Subjects and lessons to explore.

Law

UX/UI Design · Foundations

Surveys

9 min read 5 practice questions
10 key termsSourcesReport an error
Want it in plain words first? Jump to Eli explains — the same idea, no jargon.
On this page 9 sections
  1. In 30 seconds
  2. Why this matters
  3. The college version
  4. Eli explains
  5. Worked example
  6. Key takeaway
  7. Quick check
  8. Study tools
  9. Sources & references

In 30 seconds

A is a structured set of questions given to many people at once. It buys breadth: dozens or hundreds of answers, all to the same questions. Good survey questions are clear, neutral, and ask one idea at a time. Common types include open-ended questions, multiple choice, and rating scales. Watch for leading questions, double-barreled questions, , and low response rates. Remember that surveys measure what people say about themselves, not what they actually do.

Why this matters

Surveys are one of the fastest ways to hear from many people, which is why teams reach for them early in user research. A well-run survey can reveal patterns across a whole audience: which features people actually use, how satisfied they are, what frustrates them. But the same convenience hides traps. A leading or produces confident-looking numbers that are simply wrong, and a low means the people who answered may not represent the people you care about. Learning to write good questions and read results critically helps you build on solid feedback, and it makes you a sharper reader of every poll and feedback form you meet.

The college version

A survey: many people, one set of questions

In user research, a survey is a structured set of questions given to many people. The Interaction Design Foundation describes UX surveys as tools that gather information about users' feelings, thoughts, and behaviors related to a product or service, typically through a set of questions. 'Structured' is the key word: every sees the same questions, in the same order, with the same answer options. That sameness is what makes the answers comparable. Ask fifty people and you can count how many chose each option; ask five thousand and you can still count. This breadth is the reason surveys exist. The trade-off is depth. A survey answer is short and fixed, so a person's reasons, hesitations, and context stay hidden behind the option they picked. That missing depth is exactly what interviews are built to capture, which is why the two methods are siblings: surveys for breadth, interviews for depth.

Writing questions people can answer honestly

The craft of a survey lives in its wording. Pew Research Center, which runs national surveys, treats question writing as the most important part of the survey process, because questions must accurately measure opinions, experiences, and behaviors. Three rules cover most of the craft. First, be clear and specific. Before: 'Do you find our app's settings menu layout arrangement user-friendly?' After: 'When you last changed a password in the app, how easy was it to find the Settings menu?' Second, be neutral. A neutral question does not imply a particular answer or give away what the team expects, as Nielsen Norman Group puts it. Before: 'How much did you enjoy our redesigned home screen?' After: 'How would you rate the redesigned home screen?' Third, ask one idea at a time. Before: 'How satisfied are you with the price and the delivery time of your order?' After: two separate questions, one about price and one about delivery time. Pew calls a question that asks for two evaluations at once a double-barreled question: respondents find it hard to answer, and the results are hard to interpret.

Three question types to know

Most surveys mix three question types. An lets respondents answer in their own words, for example 'What, if anything, stopped you from completing your purchase today?' These answers are rich but slow to analyze. A multiple-choice question offers a fixed set of options, for example 'How often do you use the search bar in the app?' with answers like 'Never,' 'A few times a month,' 'A few times a week,' and 'Daily.' A asks for a position on a numbered or labeled range, for example 'How likely are you to recommend this service to a friend?' on a zero-to-ten scale. Nielsen Norman Group notes that closed questions used in surveys usually take the form of multiple-choice or rating-scale items, and the Interaction Design Foundation recommends mixing types so that numbers and stories both come out of the same survey.

Four traps that quietly poison results

Four failure modes deserve names. A nudges respondents toward a particular answer; 'How much did you enjoy our award-winning checkout?' is really an opinion wearing a question mark. A double-barreled question packs two ideas into one, so a 'no' answer to 'Is the app reliable and fast?' is ambiguous, because nobody can tell which half failed. Response bias covers the ways answers drift from the truth: respondents who are tired of long surveys, respondents who do not understand the question, and respondents influenced by the order of the questions. Pew's example is striking: in a 2008 poll, when 'the economy' was offered as a listed answer option, 58 percent chose it, while only 35 percent of those answering in their own words volunteered it. Finally, low response rates: Nielsen Norman Group warns that every extra question reduces the response rate and makes all results suspect, so keep surveys short and test them with a few people from the target audience before sending them out.

Who you send it to matters as much as the questions

A survey's results describe only the people who received it and answered it. If a team sends a checkout survey to people who finished buying, they hear from satisfied buyers and miss everyone who left. If they post the survey link on their own social feed, they reach existing fans rather than confused newcomers. Pew's point is the mirror image: even accurate random sampling is wasted if the questions themselves are biased. Choosing who receives a survey, the sample, is its own topic taught in data-science-and-ai-literacy; here the practical rule is simpler. Before reading any results, ask who was asked and whether that group matches the people the team needs to understand.

Surveys measure what people say

Every survey answer is a statement about the person who gave it, not a recording of what they did. Nielsen Norman Group makes the point bluntly: quantitative surveys count how many people say they do this or that. Someone may report exercising daily, being very satisfied, or never using a feature, and all of it may be true, exaggerated, or simply what they wished were true. That is why survey results work best as one layer of evidence, combined with other methods: interviews add depth to the numbers, and usability testing shows what people actually do with a product. This honest framing is not a criticism of surveys; it is what makes them useful. Know what a survey can and cannot tell you, and you will know how much weight to give it.

Eli, the EliExplains learning guide

Eli explains

The same idea, in plain words

Explain it like I’m 10

A survey is like handing the same question card to lots of people at once. Everyone gets the same card, so you can count answers and see patterns: eight out of ten people circled 'easy,' two circled 'hard.' The trick is the card itself. If the card says 'How much do you love our easy checkout?' people will mostly circle 'a lot,' because the question already told them the answer it wanted. If the card asks two questions at once, like 'Was checkout fast and clear?', you cannot tell whether a 'no' means slow or confusing. If only a few people hand the card back, the answers you have may not represent everyone else. And the answers are what people say about themselves: someone can circle 'very healthy' while skipping every workout.

Picture it like this

Think of a survey as a crowd vote at a town meeting. Everyone holds up a colored card for the same question, and you count the colors. It is fast, and it tells you what the crowd says. But the crowd is only the people who showed up, the question must be one you can answer with a single card, and a card raised in the air is a claim, not proof: someone can hold up the green 'I recycle every week' card and still toss cans in the trash.

Where the picture stops working

Real surveys are not votes: people answer privately, answers are often graded scales rather than colors, and the counting requires care. The analogy also hides that question wording, not just the crowd, shapes what people say; a slightly different card can produce a different count.

Worked example

A team wants to know why shoppers abandon their carts. They draft: 'How much did you love our fast, easy checkout?' A colleague flags it: the question hands out the answer. They rework it into a neutral, single-idea question, 'How easy was it to complete checkout on your last order?' with a four-point scale from 'Very easy' to 'Very difficult,' plus one open-ended follow-up: 'What, if anything, made checkout difficult?' They keep the survey to six questions, pilot it with five colleagues and three recent customers, fix a confusing option, and send it to the 900 people who abandoned a cart in the past month. When reading the results, they remember the answers describe what those shoppers said, not what they did.

Key takeaway

A survey collects many self-reported answers to the same questions. Write questions that are clear, neutral, and single-idea, keep the survey short, choose the audience deliberately, and treat results as one layer of evidence.

Quick check

3 questions here, of 5 in this lesson’s practice set. Answers stay hidden until you check.

Question 1 of 3foundational

In user research, what is the working definition of a survey?

Choose an answer, then check it.
Question 2 of 3intermediate

A team drafts the question 'How satisfied are you with the price and the delivery time of your order?' What is the main problem with this question?

Choose an answer, then check it.
Question 3 of 3advanced

A survey asks, 'How much did you enjoy our newly redesigned, award-winning home screen?' The most serious flaw in this question is that it:

Choose an answer, then check it.
Practice all 5

Keep learning

Ready to build on this? Continue to the next lesson.

Practice this lesson
Study tools & related lessonsYou’ll learn to · Common mistakes · Easily confused · Key vocabulary · Related

You’ll learn to

  • Define a survey as a structured set of questions given to many people and contrast it with interviews in terms of breadth versus depth.
  • Describe the qualities of a good survey question: clear, neutral, and one idea at a time.
  • Distinguish open-ended, multiple-choice, and rating-scale questions and give an example of each.
  • Recognize leading questions, double-barreled questions, response bias, and low response rates in a survey.
  • Explain why the people who receive a survey matter as much as the questions it asks.
  • Interpret survey results as self-reported statements that should be combined with other research methods.

Common mistakes

  • Writing a leading question that plants the expected answer, such as 'How helpful is our friendly support team?' instead of 'How would you rate the support you received?'

  • Folding two ideas into one double-barreled question and then not knowing which half a 'no' refers to.

  • Making the survey so long that people quit partway, which drags the response rate down and makes the answers suspect.

  • Sending the survey to whoever is easy to reach and then reporting the results as if they describe the whole audience.

  • Reading self-reported answers as observed behavior, treating 'I check the app daily' as a fact instead of a claim to combine with other evidence.

Easily confused

Survey vs. Interview

A survey asks the same fixed questions of many people for breadth; an interview explores a few people's experiences in depth.

Open-ended question vs. Closed question

Open-ended questions let respondents answer in their own words; closed questions restrict them to a fixed set of options.

Leading question vs. Neutral question

A leading question suggests the expected answer; a neutral question does not imply any particular answer.

Self-report vs. Observed behavior

Surveys capture what people say about themselves; observation-based methods capture what people actually do.

Key vocabulary

survey
A structured set of questions given to many people so their answers can be compared.
respondent
A person who receives a survey and answers its questions.
open-ended question
A question that lets the respondent answer freely in their own words.
closed question
A question that restricts the respondent to a fixed set of answer options.
rating scale
A closed question format that asks respondents to place their answer on a numbered or labeled range.
leading question
A question worded so that it nudges respondents toward a particular answer.
double-barreled question
A question that asks about two different ideas at once, making the answers hard to interpret.
response bias
Any tendency in a survey that makes answers drift from the truth, such as fatigue or question order.
response rate
The share of the people who received a survey who actually answered it.
self-report
Information people provide about themselves, which may differ from what they actually do.

Sources & references

  1. 28 Tips for Creating Great Qualitative Surveys — Nielsen Norman Group
  2. Open-Ended vs. Closed Questions in User Research — Nielsen Norman Group
  3. Surveys (UX Surveys) — What are UX Surveys? — Interaction Design Foundation
  4. Writing Survey Questions — Pew Research Center

EliExplains lessons are original prose written from the open, credible references above. See Copyright & Licensing.

Researched 2026-08-21

Educational content only. It is not medical, legal or professional advice. Found an error? Tell us.