Introduction to Ashtanga Collider Theory
Introduction: When Effects Distort Our Understanding of Causes In 1946, statistician Joseph Berkson noticed something peculiar in hospital patient data: diseases that should have been independent appeared to be negatively correlated. Cancer patients seemed less likely to have diabetes, and vice versa. This observation seemed to contradict everything known about disease independence. The answer to this paradox would eventually reveal one of the most counterintuitive principles in causal reasoning—the collider effect. A collider is a variable that sits at the convergence point of two or more independent causal pathways. The structure looks deceptively simple:Variable X → Collider Z ← Variable Y Two independent causes (X and Y) flow into a common effect (Z). But here’s the paradox: when we condition on the collider—when we only…
