Reverse causation
Also known as: Reverse causality
The arrow runs the other way — the supposed effect is really the cause.
What it means
Reverse causation is the error of inferring that X causes Y when in fact Y causes X, or when the two reinforce each other. It is a pervasive threat to causal claims from cross-sectional and observational data, where the temporal order of variables is unknown or ambiguous. Distinguishing direction requires evidence that the cause precedes the effect, ideally through longitudinal data, experiments, or instrumental variables that fix the direction of influence. It is closely related to, but distinct from, confounding: here there is a real causal link, but its direction is misread.
Examples
Observing that troubled people see therapists, one might wrongly conclude therapy causes distress, reversing cause and effect.
Shops with the most security guards report the most theft, so guards must cause theft — or, far more likely, shops with a theft problem are the ones that hire guards.
Companies with the biggest advertising budgets post the highest sales, so advertising must drive sales, though firms plainly spend more on ads once the sales money is already coming in.
First described in Foundational to causal reasoning.