Causal inference when you can't experiment: difference-in-differences and synthetic controls
Linear Digressions
00:00
How to Pair Up Cities That Look Like Each Other
Before monkeying with any of the ads, you try to pair up cities that are as close as possible. So we were talking a little bit right before we started recording, and i was trying to brain storm about cities that might look sort of similar. The ones i came up with were like cleveland and saint louis. These are both cities that are in the midwest. They're kind of post industrial. I don't know anything about their race or socio economic profiles but i don't have any information that tells me they're dramatically different.
Play episode from 08:49
Transcript


