A questionnaire is just the list of questions. That’s it. A survey is the whole project around that list: who you ask, how you send it, how you collect the answers, and what you do with them once they’re in. Every survey contains a questionnaire. A questionnaire on its own isn’t a survey.
So that’s the survey vs questionnaire answer in two sentences, and if you came here to settle an argument, you’re free to go.
But stick around, because the difference pays off in actual numbers, and that’s the fun part of this topic. I spend my days building survey and form software (Uplup), so these projects cross my desk constantly, and I keep seeing the same failure: someone writes a beautiful set of questions and never designs the survey around it. I’ll show you exactly what that costs below. Also, we’ll cover quizzes, because they’re the third sibling everyone leaves out of this comparison.
What is a questionnaire?
A questionnaire is a structured set of written questions designed to gather information from people. Just the questions, their order, and their answer formats (multiple choice, scales, open text). Nothing else. It can live on paper, in a PDF, or as an online form, and it can sit inside a survey or do its job entirely on its own.
That standalone part is the bit people miss. So here are some questionnaires with no survey anywhere in sight:
- A patient intake questionnaire at a doctor’s office
- A client onboarding questionnaire an agency sends after signing
- A job application questionnaire
- A gym’s new-member health screening
- A visa or immigration application form
Here’s what all five have in common: each response matters individually. The doctor reads your intake form to treat you, not to compute the average blood pressure of Tuesday’s patients.
The three types of questionnaires
Three flavors, and here’s why you’d care: each one maps to how the answers get used.
- Structured questionnaires use fixed questions with predefined answer options (multiple choice, scales, yes/no). Everyone sees the same thing, so you can actually compare answers across people. This is the right default for anything you plan to count.
- Unstructured questionnaires use open-ended questions and let people answer freely. They’re really interview scripts in written form: rich answers, brutal to compare across people. Great for exploration, when you don’t yet know what the categories should even be.
- Semi-structured questionnaires mix both: fixed questions for the countable facts, open questions where the interesting variation lives. Most real-world questionnaires land here, usually around 80% structured to 20% open.
What is a survey?
A survey is the full research process of collecting information from a sample of people to understand something about the whole group. The questionnaire is one piece of it. The rest of the survey looks like this:
- Sampling: deciding who gets asked, and how many people you need
- Distribution: email, link, embedded form, in person
- Collection: gathering responses over a defined window
- Analysis: turning hundreds of individual answers into a finding
Survey examples: a customer satisfaction survey after support tickets close, an employee engagement survey each quarter, a market research survey before a product launch. In every case, nobody cares much about respondent #47’s individual answers. The value lives in the aggregate, in statements like “62% of customers found onboarding difficult.”
And that’s the cleanest practical test I know: if individual answers matter, you’re running a questionnaire. If only the aggregate matters, you’re running a survey.
The main types of surveys
Surveys come in a few shapes too, and picking the right one is half of the “design the survey, not just the questionnaire” idea.
- Cross-sectional surveys take one snapshot of a group at one point in time. “How do our customers feel about pricing right now?” Most business surveys are cross-sectional, and that’s totally fine.
- Longitudinal surveys ask the same questions repeatedly over time to measure change: the quarterly engagement survey, the NPS you send every six months. The trick is simple: freeze the wording between waves and the trend line stays honest. That’s a free win! (Reword a question mid-stream and you can’t tell whether the score moved or the wording did.)
- Descriptive vs. analytical: descriptive surveys measure what is (“40% of users are on mobile”); analytical surveys probe why, testing relationships between answers (“do mobile users churn more?”). The analytical kind takes more planning, because the comparisons you’ll want to run decide which questions have to exist.
Survey vs questionnaire, side by side
If you’re a table person, here’s the whole article in seven rows.
| Questionnaire | Survey | |
|---|---|---|
| What it is | The question set itself | The process: sampling, distribution, collection, analysis |
| Scope | A document or form | A project |
| Unit of interest | Each individual response | The aggregate pattern |
| Contains the other? | Can exist alone | Always contains a questionnaire |
| Example | Patient intake form | Quarterly engagement survey |
| Output | A filled-in record per person | A finding about the group |
| Success looks like | Complete, accurate individual records | Enough representative responses to support a decision |
Why the difference actually matters
My stance: this isn’t just vocabulary, because the most common way survey projects fail is that someone designs a questionnaire and forgets to design the survey. The questions get written, the form gets built, the link goes out, and only when responses arrive does anyone ask: how many did we need? Was the sample representative? Who’s analyzing the open text? What decision was this supposed to inform?
I love the sample-size math here, because it settles all of that on day one. Say you want results accurate to a ±5% margin of error at 95% confidence for a large population. The standard sample-size math says you need roughly 385 completed responses. If your email list responds at a typical 10%, the questionnaire needs to reach about 3,850 people before the first question is ever answered. Someone who planned the survey knows that number from the start. Someone who only wrote a questionnaire finds out at the end, with 60 responses and a margin of error near ±13%, which is a polite way of saying “we learned very little.”
Run it the other direction too, because it’s genuinely useful for expectation-setting: if your total reachable audience is 500 people and 15% respond, you’ll have 75 responses and roughly a ±11% margin. That might still be fine! A result like “82% want feature A” survives an 11-point margin comfortably. A result like “52% prefer option 1” does not. Knowing which kind of question you’re asking, before you ask it, is the whole game.
So when you’re the person building it: write the questionnaire second. Decide the decision, the audience, the sample, and the analysis first, and the questions almost write themselves.
Advantages and disadvantages of each
Being fair to both sides means admitting the weak spots, so here’s the honest ledger.
Questionnaires (as standalone tools):
- Strengths: cheap to run at any scale, identical wording for every respondent, answers arrive pre-structured, and people can respond privately on their own schedule, which helps honesty on sensitive topics.
- Weaknesses: no interviewer means nobody’s there to clarify a confusing question, literacy and language mismatches can skew answers, and a written form can’t chase an interesting answer the way a human can. The sharp one: a badly worded question fails silently, thousands of times. But the way out is cheap; pilot the form on five people and the confusing question shows itself before it costs you anything.
Surveys (as a method):
- Strengths: the only affordable way to learn about a large group without asking everyone, and the aggregate view surfaces patterns no individual conversation reveals.
- Weaknesses: everything depends on the sample. A survey answered only by your happiest customers measures your happiest customers, however big the response count looks. Response bias, survey fatigue, and leading questions all bend the aggregate, and the final chart looks equally confident either way. So build the defense in up front: a representative sample and neutral wording buy you more truth than any pile of extra responses.
How to design a questionnaire that serves its survey
When the survey is the goal, the questionnaire is a part you get to engineer, and the order of operations does most of the work. Here’s the order I use:
- Write the decision first. One sentence: “This survey exists so we can decide X.” Every question has to serve it.
- Do the sample math. Target responses, expected response rate, required reach. If the reach isn’t realistic, shrink the ambition now, not after.
- Draft questions against the analysis. For each question, name the chart or comparison it’ll feed. No chart, no question.
- Order for momentum. Easy and engaging questions first, the core measurements in the middle, demographics last and optional. People decide whether to finish in the first three questions.
- Keep it short. Closed questions take 10 to 15 seconds each, so a 10-question survey is a 2-to-3-minute ask, which is the zone where completion rates stay healthy. Every question past that costs you finished responses.
- Pilot it on five people. Not to collect data; to find the question everyone misreads. There’s always one, and catching it before launch is the cheapest quality win in the whole process.
- Freeze the wording. Especially for anything longitudinal. Future-you will thank present-you for the unbroken trend line.
Where does a quiz fit?
A quiz uses the same raw material (a series of questions) for a completely different purpose, and it’s my favorite of the three. Surveys and questionnaires collect information; a quiz evaluates or entertains, and the answers are scored against a key. Three differences do all the work:
- Right answers exist. A survey has no wrong answers; a quiz is built on them (or on mapped outcomes, for personality-style quizzes).
- Feedback flows backward. In a survey, information flows from respondent to researcher. In a quiz, the respondent gets something back: a score, a grade, a result, “you’re 8/10 on Boolean logic.”
- The respondent is the beneficiary. People take quizzes for themselves, which is exactly why quizzes get completion and sharing numbers surveys can only envy.
The practical overlap: the three formats are cousins, and modern tools treat them that way. In Uplup, a survey, a standalone questionnaire, and a scored quiz are the same builder with scoring switched off or on, which is handy the day your marketing team wants to turn the “customer needs questionnaire” into a “what’s your automation readiness score?” quiz. Same questions, different physics: one collects data, the other creates an experience people finish.
When to use which
Most real cases sort themselves in about ten seconds once you ask what the output is.
- Use a questionnaire when you need structured information about each individual: intake, onboarding, applications, screening, registration. Success is a complete record per person.
- Run a survey when you need to understand a group: satisfaction, engagement, market research, feedback trends. Plan sample size and analysis before writing questions, and treat the questionnaire as one engineered part.
- Build a quiz when you want engagement, assessment, or lead generation: the respondent gets a result, and you get attention (and, with a lead capture step, contacts).
And a few quick scenarios, because these sort fast too:
- New client kickoff at an agency: questionnaire.
- “Should we change our pricing?”: survey, with the sample math done first.
- Quarterly team morale tracking: longitudinal survey, wording frozen.
- Conference booth engagement: quiz, results on screen.
- Doctor’s office intake: questionnaire, forever.
And when a case seems to straddle two labels, follow the output. “We want feedback from every client after each project” sounds like a survey. But if the account manager reads each response and acts on it individually, it’s a questionnaire with a recurring schedule, and it should be designed like one: complete answers per client matter more than a response rate. The label isn’t bureaucracy; it tells you which numbers to care about.
Frequently asked questions
The same few questions come up every time this topic does, so here are the direct answers.
Is a questionnaire the same as a survey?
No. The questionnaire is the set of questions; the survey is the complete process of sampling, distributing, collecting, and analyzing. A survey always includes a questionnaire, but a questionnaire can be used alone, like a patient intake form.
Can a questionnaire exist without a survey?
Yes. Any form where individual answers are used individually (intake forms, applications, onboarding forms, screenings) is a questionnaire operating outside a survey. Nobody aggregates job applications into a percentage.
What is an example of a survey vs a questionnaire?
A quarterly employee engagement survey is a survey: hundreds of responses analyzed as a group. The health history form you fill in at a new dentist is a questionnaire: one person’s answers, read as that person’s record.
What are the types of questionnaires?
Three, by structure: structured (fixed questions, predefined answers, directly comparable), unstructured (open-ended, exploratory, hard to compare), and semi-structured (a mix, which is where most practical questionnaires land).
Is a quiz a type of survey?
No. Both present a series of questions, but a survey collects information with no right answers, while a quiz scores responses against a key and gives the respondent a result. The direction of value differs: surveys serve the asker, quizzes serve the taker.
Do the words really matter in everyday use?
In casual conversation, no; whether someone frames it as questionnaire vs survey or the other way around, people say “survey” for both and everyone understands. In research and planning, yes: the words mark the difference between writing questions and designing the full process (sample, distribution, analysis) that makes the answers mean something.
How many responses does a survey need?
For a large population, roughly 385 completed responses gives a ±5% margin of error at 95% confidence; about 1,070 tightens it to ±3%. Small, defined populations (like one company’s staff) need proportionally fewer, and any survey’s real minimum is “enough to make the decision it exists for.”
What are the advantages of a questionnaire?
Scale, consistency, and privacy: identical wording for every respondent at near-zero marginal cost, with answers arriving pre-structured for analysis, and respondents answering privately on their own schedule. The trade-off is that nobody’s present to clarify a confusing question or probe an interesting answer.
Which comes first, the survey design or the questionnaire?
The survey design. Decide the decision it serves, the audience, the sample size, and the analysis plan first; the questionnaire is then engineered to feed that plan. Writing questions first and retrofitting a survey around them is the single most common cause of unusable results.
Is an interview a survey or a questionnaire?
Neither, strictly: an interview is its own method where a person asks the questions live and can probe answers. Interviews often use a questionnaire as their script (that’s the semi-structured interview), and a set of interviews analyzed together can serve the same role as a survey, with richer answers and far smaller numbers.
Do this next
Take the form you’re about to send and label it honestly: questionnaire, survey, or quiz. Then give the label its five minutes. For a survey, write the sample-size line before anything else (“we need N responses, so it has to reach M people”). For a questionnaire, decide what happens to each individual response. For a quiz, decide what the taker gets back.
If you’re building any of the three, the same free builder handles all of them, scoring on or off. And those five minutes are the difference between a pile of answers and a decision you can actually make.
