Data & analysisQualitative analysisdata
Code open-text feedback without erasing dissent
Creates a transparent themebook with counts and counterexamples.

Ready to use
Prompt
Given consented [FEEDBACK RESPONSES] and [RESEARCH QUESTION], define a small themebook with inclusion and exclusion rules, then code a sample twice to refine it. Output theme counts, representative anonymised paraphrases, contradictory comments and a not-coded bucket. Do not present frequencies as population prevalence when the sample is self-selected. Keep sensitive identifiers out of examples. Provide an audit trail of coding changes and a question for human review. Check that a response can carry multiple themes and that negative cases are not silently merged away.