Plan how I’ll code my qualitative data
Choose your analysis approach — thematic, grounded theory, content, framework or IPA — and get its coding sequence, a codebook template you can copy, and the rigour checks that particular tradition expects.
Free: the Methodology Selection Decision Tree
Your coding approach should follow from your methodology, not the other way round. This decision tree walks the choice from the question down. We’ll email you the PDF.
Coding means different things in different traditions
“I’ll code the interviews” describes five quite different activities depending on which tradition you are working in. The steps differ, the goal differs, and — the part that catches candidates out at viva — what counts as rigour differs too. Applying content analysis’s standards to an IPA study, or the reverse, reads as methodological confusion however well the coding itself was done.
- Thematic analysis — flexible and theme-driven; the sensible default for applied research, and the one with the clearest published procedure.
- Grounded theory — builds theory from data through open, axial and selective coding with constant comparison; demands theoretical sampling, which affects your recruitment plan.
- Content analysis — systematic categorisation, often with frequencies and a reported intercoder agreement statistic.
- Framework analysis — matrix-based, favoured in applied and policy research where findings must map onto pre-existing questions.
- IPA — detailed idiographic analysis of lived experience, with small samples analysed case by case before any cross-case work.
Choose on the basis of your research question and your epistemological position, not on which is quickest. Then code consistently and document as you go: a codebook and an audit trail support rigour in every one of these traditions, even though what you report on top of them — an agreement coefficient, or a reflexive account of your own position — depends on which you chose. Being explicit about that choice, and why, is what examiners are looking for.
Frequently asked questions
Which approach should I choose?
Let the question decide. If you want to describe patterns across accounts, thematic analysis. If you want to build a theory or model where none exists, grounded theory. If you need to quantify how often categories appear, content analysis. If findings must map onto pre-set policy or practice questions, framework analysis. If you want the texture of how a few people experienced something, IPA. Choosing for convenience and justifying afterwards is visible to examiners.
What goes in a codebook?
For each code: its name, a one-sentence definition, explicit inclusion and exclusion criteria, and an example quotation. The exclusion criteria matter most and are the part people skip — they are what keeps a code from quietly expanding to mean everything by the fortieth transcript. Version it, and date your changes.
How do I justify saturation in my methods chapter?
Define it before you claim it. State what you were watching for - new codes, new themes, or new dimensions of existing themes - and at what point new data stopped producing them. Say how many interviews you conducted after that point to confirm it. A bare assertion that 'saturation was reached' invites the obvious follow-up question. Note too that saturation belongs to some traditions and not others; in IPA it is not the relevant standard at all.
Do I need a second coder?
It depends entirely on your approach, and this is worth getting right. Content analysis conventionally reports an agreement statistic such as Cohen's kappa. Interpretivist approaches, including IPA and much thematic analysis, treat the researcher's interpretation as part of the analysis and value reflexivity, an audit trail and peer debriefing over agreement coefficients. Justify what you did in the terms of your own tradition rather than importing another's.
Can I use this alongside NVivo or ATLAS.ti?
Yes, and that is the intended use. The plan and codebook template are software-agnostic: decide the approach and structure your codebook here, then implement it in whichever package you use. Deciding the analytic approach inside the software, driven by what its interface makes easy, is how candidates end up with a codebook that does not match their stated methodology.
Is anything I enter stored or uploaded?
No. The plan is generated in your browser. Nothing you select is transmitted or saved, and no data of yours is ever uploaded.