We really should teach courses on causal inference
May 12, 2011 § Leave a comment
There are a few of these courses around, but not many. Do we really want our youth learning this stuff on the street corner? From economists?
1. What is causation? What isn’t?
2. Experiments and extrapolation
3. Modeling experiments
4. The key concepts
5. The problem with regression
6. If you must do regression
7. Setting up a prospective study
9. Instrumental variables
10. The front door
11. The back door
12. General graphical models
13. Causal discovery is hard
14. Attempts at causal discovery
I would teach the controversy. I’d include a variety of perspectives: Freedman, Glymour, Pearl, Rubin. Fitting these perspectives into a coherent course would not be straightforward.