Economics
Saved & recent

Correlation and causation

Correlation describes association; causation asks how an intervention changes outcomes. Common causes, selection and reverse causality can produce association without the proposed causal effect.

Read first
Distributions, means and medians
Related
Causal identification, Randomized experiments
Further reading
CORE Econ: Collecting and analysing data from experiments
X / arbitrary units Y / arbitrary units possible causes X Y direct X Y common cause Z X Y reverse
Hypothetical observations in arbitrary X and Y units. Arrows are competing causal hypotheses, not effects identified by the plotted association. Timing, design and substantive assumptions must distinguish them.

View in the full Economics collection