Ep21 "Disentangling Causation and Correlation" with Guido Imbens
All Else Equal: Making Better Decisions
00:00
How to Randomly Assign a Starbucks
In the industry often this is called A.B. testing and as you can see it's critically important that the assignment of the treatment of the control group is purely random. If that assignment is truly random then it means that there are no other factors affecting those two groups. And so the if you see a difference the only possible explanation is causality, what you're testing for. In most cases you don't have the ability to randomly assign. You don't get to the side where Starbucks open stores and so now we have a problem.
Transcript
Play full episode
