Key Idea: A correlation shows that two variables change together. A causal claim needs a design that fixes the order of events and removes the alternatives.
A link is not a cause
- Correlation: two things change together
- A third variable may cause both
- The arrow may point the other way
How to show a cause
- Change one thing, keep the rest the same
- Random allocation: groups alike at the start
- Placebo and blinding: against expectation
Checking the claim
- Internal validity: was it really the IV?
- Significance: is it more than chance?
- External validity: does it hold elsewhere?
When it is not simple
- Bidirectional: each affects the other
- Interaction: an effect depends on another factor
- One simple cause is rarely the whole story
The words to know
| Word | What it means |
|---|---|
| Correlation | Two things tend to change together |
| Causation | One thing makes another happen |
| Third variable | A hidden factor that causes both of the things measured |
| Reverse causation | The effect is really the cause |
| Independent variable | The one thing the researcher changes on purpose |
| Random allocation | Chance decides which group each person joins |
| Placebo | A fake treatment that looks like the real one |
| Internal validity | Confidence that the IV, and nothing else, caused the change |
| Statistically significant | Unlikely to be due to chance alone |
| Bidirectional relationship | Each of two things affects the other |
Was anything changed, or only measured? What else could explain it, and which way does the arrow point?
Exam-style question
Your class practical was an experiment. Your class tested whether the colour of a drink affects how sweet it tastes. Thirty participants each tasted the same lemonade: a coin flip decided whether theirs was dyed red or left clear. Each rated its sweetness from 1 to 10. Participants tasted in the same room and could see one another's drinks. Explain the concept of causality in relation to the experiment in your class practical.
Model answer plan
See the mark-by-mark plan — for / against / judgement, with marking guidance — in study mode.
Important: "Causes", "improves" and "leads to" need an experiment with a fair comparison. For anything else, write "is linked to".
Tap each card to check yourself.
Which of the three conditions for causation does a correlation meet? Only that the two variables vary together.
Extraneous or confounding: a variable that differs between the conditions? Confounding.
What does a wait-list control leave uncontrolled? Expectation: the waiting group knows it is not yet treated.
Does p < 0.05 show that the IV was the cause? No. It shows that chance is an unlikely explanation.
Bidirectional ambiguity or bidirectional relationship: "the design cannot tell the direction"? Bidirectional ambiguity.
What is an interaction? The effect of one variable depends on another.
Exam tips
- Ask first: was a variable changed, or only measured?
- Give a third variable or a reversed direction that fits this study.
- For an experiment, check how people got into the groups.
- Keep "significant" and "caused by the IV" apart.