Read a survey without fooling yourself

Here are survey results: {{paste questions, response counts, and open-text answers}}. It was distributed via {{channel}} to {{who}}, with {{n}} responses out of {{invited}}.

1. State the response rate and who is likely over- and under-represented given the channel. Say what conclusions this makes unsafe.
2. Flag leading, double-barrelled, or ambiguous questions and how they bias the answers.
3. For quantitative items: report distributions, not just means, and note where n is too small to segment.
4. For open text: code the responses into themes with counts and two verbatim quotes each. Distinguish what people asked for from the underlying problem they described.
5. Separate what respondents said they'd do from what they've done — and treat stated intent as weak evidence.
6. Give the three findings that are solid, the three that are suggestive, and the ones people will over-claim from this deck.

Do not report percentages on subgroups under n=30 without labelling them as directional.