I realized early on that most people who use causal methods do not entirely understand them. Do-calculus is deep math. Judea Pearl does a lot to simplify the concepts into manageable chunks of knowledge when he discusses them. He is frequently on Twitter, making clarifications to causal concepts.

Pearl has expertise that comes from the math, and he can use the math to make granular clarifications. Practitioners moving from theory to application frequently get pulled into the complexities and nuances. Exploring treatment effect estimation dives into one of these nuanced areas.

Treatment Effects Are Like Distributions

There are more than most people know, and each has use cases. Average Treatment Effect (ATE) and Conditional Average Treatment Effect (CATE) are the best known. There are others, and before getting to some of them, I need to explain the purpose of all these treatment effects.

High ROI Data Science

Why CATE ATE? Because She Was Hungry For…

Comments on this post are for paid subscribers