Causal Inference
Why a comparison between the treated and the untreated can carry the wrong sign, what randomisation actually buys, and the rule that says which variables to adjust for - including the ones that make the answer worse.
Sign in to take quizzes, earn XP, and unlock stages as you reach 90% mastery.
Confounding and the Reversal
25 min · 100 XPA treatment that helps by the same margin in every subgroup while appearing to harm overall, the mechanism that produces it, and the standardisation that recovers the right answer.
Open lesson →Sign in to take the 3-question quiz.
Randomisation and What It Buys
25 min · 100 XPWhy assigning treatment by a coin removes confounding by construction rather than by measurement, what it does not fix, and the measured difference between letting doctors choose and letting chance choose.
Open lesson →Sign in to take the 3-question quiz.
Graphs, Colliders and What to Adjust For
30 min · 120 XPDrawing the assumptions as a graph so the rule for which variables to control becomes readable, the case where controlling for a variable manufactures an association from nothing, and the adjustment that recovers a known effect exactly.
Open lesson →Sign in to take the 3-question quiz.