01 / UNDERSTAND
What is survey research?
Survey research collects structured responses from a defined group of people. In design work, it can describe reported experiences, compare responses or investigate how widespread an issue may be when the sampling supports that claim.
A survey produces answers to particular questions from the people who responded. Whether those answers represent a wider population depends on recruitment, coverage, response patterns and measurement quality.
Why use Survey Research?
A survey can make the same question available to many people and support systematic comparison. That consistency is useful only when respondents understand the question similarly and have the experience needed to answer. A large response count does not repair an ambiguous question or an unsuitable sample.
When to use it
USE IT WHEN
The team knows what it wants to measure and can define the relevant population.
A qualitative finding needs structured follow up across a broader group.
You need comparable responses using a consistent question and recall period.
When to choose another method
Use User Interviews when you first need to understand the language and possible explanations.
Use Usability Testing to observe task performance rather than ask people to predict it.
Use Diary Study when repeated episodes are difficult to recall accurately in one sitting.
What you’ll need
Write the decision the results should inform. Define who is eligible, how they will be reached and who may be excluded. Prepare an analysis plan before collecting responses. Decide how to handle partial responses, missing answers and personal information. Keep the questionnaire short enough that every question has a clear purpose.
What the method needs
Keep the elements that help answer your question and support a decision.
Who the research concerns and which people are actually invited.
The specific concept, context and period a respondent should consider.
Answers that fit the question without forcing an inaccurate choice.
How people encounter and complete the survey.
Which comparisons and claims the design can support.
02 / PUT IT INTO PRACTICE
How to plan and run a survey
Keep evidence and assumptions distinct. Revisit your work when new information changes your understanding.
01
Specify the decision and population
Translate the broad topic into an answerable research question. Define eligibility before choosing a channel.
A survey of recent applicants differs from a survey of everyone visiting the homepage.
02
Choose the recruitment approach
Document coverage and likely exclusions. Decide whether the sample supports population estimates or exploratory description.
A voluntary email response is not automatically representative.
03
Write focused questions
Ask one thing at a time with a concrete recall period. Avoid wording that praises or presupposes the proposed solution.
Ask about the last application, not whether the helpful new form was easy.
04
Check the answer options
Make categories meaningful and avoid overlap. Provide uncertainty or nonapplicability options where appropriate.
Do not force someone who never contacted support to rate it.
05
Pretest and pilot
Ask people to explain how they interpreted the question and selected an answer. Check routing and completion burden.
Repair misunderstandings before distributing the questionnaire widely.
06
Analyse with the denominator visible
Report who answered each question and distinguish missing responses from substantive answers.
A percentage among respondents is not necessarily a percentage of all customers.
07
Decide what follows
Connect the result to a design action or a narrower investigation. State limitations where the result is presented.
Use an unexpected response pattern to recruit follow up interviews.
Example: understanding application follow up
Fictional survey plan: a team wants to understand whether recent applicants contacted support after submitting a course application. It would invite eligible applicants from a defined period and ask about the most recent submission.
The questionnaire first asks whether contact occurred, then asks the reason using language explored in earlier interviews. It keeps status enquiries separate from requests to correct information. It does not ask only people who reached the support team.
If respondents frequently describe uncertainty about receipt, the team could investigate the confirmation experience. It would still need to examine who did not respond and whether the invitation excluded people with incomplete contact details. No response rates or findings are claimed in this example.
03 / BELOW THE SURFACE
Below the surface: the denominator is a design choice
A precise percentage can answer a different question from the one stakeholders think they asked.
POPULATION
Whose experience are we trying to describe?
Define eligibility independently of the response channel.
COVERAGE
Who cannot appear in this sample?
Inspect exclusions created by the invitation route.
CLAIM
What does the denominator actually mean?
Keep the interpretation within the study’s reach.
A survey shown after successful submission excludes people who could not submit. Its results may describe the successful route while saying little about the barrier the team most needs to improve. Name the population behind the denominator before interpreting the number.
Compare the intended population, invitation coverage and respondents. If the gaps align with the experience under investigation, more responses through the same route may not resolve the problem. A complementary recruitment approach or qualitative study may be needed.
This editorial check turns a reporting footnote into a design decision about whose experience the study can hear.
Survey Research companion worksheet
Use page one to connect the research question, population and questionnaire. Page two records response coverage, missingness and the claim the evidence can support. The sheet is a planning aid, not a sample size calculator.
Practical tips
Keep exact wording and routing with the results.
Report relevant response counts alongside percentages.
Common mistakes
Treating a large convenience sample as representative.
Combining two ideas in one question.
Changing wording between waves and assuming results are directly comparable.
What you’ll get
Structured evidence with explicit coverage and measurement limits. It should support a bounded decision or identify the next question, rather than lend numerical authority to an assumption.