How to design a thesis questionnaire: 8 rules and examples
Design the questionnaire for your thesis or final-year project: objectives, validated scales, question types, wording, Google Forms, pre-test and ethics.
Data Académique
10/13/20264 min read
A badly designed questionnaire brings the whole analysis down: ambiguous questions, unusable answers, a Cronbach's alpha that is too low… Yet a few simple rules are enough to get clean data that is easy to analyze in SPSS or Excel. Here is the method, from the first question to the pre-test.
In short: a good thesis questionnaire starts from the research objectives, reuses validated scales (often Likert scales), asks short and neutral questions, is organized from general to specific, is pre-tested with 5 to 10 people and respects respondents' anonymity.
1. Start from your objectives, not from question ideas
Every question must serve an objective or a hypothesis of your thesis. Before writing a single question, make a three-column table: objective or hypothesis, variable to measure, question numbers. For example:
H1 (ad credibility influences purchase intention): perceived credibility, Q8 to Q11 (Likert scale); purchase intention, Q12 to Q14.
Describe the sample: age, sex, education level, Q1 to Q4.
A question that does not fit any row should be deleted. It is also the best way to keep the questionnaire short.
2. Reuse validated scales
To measure a concept (satisfaction, stress, trust, engagement…), look for a scale already published and validated in scientific articles. Cite the author and adapt the wording to your context without changing the meaning. It is more solid than inventing your own questions, and your jury will notice.
If you translate a scale (from English into French or Arabic), have the translation reviewed and test it.
3. Choose the right question type
Single-choice closed questions: sex, age group, city. Easy to code and analyze.
Multiple-choice questions: useful but harder to analyze (one variable per option in SPSS).
Likert scales (1 = strongly disagree to 5 = strongly agree): ideal for measuring opinions and attitudes. Keep the same number of points for all scales.
Numeric questions: age in years, number of hours… Specify the unit.
Open questions: limit them to one or two; they take a long time to analyze.
4. Word each question carefully
One idea per question. "Are the staff competent and friendly?" mixes two things.
Simple words, understood by your whole population.
No leading questions: "Don't you think that…" pushes the answer.
Complete, non-overlapping options: "18-24, 25-34" and not "18-25, 25-35".
Add "Other" or "I don't know" when relevant.
5. Organize the questionnaire
A short introduction: who you are, the objective, the length ("5 minutes"), anonymity.
Easy, general questions first.
Thematic blocks (one dimension per block).
Socio-demographic questions, often at the end.
A thank-you message.
Aim for 5 to 10 minutes to complete. Beyond that, the dropout rate rises sharply.
6. Choose the distribution tool
Google Forms is free, simple and exports answers to Google Sheets, then to Excel or SPSS. For patients or people with little internet access, plan a paper version, and possibly an Arabic version. Before distributing, calculate how many answers you need: see our article on sample size.
7. Run a pre-test
Have 10 to 30 people close to your target population fill in the questionnaire. Ask them what was unclear, too long or uncomfortable. Also check that the data export works and calculate a first Cronbach's alpha on your scales. Correct, then launch the real data collection. Mention this pre-test in your methodology.
8. Think about ethics
Obtain the necessary permissions (company, hospital, school).
Inform participants and obtain their consent.
Guarantee anonymity: do not ask for names or phone numbers unless necessary.
The complete data analysis path
For a questionnaire-based thesis, the steps follow this order:
cross categorical variables with the chi-square test;
test the effect of several variables with linear regression;
for models with latent variables: SmartPLS and PLS-SEM.
Example Likert scale items
To measure purchase intention, respondents rate their agreement from 1 (strongly disagree) to 5 (strongly agree) with statements such as:
"I intend to buy a product I saw in an Instagram ad."
"I would recommend products I saw on Instagram to my friends."
"I am likely to buy a product presented on Instagram soon."
Several items per variable then let you calculate a score and check its reliability.
Nursing student? See the ISPITS nursing thesis structure, section by section.
If your questionnaire measures latent constructs, our SmartPLS (PLS-SEM) course guides you all the way to the structural model.
To process your responses afterwards, see our statistics course: Excel, SPSS and SmartPLS.
Frequently asked questions
How many questions should a thesis questionnaire have?
Aim for 10 to 15 minutes of response time, often 20 to 40 questions depending on the number of dimensions measured. Beyond that, the dropout rate rises sharply.
5-point or 7-point Likert scale?
Both are valid. The 5-point scale is simpler for respondents; above all, keep the format of the validated scale you use.
How do I code answers for SPSS?
Give each option a numeric code (1 = Male, 2 = Female) and declare the labels in Variable View. Our SPSS course covers all this preparation.
We build your questionnaire with you
Data Académique helps you design a questionnaire aligned with your hypotheses, choose the right scales and prepare the analysis. Discover our thesis support or message us on WhatsApp.
