By the Wavefield Research team · Published Sep 7, 2026
A dichotomous survey question offers exactly two mutually exclusive answers — Yes/No, True/False, Favor/Oppose. It is the fastest question in research and the cleanest to analyze, best for screening gates, factual recall, and go/no-go reads, and wrong wherever degree or genuine uncertainty is the real story.
Below: forty dichotomous survey questions, examples organized by the job they do, the four wording rules that keep binaries honest, and the part every other guide skips — what the yes-share can support statistically, and what your screener binary does to your fielding budget.
Copy what fits, but read each category note — the format changes jobs as the context changes, and two of these categories have rules of their own.
The dichotomous question's natural habitat: one clean gate, terminate on the wrong side. The yes-rate of your first screener is the incidence that sets your fielding cost.
Binary recall works when the behavior is memorable and the window is short. For frequencies, switch to a frequency scale instead of stacking yes/nos.
The registration and turnout questions every poll starts with — and the favor/oppose format that reads cleaner than agree/disagree.
Go/no-go reads on specific moments. Pair the No with an open-end follow-up and the binary becomes a diagnosis.
Binary items make honest baselines for tracking — the share saying Yes is a number leadership cannot reinterpret.
The one category where a binary is not just convenient but required — consent must be explicit, unambiguous, and its own question.
“Do you favor or oppose the plan?” beats “Do you support the plan?” — naming both options in the question counters acquiescence bias, the documented tendency to agree with whatever is asked. It is the cheapest wording fix in survey research, and the reason professional ballot questions read favor/oppose, approve/disapprove.
Mutually exclusive is easy with two options; exhaustive is the trap. “Do you own or rent your home?” strands respondents who live with family. Either widen the frame or add the honest third option — a binary that forces a false answer produces confident-looking noise.
“Was the product affordable and reliable?” is two questions in one costume, and a No is unreadable. Split double-barreled binaries always; each extra yes/no costs respondents two seconds, while a contaminated answer costs the analysis everything.
If a meaningful share of respondents genuinely sits in the middle — undecided voters, mixed feelings — a forced binary manufactures certainty that does not exist. The professional pattern: ask the binary, then probe the refusals (“As of today, do you lean toward Yes or No?”) and report leaners separately.
The acquiescence problem behind the first rule is well documented — Pew Research Center's questionnaire-design guidance treats balanced forced-choice wording as the standard fix, and it costs nothing to adopt. Where dichotomous items sit among the other formats — single choice, scales, matrices, open ends — is mapped in survey question types, with examples.
A dichotomous question produces a proportion, and proportions come with the cleanest error math in survey research. At 95% confidence, the margin on a yes-share is widest at 50/50: roughly ±9.8 points at 100 completes, ±6.9 at 200, ±4.9 at 400, ±3.1 at 1,000. Plan backward from the read you need — the arithmetic is in the sample size calculator. Comparing the yes-share between segments is a two-proportion z-test, and on weighted data it should run on effective bases, not raw counts — the convention our crosstabs apply automatically, with columns under 30 effective base excluded from testing rather than quietly over-trusted.
The budget math matters just as much and gets written about even less. A screener binary's yes-rate is your incidence: if 30% of general-population respondents pass “Have you purchased [category] in the past 3 months?”, you will pay for roughly three contacts per usable complete, and panel pricing scales accordingly. Every screener you add multiplies. The full cost mechanics — screen-out versus quota-full, the incidence table — are in survey screening questions, and they start with a yes/no.
A closed question offering exactly two mutually exclusive answers. The classic example: “Have you purchased from us in the past 12 months? Yes / No.” Variants include True/False, Favor/Oppose, and Approve/Disapprove. The format's strengths are speed, near-zero misinterpretation, and clean percentages; its cost is nuance — strong and mild positions collapse into the same answer.
It depends on whether the middle is real. For factual recall (Did you vote? Have you purchased?) a Not sure option mostly harvests laziness, so omit it. For attitudes where genuine uncertainty exists — an unfamiliar policy, a new concept — forcing a binary manufactures opinion; offer Not sure, then probe leaners as a follow-up. The test: would a thoughtful respondent honestly need the third option, or just prefer it?
The margin of error on a proportion is widest at a 50/50 split, which makes the math simple to plan: about ±9.8 points at 100 completes, ±4.9 at 400, ±3.1 at 1,000, at 95% confidence. If you will compare the yes-share between segments, each segment needs enough of its own — differences under roughly 10 points are rarely significant between groups of 100. Work backward from the smallest comparison that matters.
When the honest answer space has more than two regions. Binary formats suit gates, facts, and go/no-go reads; the moment you care about degree (how satisfied), frequency (how often), or selection among options (which of these), a scale or multiple choice carries more information for the same respondent effort. The common failure is stacking five yes/nos to fake a scale — one 5-point question does that job better.
Related: survey question types · 50 open-ended survey questions · survey screening questions · Likert scale examples
Brief the agent and it builds the screener gates, wires the terminates, and reports every yes-share with its margin of error. From $99 per study.
Start a study