Two variables are correlated when they move together: as one rises, the other tends to rise (a positive correlation) or fall (a negative correlation). A correlation on its own never shows that one variable causes the other.
- Pick a scenario in the title bar. Each dot is one case (a day, a child, a fire).
- Predict first: does one variable cause the other? Then reveal the hidden variable. The dots are recoloured by it, from low (blue) to high (red), and you can see it drives both.
- r is the correlation coefficient, from −1 (perfect negative) through 0 (none) to +1 (perfect positive). It measures how closely the dots follow a straight line, not whether one causes the other.
Why a correlation might not be causal
- A third (confounding) variable drives both, as in every scenario here: temperature, age, how serious the fire is.
- Reverse causation: B might cause A, not A cause B.
- Chance: with enough variables, some will correlate by coincidence.
To show causation, researchers use an experiment: they change the independent variable themselves, keep other variables the same (or assign participants at random), and measure the effect on the dependent variable.
Common exam mistakes: writing "proves" for a correlation; forgetting that a correlation can be negative; and saying a correlational study can't be useful, when it can show that variables are related and suggest what to test in an experiment.
Objective: explain the difference between correlation and causation, and how a third variable can produce a correlation (for example, IB Psychology research methods; Cambridge International AS Level Psychology 9990, research methods; IB Theory of Knowledge, knowledge and evidence in the human sciences).
