Codebook
A codebook is the agreed set of rules a team uses to code qualitative data, so that two analysts reading the same passage label it the same way. Each code gets a name, a meaning, boundaries that say where it stops, and a passage that shows it in use.
How it works
A codebook usually starts with codes drawn from the research questions or an existing framework, then grows as analysts read the data and add codes for patterns the framework did not anticipate. Each entry carries a short label, a full definition, inclusion and exclusion criteria, and a typical quote.
The team tests the draft codebook on a sample of transcripts, compares how different coders applied it and tightens any definition that produced disagreement. Codes are often arranged in a hierarchy of parent and child codes. Keeping a record of each change creates an audit trail, which lets reviewers see how the analysis developed.
A worked example
For 45 interviews about airport parking, an operator's customer team starts with a codebook of 12 codes. After two coders test it on 5 transcripts, they disagree most on "confusion", so the team splits it into "shuttle confusion" and "payment-kiosk confusion", each with its own rules and example quote.
Related
Automate this in Pulse
Draft entries for a codebook can be taken from Pulse's theme names, with the quotes under each serving as the inclusion examples a good code definition needs.