“In many empirical studies in social sciences, causal effects are estimated through linear regression [which has an] exogeneity assumption [that] combines unconfoundedness with functional form and constant treatment effect assumptions that are quite strong, and arguably unnecessary.'' -- Imbens and Rubin (2015)
I think this will be the last in the series, as it’s kind of evolved a little anyway and I’m moving beyond Tymon’s OLS theorem anyway. What I’m trying to do though is just better understand the properties of a few estimators with each of their assumptions slightly messed with. And so I guess today what I’ll do is just remind you of what these estimators are (as that’ll differentiate it from the previous entries), and then show you those same simulations as before. This is the last entry in “Lies, Damn Lies and OLS Weights”, which has been my journey down a rabbit hole involving OLS as well as revisiting my interest in matching and weighting in causal inference.
Now how am I goi…
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.