Correlation mistaken for causation
Cum hoc ergo propter hoc, Corrélation prise pour causalité
Concluding that two phenomena are linked by cause and effect because they vary together.
The essentials
Argument diagram
A and B vary together.
So A causes B.
Why the argument is invalid
A correlation can have other explanations: a common cause (a confounder), reverse causation (B causes A), or coincidence, very frequent when many variables are compared.
In practice
Example
Ice cream sales and drownings rise together. The common cause is the heat, which also increases swimming.
Famous case
Hormone therapy for the menopause: observational studies showed that treated women had less heart disease. In 2002, the Women’s Health Initiative randomised trial found instead a slight increase in risk. In the observational studies, the women taking the treatment were, on average, better off and in better health.
Defending yourself
How to respond
Look for a common factor, consider reverse causation, and ask whether there is a controlled study (a randomised trial).
Don’t confuse
Not to be confused with: Post hoc ergo propter hoc
Mistaking correlation for causation relies on joint variation; post hoc, on mere succession in time.
Going further
Source
Tyler Vigen, “Spurious Correlations” (2015): for example, the number of films starring Nicolas Cage and the number of swimming-pool drownings in the United States.
Practise
In the app, each entry offers ways to practise recognising the concept. No sign-up: your progress stays on your device.