How to Design a Questionnaire for Your Thesis: A Step-by-Step Guide
Good data starts with good questions. A practical, step-by-step guide to designing a thesis questionnaire that answers your research questions and produces clean, analysable data.
Your questionnaire is the bridge between your research questions and your results chapter. If a question isn't on the form, you can't analyse it later. If it's badly worded, the answers can't be trusted. Yet many students draft their questionnaire in an afternoon.
This guide walks through designing a questionnaire that your supervisor will approve, your respondents can answer, and your statistics software can analyse.
Step 1: Start from your objectives, not your questions
Write down each research objective or hypothesis. Under each, list the variables you need to measure, then draft the questions that measure them. This simple map, objective → variable → question, does three things:
- It shows you which questions are missing.
- It exposes "nice to know" questions that don't serve any objective. Cut them, because shorter surveys get better responses.
- It becomes the basis of your analysis plan and codebook.
Where possible, use validated scales from published studies (for example, an established job-satisfaction or food-security scale) rather than inventing your own. Cite the source and note any adaptations.
Step 2: Choose the right question type
| Question type | Example | Use it for |
|---|---|---|
| Single choice | Marital status | Mutually exclusive categories |
| Multiple choice | Sources of income (tick all) | Several answers can apply |
| Likert scale | Strongly disagree … Strongly agree | Attitudes and perceptions |
| Numeric | Age in years, household size | Counts and measurements |
| Open-ended | "What challenges do you face…?" | Exploring reasons, quotes |
Collect numbers as numbers. Asking age in years gives you far more analytical options than age bands like 18–25 or 26–35. You can always group exact ages later, but you can't recover exact ages from bands.
Step 3: Word questions clearly
Most questionnaire problems are wording problems. Avoid these common traps:
- Double-barrelled questions: "Is the service fast and affordable?" asks two things. Split it into two questions.
- Leading questions: "Don't you agree that mobile money has improved your life?" pushes respondents towards yes.
- Vague time frames: "Do you often buy fertiliser?" Ask "In the last 12 months, how many times did you buy fertiliser?"
- Jargon: write for your respondents, not your examiners. Translate technical terms into everyday language.
- Overlapping options: income bands of 0–10,000 and 10,000–20,000 overlap at 10,000. Make categories exclusive and exhaustive, and add "Other (specify)" where needed.
Step 4: Order and structure the questionnaire
- Introduction and consent: who you are, the purpose, confidentiality, and the right to withdraw.
- Easy, engaging questions first to build rapport.
- Group questions by topic into clear sections.
- Sensitive questions later, once trust is established (income, health, personal experiences).
- Demographics at the end for self-completed surveys, or at the start for interviews where they control later questions.
Step 5: Plan your skip logic
Skip logic (also called branching) shows a question only to the people it applies to. For example, only respondents who answer "Yes, I own a business" should see questions about their business. Write skip instructions on your paper draft, such as "If No, go to Section D", before building anything digitally. This logic is exactly what a digital form will enforce automatically.
Step 6: Code every answer before you collect data
Give each question a short variable name (hh_size, q12_income) and each answer option a numeric code (1 = Male, 2 = Female). Agree on codes for "Don't know" and "Refused" (for example 98 and 99) and use them consistently. This becomes your codebook, and it makes entry and analysis in SPSS, Stata or R much smoother.
Step 7: Pilot test, always
Test the questionnaire with 5–15 people similar to your sample (but not part of it). Time how long it takes, note questions people hesitate over or misunderstand, and check the answers actually give you the data your analysis needs. Revise, then get final approval from your supervisor or ethics committee.
Step 8: Go digital (carefully)
Digital forms remove manual data entry and can block impossible answers as they're typed. But a digital form is only as good as its build: missing skip logic, untested validation or messy variable names cause problems that surface during analysis, when it's too late to re-collect.
If you're deciding between platforms, read our comparison of KoboToolbox, ODK, Google Forms and Microsoft Forms.
Already have an approved questionnaire? We'll turn it into a tested Google Form, Microsoft Form, KoboToolbox or ODK form, with skip logic, validation and analysis-ready variable names built in. Your questionnaire content stays exactly as approved.
Quick checklist
- Every question maps to an objective or variable.
- No double-barrelled, leading or vague questions.
- Answer options are exclusive, exhaustive and coded.
- Skip instructions are written down.
- The questionnaire has been piloted and revised.
- The consent statement is clear, and ethics approval is in place.
Need your questionnaire built properly?
We design and test data collection tools on Google Forms, Microsoft Forms, KoboToolbox, ODK and SurveyCTO. Free quote within 24 hours.
Get a free quote →