Name the denominator
A percentage needs a clearly defined base. Is it all invited customers, all submitted forms, all completed ratings or only those who answered one question? These can be different groups. In a fictional collection, 20 forms arrive but only 12 contain a rating. Nine positive ratings are 75% of those 12 ratings, not 75% of all customers.
Show how one answer changes the picture
For five answers, a single answer is twenty percentage points of the set. For fifty, it is two. That arithmetic does not provide a confidence interval or fix who chose to respond. The local count tool below shows these proportions and keeps the raw counts visible. It does not infer population opinion or determine whether a change is statistically meaningful.
Read the missing group
Ask who was eligible, invited, able and willing to answer. A convenient collection may omit important experiences. Increasing the count does not automatically remove that selection problem. Keep comments and operational context alongside the summary rather than letting a large percentage substitute for a clear explanation.
Use the result proportionately
A small set can reveal a question worth investigating. It may justify checking a confusing instruction or a broken link. It does not justify declaring a universal preference or penalizing a person. Record the proposed check separately from the claim the survey actually supports.
A count table you can explain
For the fictional twelve-rating set, report nine positive and three other valid ratings, plus eight submissions without a rating if there were twenty forms overall. Keep those categories clear instead of displaying only 75%. If the positive definition changes, recalculate transparently and preserve the earlier definition. The tool below cannot decide which answers belong in a category; that rule must come from the question and your declared review method, not from whichever grouping produces a more attractive result.
Sources used for this page
These records support the facts and comparisons above. Merchant-controlled records are labelled so you can separate product claims from independent evidence.
- AAPOR survey-research best practices — Standards and certification reference · aapor.org · Publisher independence not verified · checked 2026-10-01