Causal Inference: the Mixtape, the new unconfoundedness chapter (Taylor's version)
Listing out all the changes I'm planning on the unconfoundedness chapter
Apologies for two posts in same day — I messed up the automatic scheduling.
When I got the contract to write a revision to the Mixtape, I was excited, and I still am. But I am behind, and it was partly caused by (I think I said this already) dropping the probability and regression chapter. But many people, particularly those in industry, said the regression chapter was helpful. So I decided to move OLS into the chapter previously called Subclassification and Matching and just change the title to “Unconfoundedness”. Unconfoundedness is what most people think of when they think of estimating for causal effects after controlling for many covariates. It’s what they think of when they think of exogeneity from econometrics.
But the second I moved regression into that chapter, it completely overturned the chapter as there really was no way to put regression into the chapter and not address the elephant in the room which was why should one ever estimate anything using propensity scores or…
Keep reading with a 7-day free trial
Subscribe to Scott's Substack to keep reading this post and get 7 days of free access to the full post archives.