Skip to content
UX Atlas
Plate RES·004 · Research methodGrid H2 · Rev. 2026-08-14

Surveys

Good for measuring how widespread something is, and poor for discovering what that something is.

Research methodEstablished researchIntermediate

Definition

A survey collects structured responses from a sample. It quantifies attitudes and self-reported behaviour, and it can only measure things you already know to ask about.

Established research. Originates in published research and has been replicated. The effect is real. Its size in your product still needs measuring.

On this plate
  1. 1Definition
  2. 2Design rules
  3. 3Writing one that survives contact
  4. 4Question wording
  5. 5In practice
  6. 6Commonly confused with
  7. 7Local chart
  8. 8Sources
  9. 9Continue from here

Reach for this when

  • Quantifying something you already understand qualitatively.
  • Tracking a measure over time.
  • Segmenting a known population.
  • You need a number you can track over time rather than an explanation.

Common mistakes

  • Surveys used for discovery.
  • Biased sampling.
  • Leading or compound questions.
  • Reporting percentages without the sample size.
  • Treating an open-text field as qualitative research. It collects opinion without the chance to ask why.

Design rules

  • One idea per question. Double-barrelled questions produce uninterpretable answers.
  • Avoid leading wording and loaded adjectives.
  • Offer a neutral midpoint and a 'not applicable' option where they are legitimate answers.
  • Randomise option order where order effects are plausible.
  • Keep it short. Completion falls sharply past a few minutes, and the dropouts are not random.
  • Pilot with five people before sending. It always finds an ambiguous question.

Writing one that survives contact

  1. 1

    Decide the decision first

    Write down what you will do differently at each plausible result. Any question that fails this test is cut.

  2. 2

    Ask about behaviour with a time frame

    How many times in the last seven days beats how often do you usually, which measures self-image.

  3. 3

    Pilot on five people out loud

    Have them narrate their interpretation as they answer. Ambiguity you cannot see is the main cause of unusable data.

  4. 4

    Keep it short enough to finish

    Completion falls sharply past a few minutes, and the people who drop out are rarely a random subset.

  5. 5

    Report the denominator

    Sent, started, completed. A result without its response rate cannot be weighed.

Question wording

Do

  • Offer a genuine opt out, such as does not apply, so people are not forced into a false answer.
  • Balance scales, with as many positive points as negative and a neutral midpoint.
  • Put the open text box at the end, where it costs nothing to leave blank.

Do not

  • Ask two things in one question, as in fast and easy to use.
  • Lead with an adjective you hope for, as in how helpful was this.
  • Make every question required, which converts abandonment into invented answers.

In practice

  1. i.

    Interview first, then survey

    Mixed methods

    Interviews tell you what to ask about. Surveys tell you how common it is. Running the survey first usually measures the team's assumptions rather than the users' reality.

  2. ii.

    The intercept that only caught the happy

    Sampling

    An in-product survey shown after checkout measured satisfaction at ninety percent. Everyone who failed to check out never saw it, which is the population the team was actually asking about.

Commonly confused with

Close neighbours that get substituted for each other. The difference is stated first, then the condition that picks one over the other.

vs.

User Interviews

Research method

An open conversation that can follow whatever turns out to matter.

Key difference

A survey measures how widespread something is among people you can reach. An interview finds out what that something is. A survey can only ask about what you already thought to ask, which is why it is the wrong first move.

Use Surveys when

  • You already know what to ask and need to know how common it is.
  • The sample has to be large enough to generalise from.
  • The output is a number that will be tracked over time.

Use User Interviews when

  • You do not yet know the right questions.
  • The answer is a reason, a story or a workaround.
  • A handful of people, explored properly, will settle it.

Learn about User Interviews

Local chart

RES·004 · Region IV · Grid H2

This entry and the entries it is stated to connect with. Each line is labelled with what the connection means.

Confirmation BiasBIAS·009 · PRINCIPLES BEHIND THISSurvivorship BiasBIAS·010 · PRINCIPLES BEHIND THISResearch PlanTPL·001 · OFTEN USED WITHUser InterviewsRES·003 · ALTERNATIVE APPROACHQuantitative ResearchTERM·020 · SHORT DEFINITIONSample SizeTERM·024 · SHORT DEFINITIONA/B TestingRES·009 · RELATED CONCEPTProduct AnalyticsRES·010 · RELATED CONCEPTResearch SynthesisRES·014 · RELATED CONCEPT

Sources

Where a source establishes something narrower than the popular reading of it, the note says so.

  1. [1]

    Surveys That Work

    C. Jarrett, Rosenfeld Media, 2021

    A process for surveys built around the decision they inform, including why most rating scales collect noise.

Continue from here

Each link says what the connection is, so you can tell a principle from an alternative from a thing people mix this up with.

Principles behind this

The reasoning this solution is an application of.

Often used with

These usually appear in the same screen or the same decision.

Short definition

The one-paragraph version, for when that is all you need.

Related concept

Connected closely enough to change how you apply this.