Directed acyclic graphs

Part 1: Identifying structural sources of bias

Learning objectives

1. Identify features of a DAG.
2. Understand the rules of d-separation.
3. Use a causal DAG to identify bias due to confounding and selection bias.
4. Identify control strategies to account for bias due to confounding and selection bias.
   
   

Part 2: Information bias and time-varying treatments

Learning objectives

1. Identify different types of information bias on a DAG
2. Recognize the structure of treatment-confounder feedback on a DAG.
3. Identify situations when stratification-based methods fail.
4. Devise an approach to draw your own causal DAGs.
   
   

Additional Readings or Resources

1. Causal Diagrams: Draw Your Assumptions Before Your Conclusion
2. Causal Inference: What If (Chapters 6-9, 19-20)