A questionnaire looks simple. A list of questions, a link, a few hundred responses — and then the trouble starts. Answers that contradict each other, half-finished surveys, and results you cannot quite trust. In most cases, the problem is not the sample size or the statistics. It is the questionnaire itself.
Good questionnaire design is not about writing clever questions. It is about designing a conversation that a stranger can follow easily, answer honestly, and finish without getting tired or confused. This guide walks you through 12 practical rules that will help you design a research questionnaire that collects clean, usable data — whether you are working on a class assignment, a thesis, or a larger study.
Start With Your Research Question, Not Your Questions
Before you write a single survey item, write down your research question and the variables you need to measure. Every question in your questionnaire should earn its place by helping you answer that question. If a question is merely interesting, leave it out.
A useful test is to make a simple table with three columns: your research question, the information you need, and the questionnaire item that will provide it. If an item does not map to anything in the first two columns, it is a candidate for deletion. Long questionnaires do not just annoy respondents; they lower completion rates and increase careless answering toward the end.
Rule 1: One Idea Per Question
Each question should ask about one thing only. Double-barreled questions — questions that combine two ideas — are one of the most common questionnaire mistakes, and they make answers impossible to interpret.
For example, “How satisfied are you with your supervisor’s guidance and feedback?” is two questions in one. A respondent might be happy with the guidance but unhappy with the feedback, and no single answer can capture that. Split it into two separate questions, one about guidance and one about feedback.
Rule 2: Use Plain, Concrete Language
Write the way you would explain the question to a friend who is not in your field. Avoid jargon, abbreviations, and academic phrasing unless your respondents definitely share that vocabulary. Words like “frequently,” “often,” and “rarely” also mean different things to different people, so give concrete time frames instead: “in the past week” is clearer than “often.”
Read each question aloud. If you stumble over it, your respondents will too. Aim for short sentences and familiar words, especially if your sample includes people with different educational backgrounds.
Rule 3: Keep Response Options Balanced and Complete
For closed questions, the answer options must cover all reasonable possibilities and must not overlap. If you ask about age with the ranges 18–25, 25–35, and 35–45, a 25-year-old or 35-year-old could fit two categories. Use non-overlapping ranges such as 18–24, 25–34, and 35–44 instead.
Also include an “Other (please specify)” option where you cannot list every possibility, and a “Prefer not to say” option for sensitive questions. Forcing an answer that does not fit produces false data, which is worse than a missing answer you can see and handle.
Rule 4: Match the Scale to the Question
If you use a Likert scale, keep it consistent throughout the questionnaire. Switching between a 5-point scale in one section and a 7-point scale in the next confuses respondents and complicates analysis. Label every point on the scale, not just the endpoints, so respondents know exactly what “3” means.
Decide deliberately whether to include a neutral midpoint. A midpoint is appropriate when a genuinely neutral position exists, such as “neither agree nor disagree.” If you force a choice by removing it, you may push undecided respondents into a false opinion.
Rule 5: Avoid Leading and Loaded Questions
A leading question nudges the respondent toward a particular answer. “Don’t you agree that online learning is more convenient?” is not a question; it is an opinion wearing a question mark. Ask neutrally instead: “How convenient do you find online learning?”
Loaded questions are similar but sneak in an assumption, such as “How often do you skip lectures?” when the respondent may never skip lectures. Use a filter question first — “Do you ever skip lectures?” — and only then ask follow-up questions of those who answer yes.
Rule 6: Order Questions Thoughtfully
Start with easy, non-threatening questions. This builds confidence and momentum. Put sensitive questions — about income, health, or personal behaviour — later in the questionnaire, once the respondent is already engaged. Group related questions together under clear section headings so the survey feels organised rather than random.
Be careful about order effects: an earlier question can influence how people answer a later one. If you ask about a serious problem first, respondents may rate everything that follows more negatively. When in doubt, keep the flow logical: general to specific, factual to opinion-based.
Rule 7: Use Filters and Skip Logic Carefully
Not every question applies to every respondent. A question about thesis supervision only makes sense for students who have a supervisor. Clear instructions such as “If you answered No, please skip to Question 12” prevent irrelevant answers. In an online survey tool, use skip logic to hide irrelevant questions automatically.
Test the skip logic thoroughly. One broken skip can send respondents in circles or hide questions they should have answered, and you may only discover the damage after data collection has closed.
Rule 8: Keep It As Short As You Honestly Can
Respondent fatigue is real. After about 15 to 20 minutes, attention drops, answers get shorter, and people start choosing the same response option just to finish. Before finalising your questionnaire, ask of every section: could this be shorter without losing essential information?
Shorter does not mean shallow. A focused 20-item questionnaire that measures three variables well is far more useful than a 60-item questionnaire that measures ten variables badly. If you need more depth, consider interviews or a follow-up study rather than stretching one survey to breaking point.
Rule 9: Design for the Device Your Respondents Will Use
Many respondents will complete your survey on a phone. Long matrix questions with many columns are painful on a small screen, and open text boxes that are too small discourage detailed answers. Preview your questionnaire on a phone as well as a computer, and check that buttons are easy to tap and that no question is cut off.
Also keep the visual layout clean: one question per screen or a small group of related questions, plenty of white space, and a progress indicator so respondents know how much is left. A survey that looks effortful will be abandoned; a survey that looks manageable will be finished.
Rule 10: Pilot Test Before You Launch
Never send a questionnaire to your full sample without testing it first. Ask five to ten people similar to your target respondents to complete it while thinking aloud, and note every question they hesitate over, misunderstand, or find awkward. Then ask them what they thought each question was asking — you may be surprised.
Pilot testing also reveals practical problems: how long the survey really takes, whether any skip logic fails, and whether any response option is missing. Fix these issues now. After launch, it is too late.
Rule 11: Protect Anonymity and Be Transparent
Start with a short introduction explaining the purpose of the study, who is conducting it, how long it will take, and how the data will be used. State clearly whether responses are anonymous or confidential, and do not collect identifying information you do not need.
This is not just good ethics; it improves data quality. People answer sensitive questions more honestly when they believe their answers cannot be traced back to them. If your study requires ethical approval, include the approval details and a contact address for questions.
Rule 12: Check Reliability and Validity, Not Just Grammar
A questionnaire can be beautifully written and still measure the wrong thing. Validity asks whether your questions really capture the concept you intend to study; reliability asks whether they do so consistently. If you are adapting an existing scale, keep its wording and scoring intact unless you have a good reason to change them, and report the change.
After data collection, check internal consistency for multi-item scales and look for items that behave strangely — for example, an item almost everyone answers the same way, or one that correlates poorly with the rest of its scale. These checks belong in your methods chapter, not just in your head.
Worked Example: From Research Question to Questionnaire
Suppose your research question is: “How does social media use relate to study habits among undergraduate students?” Your variables are social media use and study habits, plus basic background information such as year of study.
You might start with background items (year of study, field of study), then a section on social media use (platforms used, average hours per day, purpose of use), then a section on study habits (study hours, planning, distraction). Notice what is missing: no questions about grades, sleep, or mental health, even though all three are interesting. Unless they are part of your research question, they only lengthen the survey and dilute your focus.
For the social media section, a good item is: “On a typical weekday, about how many hours do you spend on social media?” with clear ranges (less than 1 hour, 1–2 hours, 3–4 hours, more than 4 hours). A weaker version would be: “Do you use social media a lot and does it affect your studies?” — that is double-barreled, vague, and leading all at once.
Common Mistakes to Avoid
First, asking everything you can think of instead of everything you need. Second, using vague frequency words instead of concrete time frames. Third, writing questions that assume the answer, such as asking how much someone enjoys something before checking whether they do it at all. Fourth, skipping the pilot test because of time pressure — the pilot almost always saves more time than it costs. Fifth, changing a validated scale’s wording casually and then comparing your results with studies that used the original version.
If you catch yourself defending a question with “but I really want to know,” pause. Wanting to know is not the same as needing to know for this study. Save the extra questions for a future project.
Frequently Asked Questions
How many questions should a research questionnaire have?
There is no magic number. It depends on your research question, but as a guide, most student surveys work well with 15 to 30 focused items and a completion time under 15 minutes. If your questionnaire takes longer, check whether every item is essential.
Should I use open-ended or closed questions?
Use both, but mostly closed questions. Closed questions are easier to answer and analyse, while one or two open-ended questions at the end let respondents add anything important you missed. Too many open questions increase dropout, especially on phones.
What is the difference between a questionnaire and a survey?
A questionnaire is the set of questions itself; a survey is the whole process of collecting data with that questionnaire, including sampling and distribution. People use the words interchangeably in everyday speech, but in a methods chapter the distinction is worth keeping.
Do I need ethical approval for a questionnaire study?
In most universities, yes, if you are collecting data from human participants — even for a simple anonymous questionnaire. Check your institution’s ethics requirements before distributing your survey, and include a consent statement at the start.
How do I increase my response rate?
Keep the survey short, send a clear invitation explaining the purpose, use reminders sparingly, and make the first questions easy. A progress bar and a mobile-friendly layout also help. Incentives can help, but they should be modest and appropriate to your context.
Final Thoughts
A good questionnaire respects the respondent’s time and intelligence. It asks one clear thing at a time, in plain language, in a sensible order, and only when the answer genuinely matters for the research question. Follow the 12 rules above, pilot test with real people, and fix what they stumble over — and your data will be far easier to analyse and far easier to defend.
Start with your research question, draft your items, test them, and revise. That cycle, repeated once or twice, is what separates a questionnaire that merely collects answers from one that produces trustworthy evidence.