HandBook: https://matheusfacure.github.io/python-causality-handbook/01-Introduction-To-Causality.html

All content on this page is sourced from the above handbook. The information here won’t make much sense if you haven’t already gone through above handbook.

01 - Introduction To Causality

ATE: Average Treatment Effect

$$ ATE = E[Y_1 - Y_0] $$

ATT: Average treatment effect on the treated

$$ ATT = E[Y_1 - Y_0 | T=1] $$

Association is equal to ATT + Bias

$$ E[Y|T=1] - E[Y|T=0] = \underbrace{E[Y_1 - Y_0|T=1]}{ATT} + \underbrace{\{ E[Y_0|T=1] - E[Y_0|T=0] \}}{BIAS} $$

02 - Randomized Experiments

Randomized experiments is the best way to remove bias

03 - Stats Review

We find ATE, but is it statistically significant?

04 - Graphical Causal Models

3 types of graph structures: Linear, Fork, Collider. We saw how to think about conditioning/blocking through graph and how that relates to independence of events.

1. Confounding: Happens when treatment and outcome have a common cause