> Qualitative Analysis
Path B — Reflexive Coding
The same data, analyzed reflexively — and why this path rejects IRR. · 11 min
This is the reflexive thematic analysis (RTA) path (Braun & Clarke, 2006) (Braun & Clarke, 2022). Same six broad phases as any TA — familiarize, code, generate themes, review, define, write — but a different philosophy: the analyst is the instrument, and themes are actively constructed, not discovered.
Semantic vs. latent coding
- Semantic codes stay close to what’s said.
- Latent codes interpret what’s underneath — assumptions, meanings, framings.
RTA leans into latent coding. Watch the same excerpts from Lesson 4, recoded reflexively. Click a span: the codes now read the data interpretively (identity, social risk) rather than just labeling surface content.
It was a linked-list thing, a null pointer exception. . Honestly for missing something basic. Before that I .
I never really felt in control of it. . The compiler kept throwing an , and I , so I .
The error said , so I , and eventually I just .
Click a highlighted segment to see how it was coded and why.
Compare to Lesson 4: "think I was dumb" was a tidy in vivo code there; here it becomes
managing how I am seen — a reading of identity work. Neither is “more correct”; they
serve different analytic aims.
Why RTA rejects inter-rater reliability
If themes are the product of one researcher’s situated interpretation, then asking a second coder to reproduce them — and scoring the “agreement” — measures the wrong thing (Braun & Clarke, 2019). A low κ wouldn’t mean the analysis is wrong; a high κ wouldn’t make it insightful.
In the next lesson we generate, review, and define this path’s themes; then we put them side by side with Path A’s.