What you'll learn
This revision guide covers everything you need to know about designing effective questionnaires and conducting surveys for AQA GCSE Statistics. You'll learn how to write clear, unbiased questions, choose appropriate response formats, identify common design flaws, and understand how data collection methods affect the quality of statistical results.
Key terms and definitions
Questionnaire — A written set of questions used to collect data from respondents, typically self-completed without an interviewer present.
Survey — The overall process of collecting data from a sample of people, which may use questionnaires, interviews, or observations.
Response options — The predetermined choices provided for answers in closed questions (e.g., tick boxes, rating scales, multiple choice).
Leading question — A question worded in a way that suggests or encourages a particular answer, introducing bias into the data.
Hypothesis — A testable statement or prediction that the survey aims to investigate or prove/disprove using collected data.
Pilot survey — A small-scale trial run of a questionnaire with a few respondents to identify problems before full distribution.
Closed question — A question with fixed response options that limit answers to predetermined categories.
Open question — A question allowing respondents to answer freely in their own words without restricted choices.
Core concepts
Purpose and planning of questionnaires
Before designing any questionnaire, you must establish a clear purpose. Every question should relate directly to your hypothesis or research aim. Irrelevant questions waste respondents' time and reduce response rates.
Key planning steps:
- Define the hypothesis clearly (e.g., "Students who eat breakfast perform better academically")
- Identify what data you need to test this hypothesis
- Consider your target population and how you'll sample them
- Decide whether questionnaires or interviews are most appropriate
- Plan how you'll process and analyze the data collected
The target population determines question wording and complexity. A questionnaire for Year 7 students requires simpler language than one for adults. Similarly, sensitive topics (health, income, personal habits) need careful wording to encourage honest responses.
Types of questions and response formats
Closed questions generate quantitative data that's easy to analyze and compare. They work well for:
- Categorical data: "Which form are you in? Year 10 / Year 11"
- Ordinal data: "How often do you exercise? Never / Sometimes / Often / Always"
- Numerical data with options: "How many hours of sleep do you get? 0-4 / 5-6 / 7-8 / 9+"
Response options must be:
- Mutually exclusive — no overlap between categories (avoid "10-15" and "15-20" as 15 appears twice)
- Exhaustive — covering all possible answers, often including "Other" or "Prefer not to say"
- Equal in range for numerical data (except possibly the first/last categories)
Open questions allow detailed responses and can reveal unexpected information. However, they're time-consuming to analyze and difficult to quantify. Use them sparingly, typically when you cannot predict all possible answers.
Example: "What improvements would you suggest for the school canteen?" requires an open response.
Rating scales (e.g., 1-5 or 1-10) provide ordinal data. Odd-numbered scales (1-5) include a neutral middle option; even-numbered scales (1-4) force respondents toward positive or negative.
Characteristics of good questions
Effective questions are:
Clear and specific — Avoid ambiguity. "Do you exercise regularly?" is poor because "regularly" means different things to different people. Better: "How many times per week do you exercise for at least 30 minutes?"
Unbiased and neutral — Leading questions should be avoided. "Don't you agree that homework is excessive?" suggests the expected answer. Better: "Do you think you receive too much, too little, or about the right amount of homework?"
Not double-barreled — Each question should ask one thing only. "Do you enjoy Maths and English?" is problematic if someone likes one but not the other. Use separate questions.
Appropriate for the time period — Specify timeframes clearly. "How many fizzy drinks do you consume?" is vague. Better: "How many fizzy drinks did you consume yesterday?"
Free from assumptions — "How much do you enjoy watching football?" assumes the respondent watches football. Add a response option: "I don't watch football."
Common sources of bias
Response bias occurs when question design influences answers:
- Leading questions suggesting desired answers
- Loaded language with emotional connotations ("dangerous," "unfair")
- Embarrassing questions encouraging dishonest responses
- Missing response options forcing inappropriate choices
Sampling bias happens when the sample doesn't represent the population:
- Convenience sampling (asking only friends or people nearby)
- Voluntary response (self-selection attracts those with strong opinions)
- Time/location bias (surveying only during lunch excludes students in clubs)
Non-response bias arises when certain groups don't complete questionnaires:
- Overly long questionnaires with low completion rates
- Difficult questions that respondents skip
- Sensitive topics some people refuse to answer
Pilot surveys and improvements
A pilot survey tests your questionnaire on a small sample (10-20 people) before full distribution. This identifies:
- Ambiguous or confusing questions
- Missing response options
- Questions respondents find difficult or offensive
- Technical problems with online forms
- Time required for completion
After piloting, revise questions based on feedback. Common improvements include:
- Adding "Other (please specify)" to closed questions with unexpected answers
- Clarifying vague terms identified by respondents
- Removing or combining questions if the survey is too long
- Adjusting response scales if most answers cluster in one category
- Rewording questions that respondents found unclear
Data collection methods
Self-completion questionnaires are distributed for respondents to complete independently:
Advantages:
- Cost-effective for large samples
- Respondents may answer sensitive questions more honestly
- No interviewer bias
- Can be completed at convenient times
Disadvantages:
- Lower response rates
- No opportunity to clarify questions
- Cannot verify who actually completed it
- May have incomplete responses
Interviews involve an interviewer asking questions directly:
Advantages:
- Higher response rates
- Can clarify misunderstood questions
- Can ask follow-up questions
- Can verify respondent identity
Disadvantages:
- Time-consuming and expensive
- Interviewer bias possible
- Respondents may give socially desirable answers
- Requires trained interviewers
Online surveys (e.g., Google Forms, SurveyMonkey) combine features of both:
Advantages:
- Automatic data recording and analysis
- Can include skip logic (different questions based on previous answers)
- Easy distribution to large samples
- Low cost
Disadvantages:
- Excludes those without internet access
- May receive spam or duplicate responses
- Less personal than face-to-face methods
Worked examples
Example 1: Identifying and correcting poor questions
Question: A student is investigating shopping habits. Criticize the following question and write an improved version:
"How much money do you waste on unnecessary shopping?"
Solution (5 marks):
Criticisms (3 marks):
- The word "waste" is loaded language that suggests spending is bad (1 mark)
- "Unnecessary shopping" is biased as it implies the shopping is wrong (1 mark)
- No time period specified (yesterday/weekly/monthly) (1 mark)
- No response options provided, making data difficult to analyze (accept for 1 mark)
Improved version (2 marks): "How much did you spend on non-essential items last week? □ £0 □ £1-£10 □ £11-£20 □ £21-£30 □ £31+"
(1 mark for neutral wording, 1 mark for clear response options with appropriate ranges)
Example 2: Designing a question to test a hypothesis
Question: Callum believes that students who walk to school are fitter than those who travel by car. Design a suitable question with response options to collect data about method of travel to school. (3 marks)
Solution:
"How did you travel to school today?" (1 mark for clear, unambiguous question)
Response options (2 marks for appropriate, exhaustive, mutually exclusive options):
- □ Walked
- □ Cycled
- □ Car
- □ Bus
- □ Train
- □ Other (please specify) ___________
(Award 2 marks for 4+ appropriate options including "Other"; 1 mark for 3 options or if options overlap)
Example 3: Evaluating a questionnaire design
Question: Maya wants to investigate whether students prefer online or in-person lessons. Comment on the suitability of her questionnaire:
- What year group are you in?
- Don't you agree that online lessons are worse than in-person lessons?
- How many online lessons have you attended?
Solution (4 marks):
Question 1: Suitable/appropriate (1 mark) — collects relevant demographic data to compare year groups.
Question 2: Not suitable (1 mark) — this is a leading question that suggests online lessons are worse, introducing bias (1 mark).
Question 3: Partially suitable (1 mark) — relevant to the topic but needs response options rather than being open-ended, and should specify a time period such as "this week" or "this term."
(Alternative valid criticisms: no question directly asks about preference between online/in-person lessons; only 3 questions may not provide enough data)
Common mistakes and how to avoid them
Using overlapping response options — Categories like "0-5" and "5-10" both include 5. Use "0-4, 5-9, 10-14" or "0-5, 6-10, 11-15" to make options mutually exclusive. For continuous data, specify "5 ≤ x < 10" or use "up to 5, more than 5 to 10."
Forgetting to specify time periods — "How many times do you eat fast food?" is meaningless without "per day," "per week," or "per month." Always include a clear timeframe.
Creating questions that don't help test the hypothesis — Every question must serve a purpose. Asking someone's favorite color when investigating exercise habits wastes time and reduces response rates.
Writing vague or subjective terms — Words like "often," "regularly," "expensive," or "many" mean different things to different people. Use specific numbers or defined categories instead.
Missing the "prefer not to say" option — For sensitive questions (age, income, ethnicity, health), always include an opt-out option to avoid non-response or dishonest answers.
Not piloting the questionnaire — Even experienced researchers test questionnaires first. Piloting reveals problems you won't spot yourself, saving time and improving data quality.
Exam technique for "Data Collection: Questionnaires and Surveys"
"Criticize" or "Comment on" questions require you to identify specific problems (bias, ambiguity, missing options) and explain why they're problematic. Don't just say "it's bad" — explain the effect on data quality. Typically 1 mark per valid criticism with explanation.
"Design a question" tasks need both the question wording AND appropriate response options. Even if the question itself is perfect, you'll lose marks without suitable options. Check they're mutually exclusive, exhaustive, and match the data type needed.
"Improve this question" means you must write a better version, not just say what's wrong. Show both criticism of the original AND your improved question. Usually 2-3 marks: 1-2 for identifying problems, 1-2 for a correct improved version.
Use technical vocabulary accurately — terms like "leading question," "biased," "mutually exclusive," "exhaustive," and "hypothesis" demonstrate understanding and can earn marks. Vague statements like "it's not very good" won't score.
Quick revision summary
Effective questionnaires use clear, unbiased questions with appropriate response options that are mutually exclusive and exhaustive. Avoid leading questions, double-barreled questions, and vague terms. Specify time periods and ensure questions directly relate to your hypothesis. Pilot surveys identify problems before full distribution. Choose between closed questions (easy to analyze) and open questions (detailed responses) based on your data needs. Consider sampling methods to avoid bias and maximize response rates.