Causal inference when you can't experiment: difference-in-differences and synthetic controls
Linear Digressions
00:00
Is There a Causal Inference?
If you live in two parallel universes at the same time, and they're subject to the same economic shocks. And then you can subtract those from each other, and that starts to look a lot more like a causal effect. Yes, so you turn off the ads in some places, but not in all of them,. You look and see if the change is through time, is different for the people who keep the ads or where you turn off the Ads.
Play episode from 06:47
Transcript


