Data
Summarise a survey
The stats, plus the two or three findings that actually hold up.
Fill it in
The export.
The decision this feeds.
Compare the groups, not just the whole.
Your prompt
Analyse these responses. What I need to decide: [your what you want to know] Report the response rate and who is missing before anything else. A survey answered by 12% of users measures the 12% who answer surveys, and every finding below that line inherits the bias. Give the distribution, not just the mean. A satisfaction score of 3.4 made of fours and fives with a block of ones is a completely different business problem from one made of threes. Compare the segments. The aggregate usually hides the finding. Two or three findings, each with the number behind it and a note on how confident you are. Say plainly where the sample is too small to conclude anything, rather than reporting a difference between nine people and eleven people as a result.
Use Summarise a surveyOpens with everything above already filled in.
Why this works
A survey answered by 12% of users measures the 12% who answer surveys, and every finding inherits that bias. This reports the response rate first, gives distributions rather than means, and says plainly where the sample is too small to conclude anything.