Quiz: can you tell correlation from causation?
Two things that happen together are not always linked by cause and effect. Test your eye with everyday situations.
Content produced by a Humanevi digital AI editorial agent.
Short answer: correlation means two things vary together; causation means one produces the other. Seeing the two together does not prove that one causes the other: there may be a third cause, coincidence or a reverse relationship. This quiz helps train that eye. It is educational material, with no intention of grading anyone.
Simple definitions
- Correlation: when two variables change in an associated way, such as height and weight.
- Causation: when changing one variable produces a change in the other, which is usually tested in experiments or well-designed studies.
How to use the quiz
For each situation, ask: "does this show that A causes B?" Then look at the alternative explanations.
Situations to assess
- Ice cream and drownings. In hot months, ice cream sales and drownings rise together. Does ice cream cause drownings? Most likely explanation: heat leads more people to the water and to buy ice cream. Heat is a third variable.
- Foot size and reading in children. Children with bigger feet read better. Explanation: age influences both.
- Do coffee drinkers live longer? An observational study finds an association. Caution: there may be lifestyle differences between the groups. Only further studies help assess causation.
- More firefighters, more damage in fires. Big fires call more firefighters and cause more damage. The cause is the size of the fire.
- Schools with better results and more books at home. There may be a relationship, but income, parents' education and other factors also play a part.
- The reverse. People who exercise are happier: does exercise make someone happy, or do happy people exercise more? It may run in both directions.
Questions that help
- Is there a third variable that explains both?
- Could the relationship be reversed?
- Could it be coincidence, especially with little data?
- Was the study an experiment with a comparison group or only observation?
- Is the sample large and representative?
- Is there a plausible mechanism?
How to interpret news
Be wary of headlines that say "X causes Y" based on association. Look for terms such as "associated with" and "linked to", and check the type of study.
Counterpoints
Correlation does not prove causation, but it can suggest hypotheses. In many cases, causation was established through accumulated evidence of different kinds.
