Inductive vs. deductive coding, explained
Coding is the heart of thematic analysis, and there are two broad ways to do it: let codes emerge from the data (inductive) or apply a predefined framework (deductive). Your methods section needs to say which you did and why — choosing well shapes everything downstream.
Inductive coding (data-driven)
In inductive coding, you start with no preconceived categories and derive codes directly from what the data says. It's the usual choice for exploratory questions and for synthesising findings across studies, because it surfaces patterns you didn't know to look for. The trade-off is consistency: without a fixed scheme, codes can proliferate and drift between coders — which is why examiners ask how you kept coding systematic.
Deductive coding (framework-driven)
In deductive coding, you bring a codebook — a predefined set of codes from prior research or theory — and tag the data against it. It's fast, consistent, and comparable across datasets, which suits theory-testing designs and replication studies. The risk is tunnel vision: a fixed framework can miss anything it wasn't designed to catch, so reviewers expect you to note findings that didn't fit the codebook.
Most real analysis is hybrid
In practice researchers combine both: a deductive backbone from the research questions, plus inductive room for unexpected themes to emerge. Braun and Clarke treat the distinction as a continuum rather than a switch. The analysis tool on this site codes inductively — semantic codes derived from your studies' findings, each with a verbatim quote — which gives you a data-driven first pass you can then check against your own theoretical framework.
Try it on your own studies — free
Paste the findings of 3–15 studies, choose a framework — Braun & Clarke and more — and watch codes cluster into themes with a verbatim quote behind every one. First 3 studies free, no signup.
Start your free analysis