Qualitative coding
Qualitative coding is the process of labeling passages of qualitative data, such as interview transcripts or open-ended answers, with codes that capture their meaning, so that patterns across many responses can be found, counted and interpreted.
How it works
A code is a short label for a single idea in a passage: "price confusion", "trust in advisor". Coding can be inductive, with codes created from the data, or deductive, with codes taken from a framework, and most projects mix the two. Analysts apply codes, keep definitions in a codebook, and revise them as the data teaches them more.
Coding is the bridge between raw data and findings. Once passages are coded, analysts can pull every passage on a code, compare how different groups talked about it and group codes into themes. Rigor comes from clear definitions, more than one coder, and a record of decisions.
A worked example
The code "document back-and-forth" turns up in 27 of 36 interviews with merchants who signed up to a payments company, and comparing by segment shows it in 14 of 15 interviews with sole proprietors. The finding points the company at one form for one segment.
Related
Automate this in Pulse
Pulse's first-pass themes act as provisional codes, and a coder checking one can read every passage it groups in its full context.