By the Wavefield Research team · Published Sep 8, 2026
A multiple choice survey question offers a fixed list of answer options, either single select (pick one) or multi select (pick all that apply). Below are thirty real examples with their option sets included — because the guides ranking for this search explain the types at length and then, remarkably, provide no actual questions.
After the bank: the four option-writing rules, the single-versus-multi-select decision, and how to analyze multiple choice survey answers without the two mistakes that quietly ruin reports.
Six categories of five. Adapt the bracketed placeholders, keep the structural choices — the exclusive None options, the specify boxes, the capped select-alls — because those carry the method.
Mostly single select
Multi select territory
The awareness funnel, question by question
The classifier questions
Single select, census-comparable
Single select, aligned to analysis targets
Exhaustive and mutually exclusive: every respondent has exactly one honest home (or a defined set, in multi select). Overlapping ranges ($50–$100, $100–$250) and missing categories are the two most common questionnaire bugs in circulation.
An Other option with a specify box catches the categories you did not predict — and tells you when the option list needs fixing (an Other above roughly 10% means the list is wrong). None of these must be exclusive: selecting it should clear everything else, mechanically.
Respondents over-pick early options, so rotate the substantive list — but anchor Other, None of these, and Prefer not to say at the bottom, where convention expects them. Order effects are documented survey methodology, not superstition.
Past about seven options, respondents stop reading and start satisficing. If the honest list is longer, the question usually wants to be two questions — a category pick, then a follow-up within it — or a short open end coded afterward.
The response-order research behind the rotation rule is summarized in Pew Research Center's questionnaire guidance. On Wavefield these mechanics are declarative rather than manual: choices rotate with per-option anchors, None of these enforces its exclusivity, and specify boxes attach to any option — the agent programs all three from a plain-language brief. Where multiple choice sits among the other formats is mapped in survey question types; the two-option special case has its own 40-example dichotomous bank.
Single select analysis is the clean case: each option's share of the answered base, summing to 100%, weighted where the sample needs correction, and compared across segments with significance tests on effective bases. Multi select is where reports quietly go wrong. The base is respondents, not selections — if 400 people answer and 240 tick “Dashboard,” that is 60%, and the column will legitimately sum past 100% because people select several things. Report “% who selected,” never “% of all selections”; the second framing makes popular options look small and is the most common multi-select misread in circulation.
The second discipline: option codes are nominal, so there is no mean to take — averaging “How did you hear about us?” codes produces confident nonsense. That temptation belongs to scales, where the means debate is at least a real debate. Our crosstabs apply these conventions automatically — respondent bases on multi selects, weighted shares, significance letters, and no means where means have no meaning — and the full workflow from raw answers to report is in how to analyze survey results.
A closed question offering a fixed list of answer options — single select (pick one: “How did you first hear about us?”) or multi select (“Which features have you used? Select all that apply”). It is the workhorse format of survey research because it produces countable, comparable data with near-zero coding effort. The craft lives in the option list: exhaustive, mutually exclusive, five to seven options, with Other and None of these handled deliberately.
Multi select fits when the true answer is a set: features used, brands heard of, channels seen. Single select fits when you need the one that matters most — and “Which ONE best describes…” often extracts more decision-useful data than a select-all sweep. A robust pattern is both: a multi select for breadth, then a single select from the chosen options for priority. Cap select-alls (“up to 3”) when you want signal rather than inventory.
Five to seven for most questions. Fewer collapses real distinctions; more triggers satisficing, where respondents pick the first plausible option and move on. Brand lists and category screens legitimately run longer, which is exactly when rotation matters most, so position bias spreads evenly instead of inflating whichever brand sits first. Dropdowns hiding 30 options do not fix a long list; they hide it.
Single select: shares of the answered base, summing to 100%, compared across segments with significance tests. Multiple choice survey answers from multi selects work differently: the base is respondents, not selections, so percentages legitimately sum past 100% — report “% who selected,” never “% of all selections.” And because option codes are nominal, means are meaningless; that discipline belongs to scales. Weighted shares, honest bases, and significance letters carry the analysis.
Related: survey question types · dichotomous survey questions · 50 open-ended survey questions · how to ask about age
Brief the agent and it writes the questions, rotates the options, anchors the escapes, and reports respondent-based shares with significance letters. From $99 per study.
Start a study