Data and spreadsheets
Summarize survey results with the sample size visible
Report each result next to the number of people who answered that question, not the number who started the survey. Say how many skipped it, and describe the findings as applying to respondents. Whether they stand for a wider group depends on how people were invited.
A text to start with
I ran a survey called [survey name] with [how people were invited]. [Number] people were invited and [number] started it. Here is the response table for each question: [question, answer counts, blanks]. Summarize the results in plain language. Show the response count beside every finding, say how many skipped each question, and note who the results do not describe. Use only the numbers I supplied.
Example request. Change the details to fit your situation.Survey result summary sheet
Survey name and dates: [ ] How people were invited: [ ] People invited: [ ] People who started: [ ] People who finished: [ ] For each question, copy the block below: Question wording: [ ] Who was shown this question: [ ] Response count for this question: [ ] Skipped or blank: [ ] Result, with count and percentage: [ ] Count of people or of selections: [ ] Limitation for this question: [ ] Who the results may not describe: [ ] Reviewer and date: [ ]
Denominator and wording problems
Percentages often use the wrong base. A question shown only to people who answered an earlier one has a smaller denominator than the whole survey, so check which group each percentage covers. Multiple-choice questions can produce more answers than participants, so counting people and counting selections must stay separate. Watch for phrases like "customers think" or "most people" when the data covers only respondents. With a very small response count, give raw numbers instead of percentages. Confirm that skipped answers are reported and not silently dropped.
Ask another agent to check the result
Recompute each percentage in this summary from the raw counts I supplied. Identify any percentage that uses the wrong denominator, any place where selections are counted as participants, and any sentence that implies the findings cover people who did not respond. For each problem, quote the sentence and state the correction.
Check the numbers against the export
Compare every count in the summary with the original export and note any difference. Write one sentence on how people were invited and who may be missing, and place it near the top. Decide whether small groups should be shown as raw counts. Ask one other person to read the summary and say who they think the results describe, then revise any line they read more broadly than you intended.