intro-causal.qmd

# Causal-Inference

Statistical models can only tell us about differences in real-world distributions. But causal inference is about
comparisons between potential-world distributions.

![An illustrative example](correlation-causation.png){#fig-cor-caus-death}

@fig-cor-caus-death is a bit misleading; there isn't even a correlation between (confusing correlation and causation) and dying. This (and the usual milk-criminal example) are both more about predictive accuracy and lack of specificity.

No causes in, no causes out.