Application guide
Thematic Analysis of Interview Data
Interviews are the most common data source for thematic analysis. This guide covers how to do thematic analysis of interview data specifically — from preparing transcripts through coding to building themes that hold across participants — and the documentation examiners expect from interview-based studies.
The six-phase framework still applies; what follows is how each phase plays out when your data is interview transcripts.
Preparing interview transcripts
Good thematic analysis starts with good transcripts. Decide on a transcription convention (usually verbatim for thematic analysis, including false starts and fillers where they carry meaning) and apply it consistently. Anonymise participants with stable identifiers (P1, P2…) so you can later show themes are supported across multiple people.
Read each transcript fully before coding. This familiarisation pass is analytic work, not admin — it is where your earliest sense of patterns forms.
Coding interview data
Code across the whole set of interviews before building themes, so themes reflect all participants rather than the first few you coded. Tag every segment relevant to your research question with a short, meaningful code anchored to the extract.
Keep codes close to what participants actually said. The strength of interview-based thematic analysis is that every theme can be traced back to specific things real people said — that traceability is exactly what makes it defensible.
Building themes across participants
Cluster recurring codes into candidate themes, then check each theme against the data. A key examiner concern with interview studies is over-reliance on a single vivid participant: a theme that rests on one person's account is weak. Aim for themes supported across multiple participants, and be honest in the write-up about how widely shared each theme was.
Select extracts for the write-up that span participants and illustrate the theme clearly, woven into an analytic narrative rather than listed as standalone quotes.
Documenting rigour for interview studies
For interview-based thematic analysis, reviewers often expect reporting against COREQ (for the interview method and conduct) alongside an account of how themes were developed. Keep a codebook, an audit trail of coding decisions, and a reflexivity statement; disclose any AI use.
QualIntel OS surfaces candidate extracts for each code across all your transcripts, has you confirm each one, flags single-participant over-reliance, and generates the audit trail and AI disclosure automatically — so the rigour expected of interview studies is documented as you work.
Frequently asked questions
How do you do thematic analysis of interview data?
Prepare verbatim, anonymised transcripts; read each fully (familiarisation); code every relevant segment across all interviews with short, meaningful codes; cluster recurring codes into candidate themes; review themes against the data; define and name them; then write up using extracts that span participants. The key with interviews is coding across the whole set before building themes.
How many interviews do you need for thematic analysis?
There is no fixed number — it depends on your question, sample homogeneity, and the richness of the data. Many postgraduate thematic analyses use somewhere between 6 and 25 interviews. The guiding principle is whether you have enough data to develop themes that are well-supported across participants rather than resting on one or two accounts.
How do you code interview transcripts for thematic analysis?
Go through each transcript and tag segments relevant to your research question with short codes anchored to the text, coding inclusively across all interviews before grouping codes into themes. Keep codes close to what participants said so every theme remains traceable to specific evidence. A codebook recording what each code means supports rigour.
How do you avoid bias in thematic analysis of interviews?
Code across all participants before building themes, seek evidence that challenges your emerging themes (not just confirms them), avoid over-relying on one articulate participant, and write a reflexivity statement on how your position shaped interpretation. An audit trail of your coding decisions makes your process transparent and checkable. QualIntel OS flags single-voice over-reliance and maintains the audit trail.
Related guides
Run your thematic analysis in QualIntel OS
AI surfaces candidate evidence, you confirm every decision, and the audit trail writes itself. One project free, no card.
Start for free