Average Treatment Effects: Causal Inference Bootcamp

This module introduces the concepts of the distribution of treatment effects, and the average treatment effect.
The Causal Inference Bootcamp is created by Duke University's Education and Human Development Incubator (EHDi) at Duke's Social Sciences Research Institute.
See our other modules on many related topics: modu.ssri.duke.edu

[@rhydaleghan2549]

simple but great illustration on the subject matter

[@bradconstant3778]

Nice explanation, thanks for putting it together

[@keynesianecon2917]

Thank you for the video. Could you provide references (i.e. empirical papers or textbooks) for the theoretical part?

[@pop_x3]

super helpful & easy to understand

[@joshuaellis7121]

This is so friggen useful. Thanks!

[@kanchanpathak3555]

Very well explained

[@mateusbringel]

Note that, when your outcome is lethality, you can have the opposite. Negative is the protective factor.

[@arijitdey3112]

Are you not contradicting yourself bt saying first "We fundamentally can't learn unit level causal effect" at 1:20 and secondly, after introducing ATE, by saying "take everybody in your population and you look at their unit level causal effect" at 2:20?

[@AmanSingh-xk2lv]

I caught that too. My guess is:
that's the theoretical definition. But in real life, we estimate it by taking the sample mean for the treatment group and the sample mean for control, And then a difference between those.