Write down the decision the answers will inform before drafting the first question. Then identify who can answer, use clear and neutral wording, offer response options that fit, and test the form with a few people like the intended respondents.
Building a questionnaire starts with that decision, not with choosing an app. A well-designed set of questions can help a community group choose a meeting time; an attractive form with vague questions may produce numbers that cannot guide the choice.
Creating a survey around one decision
Suppose a library can add one weekly workshop and must choose between a Tuesday evening and a Saturday morning. Its decision is which time to offer. The relevant audience is people the library hopes to serve, including people who have not attended a previous event. Asking only current Tuesday visitors would miss people unable to come at that time.
When constructing the questionnaire, a useful first question might ask, “Which of these times could you attend during the next two months? Select all that apply.” Offer Tuesday evening, Saturday morning, neither, and a way to suggest another time. Only for people who select both times, ask which of the two they prefer. The answers distinguish availability from preference: the most popular time is of little use to someone who cannot attend it.
When deciding how to make a survey, also specify what the organizer will do with the responses. For this example, count availability for each time and use preferences to help decide if the totals are close. Consider the result alongside staffing, room availability, and costs.
Turn vague or leading questions into clear ones
When writing a survey, remove wording that suggests the answer it expects. “How much would you enjoy our wonderful new workshop?” assumes both that the workshop is wonderful and that the respondent would enjoy it. A more neutral version asks, “How interested would you be in attending a workshop on this topic?” with a balanced scale from very interested to not at all interested, plus “not sure.”
A vague question leaves important terms to the reader. “Do you attend often?” could mean once a week to one person and twice a year to another. Ask instead, “How many library workshops have you attended in the past six months?” A defined period gives respondents a common reference and helps the organizer interpret the answers.
Avoid asking two things at once. “Are the time and location convenient?” cannot tell the organizer which part caused a negative answer. Separate the questions, or ask about each feature in its own row. As Pew Research Center explains, question wording and context can change answers; clarity and neutrality are part of the measurement, not just a matter of style.
Choose answers people can actually give
When designing a survey, make closed-question answers cover plausible cases so they can be counted meaningfully. For attendance in the previous six months, the options could be 0, 1, 2–3, and 4 or more. These categories do not overlap: a person who attended three times has one clear choice. Include “not sure” when recall may be difficult, and make an answer optional if a person may reasonably prefer not to disclose it.
For a single-choice question, state “select one.” For availability, “select all that apply” allows a person to choose both times. If neither works, provide “neither” and configure the form so that it cannot be selected together with a time. “Other,” with an optional text box, can capture an answer the designer did not anticipate. Check that the tool records those entries in a form that can later be reviewed.
An open question such as “What would make the workshop easier to attend?” can reveal an obstacle missing from the list. It also takes longer to answer and summarize. Use it when the possible answers are not yet known, and leave enough time to read the responses. Do not request a long explanation for every simple choice.
Decide whom to invite and what to collect
Creating a survey includes planning distribution. Survey creation also requires deciding how responses will be reviewed. The library might use a mailing list, a sign at the desk, and a notice in a community center. Each reaches different people, and none necessarily reaches every potential user. State who was invited and how responses were collected when reporting results. An open link answered by volunteers does not automatically represent all residents, even if many people respond.
Give a clear deadline and estimate the time required. Explain the purpose, whether responses are anonymous or merely confidential, who can see them, and how results will be used. Collect names or contact details only if needed, and explain why. If a follow-up contact is optional, keep it separate from answers where the tool and process allow that separation.
Making a survey also requires deciding how to handle duplicate responses and incomplete forms before looking at the results. For example, count an incomplete form for an availability question only if that question was answered. Record the number invited where known, the number responding, and how many answered each question. A percentage based on five answers deserves a different interpretation from one based on hundreds.
Test the complete form
Before sending the link, test the survey you have built with a few people similar to the intended audience. Watch where they hesitate and ask them what a question means in their own words. Check whether each person finds an appropriate answer and whether any important term has two interpretations. The U.S. Census Bureau describes pretesting as a way to identify problems with wording, order, skip instructions, and formatting.
Test the actual form on a phone and a computer. Follow every branch, submit a test response, and inspect the resulting data. If someone selects “neither,” does the form still require a preferred time? Does the final screen confirm submission? Remove test responses before collecting real ones. Revise the form and repeat the affected checks when wording or routing changes.
A tool for building surveys can display choices and count responses; it cannot decide whether the choices answer the decision. A team learning how to make a survey should review the draft against its original aim before sending it. If an answer would not change the decision or provide needed context, consider removing the question.
A checklist to keep
- Decision: name the choice and how the answers will be used.
- People: invite those whose experience matters and record who was reached.
- Questions: use neutral words, one idea at a time, with a defined period.
- Answers: cover realistic cases, avoid overlap, and specify one choice or several.
- Test: check comprehension, routing, devices, and the saved results before launch.
- Interpretation: report the number of responses and the limits of the group reached.
