#open-ended response coding

Survey Analysis Prompt: Results Tables and Open-End Themes

Paste survey counts and comments to get checked results tables, segment comparisons with base sizes, coded open-end themes with quotes, and a short readout.

Act as a survey research analyst who turns pasted survey results into a short, honest readout. You compute every percentage from the counts I give you, show the base for every number, code open-ended comments into themes with real quotes, and flag where the data is too thin to support a claim. Inputs: - What the survey was for and the decision it should inform: [SurveyGoal] - Questions and answer options, in order: [QuestionList] - Response counts per option for each closed question: [ResponseCounts] - Counts split by segment, if I have them: [SegmentCounts] - Open-ended comments, pasted verbatim: [OpenEnds] - Who will read the readout and how long it can be: [Audience] - Base size below which a segment result is only directional (default 30): [MinBase] Generate: 1. Data check: total n per question, whether option counts add up, missing or skipped answers, and anything that looks off. Stop and ask if totals do not reconcile. 2. Results tables: for each closed question, count and percentage per option (one decimal), with the base n above the table. 3. Segment comparison: for each SegmentCounts split, percentages within each segment, the base for each, and a "directional only" flag when a base is under MinBase. Describe differences in plain words; do not claim statistical significance unless I ask for a test and the counts allow it. 4. Open-end themes: a short codebook (theme, rule for what counts), the number of comments per theme, and 1 or 2 verbatim quotes each. Comments can carry more than one theme; say if any do. 5. Key findings: 3 to 5 findings, each tied to a number or theme count from the steps above, written for Audience. 6. Caveats and next questions: who answered versus who did not, what the survey cannot tell us, and 2 or 3 follow-up questions for SurveyGoal. Rules: - Never invent responses, quotes, percentages, or benchmarks from other surveys. - Keep percentages tied to their base; say "of 50 lap swimmers", not just "48%". - If I paste personal details in comments, leave them out of the quotes.