Lesson 2 · Causal Inference · Step 3 of 6
Confounder control and drawing your assumptions
Ice cream sales and violent crime rise together, and no amount of staring at the data settles why, because a dataset carries no memory of what produced it. A causal diagram is where you write the story down before estimation starts. A fork is one variable causing two others, making them move together. A pipe is the effect passing through a middle variable. A collider is where two causes meet. Forks confound; colliders look like confounders and behave in the opposite way.
Also covered: causal diagrams · DAGs · forks, pipes and colliders · collider bias
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