Lesson 2 · Causal Inference · Step 4 of 6
Good paths and bad paths
Between any two variables there are usually several routes, and identification is knowing which carry the causal effect and which carry confounding. A backdoor path is a route that runs into the exposure rather than out of it, and leaving one open is what makes an estimate wrong. The adjustment set is the smallest group of variables you must hold constant to close them all. Some variables must be adjusted for, some must not, and the diagram — not instinct, and not the regression output — tells you which is which.
Also covered: backdoor paths · adjustment sets · d-separation · bad controls
Alongside this step
- Slides and transcript — the whole video
- Read the chapter — Causal Inference
Lesson 2 · 6 steps
The Validity Toolkit · lesson 2 of 4