Open boxes in a survey feel like free qualitative gold. Treated carelessly, they become percentage claims based on twenty people who bothered to type.
We theme comments after the closed questions are charted. Themes explain why a score moved; they do not replace the score. Counts of themes stay descriptive—“mentioned often,” “raised by a few”—unless the sample and coding rules justify more.
Bilingual surveys common in Hong Kong need a coding note. We record whether themes were coded in the respondent’s language or after translation, because that choice affects nuance.
A useful exhibit pairs a bar chart with three short verbatims that illustrate the modal theme and one contrary voice. The contrary voice keeps the room from over-reading a single story.
If the open ends contradict the closed measures, we surface the tension rather than burying it. That tension is often where the next research question lives.