets another person apply your codes. Each entry needs a definition, rules for use and an example.
| Field | Entry |
|---|---|
| Code name | Early client contact |
| Definition | Participant describes being included in real client interactions during the first four weeks of employment |
| Include when | The participant links the contact to feeling trusted, useful or part of the team |
| Exclude when | The contact is only observation with no role, or happens after month two (code as Later client contact) |
| Example | "My manager took me to a client meeting in week three and let me present the numbers." |
| Related codes | Belonging; Formal induction felt generic |
Update the codebook as you learn. Record the date and the reason for each change so the audit trail shows how your thinking developed.
Naming themes well
| Weak name (topic) | Strong name (claim) |
|---|---|
| Onboarding | Belonging comes from real work, not induction slides |
| Managers | Short, regular contact makes asking for help easy |
| Communication problems | Unclear ownership delays integration |
| Training | Training that arrives after the task is learned too late to help |
Naming a theme as a claim forces you to say what it contributes. If you cannot write the claim, return to the data: you may have a topic, or two themes merged together.
Analytic memos
Memos are short notes you write while coding about what you notice, doubt and wonder. They turn coding into analysis.
Memo (hypothetical)
Memo, interview 7 and 11. Both describe feeling like part of the team only after client contact, and both are from branch teams of fewer than ten. Interview 4 (large central office) mentions the opposite: induction was enough. Possible pattern: team size or proximity to clients changes what builds belonging. Check against interviews 2, 9 and 13, and consider adding a code for team size.
Date your memos and keep them. They show how themes were built, support your methodology chapter and often provide the sentences of your discussion.
Rigor and quality
| Criterion | What you can do |
|---|---|
| Credibility | Member checking, triangulation with documents or other participants, thick description |
| Dependability | Keep an audit trail: codebook versions, memos, decision log |
| Confirmability | Reflexive journal about your assumptions; second coder on a sample; negative case analysis |
| Transferability | Describe the setting and participants so readers can judge relevance |
Saturation, the point at which new interviews add little to the themes, is a useful guide but not a guarantee; justify your sample size by the richness of the data and the scope of the question. Be honest about your own position, especially if you are an insider researcher.
- State the approach and cite it For example Braun and Clarke.
- Show the coding process An example table and a codebook in the appendix.
- Use quotations with analysis Never leave a quote to speak for itself.
- Be explicit about interpretation Say what the data suggest, not that they prove.
If you want help with qualitative analysis, you can order DBA dissertation help or MBA dissertation help.