3 Consistency and positivity
3.1 Exchangeability in Observational Studies
Sep 17. Slides. After class, read Hernán and Robins 2020 Chapter 3.4-3.5. Optionally, read Hernán 2016.
What makes causal inference with observational data so challenging? Why is making treatment precise so important? These are the topics we’ll discuss in this lecture!
3.2 Lab: Exchangeability and Consistency Review
Sep 18 Slides.
You will go through an activity to really hone in on the concepts of exchangeability and consistency. Download the class assignment here.
3.3 Asking good causal questions
Sep 19. Slides. After class, read Hernán and Robins 2020 Chapter 3.3 & 3.6.
Good causal questions are structured so that credibility is strong for two key assumptions: positivity and consistency.
- Positivity. Every population subgroup receives every treatment value with non-zero probability
- Consistency. Potential outcomes \(Y^A\) are well-defined and linked to observable data