You build a thesis questionnaire in four steps: derive what exactly you need to measure from your thesis aim and hypotheses, assign one unambiguous question to each variable, order the questions from the easiest ones to the demographic block, and test the finished questionnaire on a small group of people from your target population. Choose a questionnaire when you need comparable answers from a larger number of respondents. The full wording then belongs in the appendix of your thesis and its description in the methodology chapter.
When a questionnaire is the right choice
A questionnaire is a tool of quantitative data collection. Every respondent gets the same questions in the same order, so the answers can be counted, compared, and processed statistically. That makes it suitable for finding out how widespread a phenomenon is, what the differences between groups are, or whether two variables are related.
It is not suitable for everything, though. If you need to understand why people decide the way they do, and you want to ask follow-up questions during the data collection, an interview will give you more.
You describe the choice of method and its justification in the methodology chapter of your thesis. "It was the easiest option" will not satisfy your examiner; an aim that requires comparable numbers will.
How to get specific questions from your thesis aim
The most common cause of a bad questionnaire is that it was written before it was clear what it should measure. Questions get collected intuitively, and during the analysis you discover that half of your data has nowhere to belong.
The working approach is the reverse. Write down the aim of your thesis, under it your hypotheses or research questions, list the variables you need to measure for each one, and only then assign questions. A table with three columns helps: hypothesis, variable, question number. A question that has no row in that table does not belong in the questionnaire. If you are still formulating your hypotheses, start with the article on hypotheses and research questions.
Question types you can use
Survey methodology distinguishes more question formats than most students will ever need: open-ended, closed-ended, semi-open, dichotomous, or projective. Pew Research Center's guide to writing survey questions shows that even the choice between an open-ended and a closed-ended format changes the answers people give, so pick each format deliberately. For a bachelor's or master's thesis you will usually get by with four types.
| Question type | What it looks like | What it is good for | How you analyze it |
|---|---|---|---|
| Closed-ended | a fixed list of options to choose from | facts, frequencies, clear-cut choices | easily, by counting frequencies |
| Semi-open | options plus an "other, please specify" field | when you are not sure your list covers everything | like closed-ended, the extras get categorized |
| Open-ended | an empty field for a free answer | reasons, suggestions, things you cannot predict | laboriously, you have to code answers into categories |
| Scaled | an agreement or rating scale | attitudes, satisfaction, degree of agreement | averages, medians, group comparisons |
Scaled questions are most often built on a Likert scale, meaning a statement plus a scale running from full agreement to full disagreement. It is the most common way to measure an attitude with a number, and it has its own rules for the number of points.
Use open-ended questions sparingly. Each one means dozens of text answers that you have to sort into categories, otherwise they cannot be analyzed at all.
How to word questions so they produce usable data
Wording decides whether you measure what you think you are measuring. Check every question against a short list: is it unambiguous, understandable to every respondent, easy on their memory, and free of suggestion and abstraction? The Pew Research guide documents how even small changes in phrasing can shift the results noticeably.
Turn that into five rules:
- One question, one thing. "Are you satisfied with your salary and your colleagues?" cannot be analyzed, because a respondent may be satisfied with only one of the two. Split it into two questions.
- No leading. "Does the constant interruption in your open space bother you too?" tells the respondent what to answer. A neutral version asks how often interruptions occur.
- Options that do not overlap and leave nothing out. Age brackets of 20 to 30 and 30 to 40 years are a classic mistake; a thirty-year-old does not know where they belong. Add "I don't know" if such an answer can exist.
- The respondent's language. A technical term from your theoretical chapters may mean nothing to people in the field. If you have to use it, explain it in one sentence.
- Nothing the respondent cannot remember. "How many times have you changed your communication tool in the past three years?" produces guesses, not data. The same goes for double negatives.
How many questions to include and how long it should take
There is no prescribed number; it follows from the number of variables you need to test your hypotheses. As a rule, a questionnaire for a bachelor's or master's thesis takes a few minutes to complete.
The longer it gets, the higher the risk that a respondent quits halfway through or starts clicking mechanically down one column, which damages your data more than a missing question would. Measure the real completion time during pilot testing and state it in your introductory text. For every question, also ask yourself what table or chart you will make from it. If you do not know, cut the question.
How to order the questionnaire from the intro to demographics
Question order is not cosmetic. It affects how many people finish the questionnaire, and sometimes the answers themselves.
- Introductory text. Who is collecting the data, at which university, for what purpose, how long it takes, and what will happen to the answers. Four sentences are enough.
- Screening question. If your research concerns only a specific group, verify right away that the respondent belongs to it.
- Easy factual questions. Start with what people can answer without hesitation.
- The core, in thematic blocks. Keep questions on the same hypothesis together, and group scale batteries under one shared scale.
- Sensitive and open-ended questions. They belong near the end, once the respondent has already invested time in the questionnaire.
- Demographics last. At the beginning it feels like an interrogation; at the end people fill it in without a problem.
Put only the attributes you will actually use to compare groups into the demographic block. Every extra one can give a respondent away in a small organization.
Sample questions for a model topic
Model topic: the impact of working from home on employees' job satisfaction. The hypothesis assumes that employees who work from home at least two days a week rate their satisfaction higher than those who work only from the office.
- Screening. Do you work from home at least one day a week? (yes / no)
- Closed-ended, single answer. How many days a week do you work from home? (1 / 2 / 3 / 4 / 5)
- Closed-ended, multiple answers. Which of these tools do you regularly use when working from home? (a list plus "other")
- Scaled. To what extent do you agree with the statement "I get more tasks done when working from home than at the office"? (strongly agree to strongly disagree)
- Scale battery. On the same scale, rate statements about communication with your team, about your equipment, and about the boundary between work and private life.
- Closed-ended, rating. How would you rate your job satisfaction over the past quarter? (scale of 1 to 5)
- Semi-open. What do you like most about working from home? (options plus "other, please specify")
- Open-ended. What could your employer change to make working from home work better for you?
- Closed-ended, preference. If the decision were entirely yours, which form of work would you choose? (fully remote / hybrid / fully office-based)
Demographic block at the end: age category, years of experience, organization size, job position. Questions 2, 4 and 6 carry the variables from the hypothesis; the rest add context for the discussion chapter.
How to test the questionnaire before sending it out
Pilot testing is the step students skip most often, and it costs them the most. Pew Research Center describes pretesting as an essential step of questionnaire design: wording problems should surface before the questionnaire goes into the field, not on your final data.
A small group of people from the same target population is usually enough. Let them fill in the questionnaire, then ask how they understood the questions: which one was unclear, where a suitable option was missing, how long it took, and whether anything felt uncomfortable. Try analyzing the pilot answers as well. Do not include them in your final data set, and state in your methodology what you changed after the pilot.
How to distribute the questionnaire and handle the returns
For online collection, Google Forms and Microsoft Forms are the most common choices. Google Forms sends responses straight into a spreadsheet, Microsoft Forms connects to Excel. The limits of the free versions change, so check the current conditions before collecting, including how the data can be exported.
Expect that some of the people you approach will not respond, so contact more people than your target number. It helps when the questionnaire is sent out by someone the target group knows, such as the HR department or a lecturer, and when you send a reminder after a few days. How many responses you actually need is covered in the article on sample size and the number of respondents.
Keep a record of how many questionnaires you sent out, how many came back, and how many you excluded as incomplete. These numbers belong in your empirical chapter; your examiner reads the quality of your sample from them.
Anonymity, GDPR and research ethics
If your questionnaire does not ask for a name, an e-mail address, or a combination of details from which a person could be identified, you are not processing personal data. The GDPR says so directly in recital 26: "The principles of data protection should therefore not apply to anonymous information" (Regulation 2016/679, EUR-Lex).
Anonymity has a quiet trap, though. In a small company or a small school, the combination of age, position, and years of experience can point to a specific person. Work with categories instead of exact values.
State in your introductory text that participation is voluntary and that answers will be processed in aggregate, not individually. If you collect data from minors, you need consent from the school and from the legal guardians in advance.
Some faculties also review thesis research ethically. Depending on the country and institution, this is handled by a research ethics committee or an institutional review board (IRB), and the approval procedure takes time. Ask your department whether this applies to you and build that time into your schedule.
What to do with the answers and where the questionnaire belongs
The questionnaire lives in three places in your thesis. In the methodology chapter you describe why you chose it and how the collection went. In the empirical part you analyze the answers. In the appendix sits its full wording, exactly as the respondents saw it.
The first step after collection is data cleaning: remove unfinished responses and obviously mechanical patterns, such as the same value across an entire scale battery. Only then start counting. The step-by-step process with tables and charts is covered in the article on analyzing survey data in Excel.
If you are not sure whether your questions measure what your hypotheses claim, our authors can help you with the analytical part. Just place a no-obligation order and describe where you got stuck.
The most common mistakes in student questionnaires
Questions with no link to a hypothesis. The questionnaire has twenty questions, but the hypotheses cover three. The rest ends up as unanalyzed charts that the examiner calls filler.
Double-barreled and leading questions. A question asks about two things at once or hints at the desired answer, and the hypothesis gets confirmed artificially.
Overlapping or incomplete options. Respondents do not know where they belong, so they pick at random.
Demographics first. The first thing a respondent sees is a series of personal questions, and some people close the questionnaire right there.
No pilot. Wording mistakes only show up in the final data, when nothing can be fixed anymore.
Missing collection numbers. The thesis does not state how many questionnaires were sent out and how many came back.
Frequently Asked Questions
How many questions should a bachelor's thesis questionnaire have?
Universities do not prescribe a number; it follows from the variables in your hypotheses. Twelve questions that you analyze down to the last chart make a better questionnaire than twenty of which you use half.
Where does the questionnaire itself belong in the thesis?
Put the full wording in an appendix and refer to it in the text, for example as "Appendix A". The description of why you chose it and how the collection went belongs in the methodology; the results go in the empirical part.
Do I need ethics approval for a questionnaire?
It depends on your institution and your topic. Some faculties review all thesis research, others only sensitive topics or vulnerable groups. Find out at your department before collecting data; getting approval retroactively is difficult.
Can I use a ready-made questionnaire from someone else?
Yes. Standardized questionnaires are an advantage in some fields, because their properties have been tested and your results can be compared with other studies. You must cite them, though, and for protected instruments you must request the author's permission or a license. If you adapt an instrument, state in your methodology what you changed.
How do I handle a questionnaire for underage respondents?
Collecting data in schools requires consent from the school leadership and from the legal guardians, and the questionnaire must not ask anything that could identify the child. Obtain the consents in advance and mention them in your methodology.
Google Forms or Microsoft Forms?
For a typical student questionnaire both are sufficient; they differ in the interface and in the limits of the free version. Decide by how easily the data can be exported from the tool, not by how the form looks.