Glossary
Open-text analysis
Open-text analysis is the systematic processing of free-form written responses — employee comments, exit-interview notes, anonymous question replies — to extract structured patterns rather than rely on the original analyst's selective reading.
Open-text analysis is the family of techniques used to convert unstructured written responses into a structured representation that can be summarised, counted, and tracked over time. In an HR context, the input is typically employee survey comments, exit-interview notes, manager 1-1 logs, or anonymous "what is getting in your way?" responses.
Modern open-text pipelines combine three steps: extraction (parsing and de-identifying the raw text), coding (assigning each statement to one or more themes), and aggregation (rolling the codes up to clusters with counts and confidence bands). LLM-assisted coding now makes this affordable on every survey rather than once a year.
The structural alternative — selectively quoting the most articulate comments at the bottom of a deck — systematically undercounts the patterns that the most disengaged employees stop articulating altogether.
How it is measured
- Coverage: the share of free-text responses that are coded vs. ignored. A defensible analysis codes 100%.
- Code consistency: inter-coder agreement (κ) above ~0.7 for human-coded analyses; for LLM-assisted coding, a stratified sample is recoded by a second model or a human and the agreement reported.
Dr. Tim Hough · ISBN 979-8-9965397-1-0 · Buy the book →