How do you keep an investigation fair and free of bias?
Explanatory and response variables, controlled and extraneous variables, control groups, and sources of bias in sampling and data collection.
A focused answer to AQA GCSE Statistics on controlling variables and bias, covering explanatory and response variables, controlled and extraneous variables, control groups and matched pairs, and the main sources of bias in sampling and data collection.
Reviewed by: AI editorial process; not yet individually human-reviewed
Have a quick question? Jump to the Q&A page
Jump to a section
What this dot point is asking
AQA wants you to identify explanatory and response variables, control extraneous variables to keep a test fair, understand the role of a control group and matched pairs, and recognise and reduce the main sources of bias in sampling and data collection. These ideas connect sampling and questionnaire design to the reliability of any conclusion you draw.
Explanatory and response variables
For example, when testing whether revision time affects a test score, revision time is the explanatory variable and the score is the response variable. Naming these correctly is the first step in designing a fair test, because everything else (what to control, what to measure) follows from knowing which variable is the cause and which is the effect.
Controlled and extraneous variables
Failing to control extraneous variables means you cannot tell whether the explanatory variable caused the change, or whether some uncontrolled factor did. If the app-using class also had the more experienced teacher, any score difference could be due to the teacher, not the app: the two effects are confounded. Listing and controlling the obvious extraneous variables (same test, same time, similar ability) is exactly what examiners reward.
Control groups and matched pairs
Without a control group you have nothing to compare against: if scores rose, you could not tell whether the app caused it or whether everyone improved anyway. Matching strengthens the comparison further by pairing a strong student in one group with a similar strong student in the other, so the groups start out alike.
Sources of bias
Common sources of bias include: a sampling frame that misses part of the population (an online-only list excludes people without internet); non-response, where people who do not reply differ systematically from those who do; self-selection, where only the most motivated respond; and leading questions that push answers one way. Random and stratified sampling, a complete and up-to-date sampling frame, and follow-ups to non-responders all reduce bias.
Exam-style practice questions
Practice questions written in the style of AQA exam questions on this dot point, with worked answer explainers. The year tag is the paper they imitate, not the source.
AQA 20194 marksA researcher tests whether a new revision app improves test scores. One class uses the app and another does not. (a) Identify the explanatory and response variables. (b) Give two variables that should be controlled to make the comparison fair.Show worked answer →
(a) Explanatory variable: whether the class uses the app (the factor being changed). Response variable: the test score (the outcome measured).
(b) Two controlled variables: the same test for both classes, and the same amount of revision time (others: same teacher, same topic, similar prior ability).
Markers reward correctly naming the explanatory and response variables and two sensible controlled variables that would otherwise confound the comparison.
AQA 20213 marksA survey about commuting is carried out by stopping people at a train station at am on a weekday. Explain why the results are likely to be biased, and suggest one improvement.Show worked answer →
The sample is biased because it only includes train commuters travelling at peak time; it misses people who drive, work from home, work different hours, or do not commute at all, so it is not representative.
Improvement: sample at several locations and times (or use a proper sampling frame of the whole population) to capture the full range of commuters.
Markers reward identifying the unrepresentative coverage (a sampling/selection bias) and a practical improvement that widens the sample.
Related dot points
- The statistical enquiry cycle, hypotheses, the stages of an investigation, and types of statistical problem.
A focused answer to AQA GCSE Statistics on the statistical enquiry cycle, covering the stages of an investigation, writing a hypothesis, the role of pilot studies, and how the cycle structures a real statistical problem.
- Populations, sampling frames, census versus sample, random, systematic, stratified, quota and cluster sampling.
A focused answer to AQA GCSE Statistics on sampling methods, covering populations and sampling frames, census versus sample, and how random, systematic, stratified, quota and cluster sampling work, with the stratified sample calculation.
- Data collection sheets, tally charts, questionnaires, open and closed questions, response boxes and avoiding leading or biased questions.
A focused answer to AQA GCSE Statistics on collecting data, covering data collection sheets and tally charts, open and closed questions, designing non-overlapping response boxes, and how to spot and fix leading or biased questions.
- Qualitative and quantitative data, discrete and continuous data, primary and secondary data, and categorical and ranked data.
A focused answer to AQA GCSE Statistics on types of data, covering qualitative and quantitative, discrete and continuous, primary and secondary, categorical and ranked data, and why the type controls which diagrams and calculations you can use.
Sources & how we know this
- AQA GCSE Statistics (8382) specification — AQA (2017)