Probability & Statistics · Statistics
Correlation vs Causation
Why correlated things often aren't causal: confounders, reverse causation, chance and selection, Simpson's paradox, how RCTs and natural experiments establish causation, and Bradford Hill's criteria.
Step 1 of 5 · Overview30 min read ·
Key idea
Prereqs: Regression; hypothesis tests. Next: Epidemiology; Monte Carlo. Sources: Judea Pearl & Dana Mackenzie, *The Book of Why* (2018); Angrist & Pischke, *Mastering 'Metrics* (2014); Bradford Hill (1965); Crash Course Statistics #8.Next: 1. Why correlations mislead