By the Wavefield Research team · Published Sep 18, 2026
Demographic survey questions collect who the respondent is: age, gender, income, education, employment, race and ethnicity, region, and household. Eight questions cover almost every study. Below: the copy-ready option sets, census-aligned so weighting stays possible, plus the placement rules and the analysis discipline the wording guides leave out.
The guides ranking for this search list between 11 and 45 demographic questions. The count is the wrong contest — most studies should ask four to six, chosen because a crosstab or a weighting target needs them. That argument, and the mechanics behind it, are the second half of this page.
Each card is copy-ready: the wording, the option set, and the analysis note explaining what the options must line up with downstream. Two of these — age and gender — are hard enough to have their own full guides, linked from their cards.
“Which age group are you in?”
18–24 · 25–34 · 35–44 · 45–54 · 55–64 · 65 or older
Age bands are irreversible — they can never be un-merged after fielding — so cut them to match your weighting targets. Birth year is the reversible alternative for trackers.
“How do you describe your gender?”
Woman · Man · Non-binary · Prefer to self-describe: ____ · Prefer not to say
Gender data with five options weights against binary census frames by collapsing for the raking step and reporting the full question — never by asking the binary.
“What was your total household income in 2025, before taxes?”
Under $25,000 · $25,000–$49,999 · $50,000–$74,999 · $75,000–$99,999 · $100,000–$149,999 · $150,000 or more · Prefer not to say
Income draws the highest refusal rate of any demographic — 10–15% pick Prefer not to say, which is why the option must exist. Name the household and the year, and keep brackets on round census-style breaks.
“What is the highest level of education you have completed?”
Less than high school · High school · Some college · Bachelor's degree · Graduate or professional degree
Education steps stay at five to seven, census-aligned. Resist enumerating credentials: education is a weighting and crosstab variable, and finer steps produce cells too small to test.
“Which best describes your employment status?”
Full-time · Part-time · Self-employed · Student · Retired · Not currently employed
Employment categories must let every adult land exactly once — the retired and student options are what keep 'not employed' from blending three very different situations.
“Which of the following describe you? Select all that apply.”
Your census bureau's current categories · multi-select · plus a self-describe write-in
Race and ethnicity categories come from the census standard — in the US, the OMB framework updated in 2024 — because invented categories cannot be weighted against any population benchmark.
“Which region do you live in?”
The level you will analyze: provinces, states, or census regions — never street address
Region should be asked at exactly the granularity a crosstab will use. Finer geography raises privacy stakes and refusals while adding nothing to the analysis.
“How many people live in your household, including you?”
1 · 2 · 3 · 4 · 5 or more
Household size caps its top category deliberately — a 5-or-more bucket keeps the tail from producing single-respondent cells while preserving the family-size signal.
The two deep dives: every age format with its analysis cost is in how to ask about age in a survey, and the gender question's full treatment — wording, the two-step, weighting five options against binary frames, small-base reporting — is in gender survey questions. For race and ethnicity, the US standard changed in 2024 — the combined question, multi-select, and a Middle Eastern or North African category — and AAPOR's summary of the OMB updates is the practitioner's reference for aligning categories.
The standard advice — put demographic questions at the end — is right for the standard reason: personal questions cost goodwill, and a respondent twenty questions deep has invested enough to finish. Leading with identity questions also primes people to answer as their demographic, a distortion survey methodologists take seriously.
The exception the advice always omits: any demographic that drives a screener or a quota must be asked up front. A study capped at 50% women cannot discover gender on the final page — the quota engine needs the answer before the respondent spends fifteen minutes it may have to throw away. The working pattern is a split block: quota-relevant items in the first three questions, everything else at the end. On Wavefield the quota cells attach to the early questions and enforce themselves mid-survey — the mechanics are covered in survey screening questions. A purpose sentence above the block (“These questions help us understand who we heard from; they are optional”) measurably softens refusals, and every sensitive item carries Prefer not to say.
Here is the through-line that makes the census-aligned option sets above non-negotiable. After fielding, a serious study weights its sample to population benchmarks — raking the demographic shares to census or electorate targets. That only works if your categories can be laid on top of the benchmark's categories. Age bands that straddle census breaks, income brackets with creative boundaries, an education scale with nine steps — each is a weighting dimension you have quietly forfeited, discovered at analysis when it is unfixable. The mechanics are in survey weighting, explained.
The same discipline runs through reporting. Demographic subgroups are where base sizes quietly collapse: a 400-complete study reads its 65-and-older cell at maybe 60 respondents, and honest crosstabs mark which columns clear the significance floor and which are display-only — the conventions worked through in cross tabulation. And the cut-what-you-will-not-analyze rule has a mechanical test: before fielding, name the crosstab banner or weighting dimension each demographic feeds. On Wavefield, the agent programs the block from a plain-language brief — census-aligned bands, anchored escape options, specify boxes — and the same ids flow into quotas, weighting, and banner books without re-mapping.
Surveying professionals swaps half the block. Age and household income rarely earn their goodwill cost in B2B work; what the crosstabs actually need is the firmographic set: role in the decision (“Which best describes your role in choosing [category] tools?” — I decide alone · I share the decision · I recommend · I use but don't choose), company size in headcount bands (1–9 · 10–49 · 50–249 · 250–999 · 1,000 or more), industry from a short standard list with a self-describe catch, and seniority in four or five steps from individual contributor to executive.
The same three disciplines transfer unchanged: bands must match whatever benchmark you will weight or quota against (headcount bands should follow the breaks your market-sizing source uses), every category must give each respondent exactly one honest home, and the ask-only-what-you-analyze test applies with more force, not less — a B2B respondent's time is scarcer and the panels are thinner. The commonest B2B mistake is the screener-demographic blend: qualifying someone as a decision-maker and then asking their role again in the closing block. Ask once, early, because the quota needs it — then let the answer flow to the banner book.
Questions that collect who the respondent is rather than what they think: age, gender, income, education, employment, race and ethnicity, location, and household composition. They do three analytical jobs — segmenting results (crosstabs by age or region), checking the sample against the population, and supplying the targets weighting corrects to. A demographic question no crosstab or weighting plan will ever use is pure cost and should be cut.
The core eight cover almost every commercial study: an age band question, a gender question with a self-describe option, household income in brackets, highest education completed, employment status, race and ethnicity using census categories, region at the level you will analyze, and household size. Each has a standard wording and option set — the ones on this page are census-aligned so the answers can feed weighting directly.
Last, with one precise exception. Personal questions cost goodwill, and respondents answer them more willingly after they have invested in the survey — so the demographic block belongs at the end. The exception is any demographic that drives a screener or quota: a study quota-ed on age must ask age up front, because screening someone out on the final page wastes their time and your sample budget. Split the block when needed: quota items early, everything else last.
Only the ones the analysis plan will use — for most studies that is four to six of the core eight, not all of them. The test is concrete: name the crosstab or weighting target each question feeds. Every demographic item spends respondent goodwill and adds a refusal opportunity, and a questionnaire that asks income, religion, and marital status it never analyzes traded completion rate for nothing.
Related: how to ask about age · gender survey questions · survey question types · multiple choice survey questions
Brief the agent and it programs census-aligned bands, quota cells on the early questions, and weighting targets that actually match your options. From $99 per study.
Start a study