Part II · Causal Frameworks & Experiments

Module 05. Structural Causal Models & DAGs

Write assumptions as graphs and let them decide which variables to control for, and which never to touch.

Social policy: job-training DAGs and collider-bias simulations about 6 hours3 lessons

Module progress0/3 lessons · 0%

Lessons

DAGs and Structural Causal Models: Chains, Forks and Colliders

· Draw a directed acyclic graph (DAG) for a policy question with networkx
· Write the structural causal model (SCM) that a DAG represents and simulate it
· Recognise the three building blocks (chains, forks, colliders) and what each implies for correlation

· State the backdoor criterion and find valid adjustment sets programmatically
· Show by simulation that controlling for a mediator removes the effect you want to measure
· Estimate an effect with the front-door formula when the confounder is unobserved

· Simulate collider bias with the classic "Hollywood actors" example
· Recognise sample selection (looking only at employed workers, surviving firms) as conditioning on a collider
· Explain M-bias, where controlling for a pre-treatment variable still adds bias