Designing Questionnaires and Surveys
A questionnaire is a set of written questions; a survey is the wider exercise of collecting that data. Good design keeps the data collected accurate, unbiased and easy to analyse.
Method
- Decide exactly what information you need before writing any questions, and keep the survey short and relevant.
- Write questions in simple, clear language and avoid leading or biased phrasing that suggests an expected answer.
- Avoid vague words like often or a lot, and replace them with defined, non-overlapping response bands, for example 0-2, 3-5, 6+.
- Give response options that are mutually exclusive and cover every possible answer, including other or prefer not to say where relevant.
- Avoid asking two questions in one (a double-barrelled question), and keep personal questions optional.
- Trial the questionnaire on a small group first (a pilot survey) and refine any confusing questions before the full survey.
Common mistakes
MistakeOffering age bands with a gap, such as 11-15 and 17-20, so a 16 year old has nowhere to answer.
InsteadCheck that the bands run on from one another with no gap and cover every possible answer. Bands such as under 11, 11-15, 16-20 and 21 or over leave nobody out.
MistakeUsing an unbalanced scale such as Excellent, Very good, Good, Poor, which has three positive options and only one negative one.
InsteadOffer balanced options with the same number of positive and negative choices and, where sensible, a neutral middle, such as Very poor, Poor, Average, Good, Very good.
MistakeAsking an open question, such as 'What do you think of the canteen?', when the answers will need to be counted and compared.
InsteadUse a closed question with set options when responses must be counted. Keep an open question for a follow-up where reasons matter, since its answers are slow to sort and hard to compare.
MistakeAsking about a long period from memory, such as 'How many films did you watch last year?'
InsteadAsk about a short, recent period such as last week, or a typical week. People find it hard to recall a whole year accurately, so many answers are guesses.
Worked examples
Example 1
A student wants to find out how much time people spend exercising each week. Critique the question 'Don't you agree that you should exercise a lot each week?' and write an improved version.
- Identify flaw 1: the question is leading, since 'Don't you agree' suggests the expected answer.
- Identify flaw 2: 'a lot' is vague and undefined, so answers cannot be fairly compared.
- Remove the leading phrase and ask neutrally: 'How many hours do you spend exercising in a typical week?'
- Give clear, non-overlapping response options: 0-2 hours, 3-5 hours, 6-8 hours, 9 hours or more.
- Check the options cover every possible answer, with no gaps or overlaps.
Example 2
A student's questionnaire asks: 'How many portions of fruit and vegetables did you eat yesterday?' with the options 0-1, 2-3, 4-5. In a pilot with eight students, one wrote '7' in the margin, two ticked two boxes and said they were unsure what counts as a portion, one said yesterday was a birthday and not a normal day, and the rest answered easily. Use the pilot to improve the question and its options.
- Treat each problem as evidence. The student who wrote 7 shows that the options stop too soon: they must cover every possible answer, so an option such as 6 or more is needed.
- The two students who ticked two boxes did not know what counts as a portion. The question needs a short definition of one portion, taken from a reliable source.
- One answer was affected by an unusual day. Replace 'yesterday' with 'on a typical school day' so that answers describe normal habits.
- New question: 'On a typical school day, how many portions of fruit and vegetables do you eat? (One portion is defined here.)' Options: 0-1, 2-3, 4-5, 6 or more.
- Check the options: they do not overlap, they leave no gap, and every whole-number answer has a place.
- Final answer: the pilot shows that the question needs a definition of a portion and a typical day, and the options need a top category so that every answer can be recorded.
Example 3
A club asks four members, 'How many hours did you spend practising last week?' The recorded answers are 0 hours, 2 hours, 4 hours and one blank. Explain why the blank must not automatically be recorded as 0 hours. Work out the mean for the three recorded answers, then show the mean that would result if the blank were incorrectly treated as 0. Suggest how the response options could make a genuine zero clear.
- A recorded 0 hours means the member answered that they spent no time practising. A blank gives no value, so it does not tell us whether that member spent 0 hours or simply did not answer.
- For the three recorded answers, add 0 + 2 + 4 = 6 hours. There are three answers, so their mean is 6 / 3 = 2 hours.
- If the blank is wrongly entered as 0, the total stays 6 hours but the count becomes four. The calculated mean would be 6 / 4 = 1.5 hours.
- The different means show that coding a missing answer as zero changes the data summary and claims the person gave an answer they did not give.
- Use a clear response option such as '0 hours' for a genuine zero and leave a separate 'No answer' or blank status for missing data. The questionnaire must distinguish the two.
- Final answer: keep the blank as missing; the mean of the three recorded answers is 2 hours, while incorrectly coding the blank as zero gives 1.5 hours.
Practice questions
Q1[1 mark]What is wrong with the question 'Do you not think school lunches are too expensive?'Show answer
Answer: It is a leading question (it implies the expected answer).
Q2[2 marks]Give one reason why the response options '0-5, 5-10, 10-15' are a poor choice for hours of homework.Show answer
Answer: The bands overlap (5 and 10 each appear in two bands).
Q3[1 mark]Rewrite the response options '0-5, 5-10, 10-15' so that they no longer overlap.Show answer
Answer: For example: 0-4, 5-9, 10-14 (or 0-5, 6-10, 11-15).
Q4[1 mark]What is wrong with the question 'How old are you and do you smoke?'Show answer
Answer: It is a double-barrelled question (it asks two things at once).
Q5[1 mark]A survey question offers the response options 'Never, Sometimes, Always'. State one improvement that could be made.Show answer
Answer: Add more precise, defined bands (for example, based on a number of times per week), since 'Sometimes' is subjective.
Q6[2 marks]Explain why a pilot survey should be carried out before the full survey is sent out.Show answer
Answer: To test that the questions are clear and work correctly, so problems can be fixed before the real data is collected.
Exam-style questions
Written in the style of a GCSE Statistics exam paper, with a full mark scheme.
Exam tips
- When asked to write a question, write the exact wording and every response option, not a description of them; a question can only be judged once it is written out.
- In a critique, say what each flaw does to the data (leading wording pushes answers one way; vague words mean different things to different people) as well as naming it, then correct it in your rewrite.
- Read each question as a respondent: could two people with the same true answer tick different options? If so, tighten the wording or the options.
Q1[2 marks]A researcher wants to know if people support a new bypass. She asks: 'Most people think the new bypass is a great idea, do you agree?' Identify two reasons this is a poor survey question.Show mark scheme
Tick each line you got.
Nothing ticked yet - 2 available
Q2[4 marks]A student is investigating how much pocket money Year 8 students receive each week. She writes this question: 'How much pocket money do you get, loads or not much?' (a) Identify one flaw in this question. (b) Rewrite the question with four suitable, non-overlapping response bands.Show mark scheme
Tick each line you got.
Nothing ticked yet - 4 available
Q3[3 marks]A school wants to survey students about how they travel to school. (a) Suggest a suitable closed question with response options. (b) Explain why a pilot survey should be used first.Show mark scheme
Tick each line you got.
Nothing ticked yet - 3 available
See real GCSE Statistics past-paper questions, with official mark schemes
This topic in real past papers
We have tagged some boards' past papers question by question (not every board or year yet). This topic comes up in:
- AQA GCSE Statistics past papers: 5 questions across 5 papers, latest tagged June 2022
- Edexcel GCSE Statistics past papers: 4 questions across 4 papers, latest tagged November 2021
Free printable worksheet
Want more practice on paper? Download the Designing Questionnaires and Surveys worksheet pack - 23 pages of exam-style questions with a full mark scheme. Free to download, no card or account. Print it for personal and classroom use.
This topic is chapter 9 of GCSE Statistics Foundation Essentials 1 and chapter 9 of GCSE Statistics Higher Essentials 1, the whole course as one free printable PDF. Its worksheet is also chapter 1 of GCSE Statistics Foundation Topic Workbook 2 and chapter 2 of GCSE Statistics Higher Topic Workbook 2, free printable practice books.
Other versions of this worksheet: