is this dimension aggregatable?

[hyunjimoon]

Andrew's notable quotes from 100 Stories of Causal Inference

• here: latent parameter (underlying truth of economy) to model bias. Three different things that people conflate: inference, difference of opinions, conflict of interest. Room full of people making decision. Inference is bunch of ppl trying to decide how much gumballs are in the jar, difference of opinion is a bunch of people in the room and they're going to have to decide whether abortion is legal in their state. conflict of interest is a bunch of ppl deciding how to divide up a certain pile of money. If you look into decision making literature, these tend to be conflated. If you can easily persuade me with data, it's an inference problem, but if you can't it's more of a difference of opinion and what about another question is would a side payment solve the problem? Side payment would solve conflict of interest but it wouldn't solve a difference of opinion. Another better example of difference of opinion is who should win the oscar this year. I think it should be movie A and you think movie B and someone comes in and says here's a side payment; i'm giving andy 100 dollars and he'll agree movie b is better -- that's ridiculous. that doesn't resolve difference of opinion question. it's a different view of the world. ... in economy people care about different things so you should be explicit rather than pretending you have a one dimensional measurement.

• people should be willing to learn (diffuse prior for different opinions) from the stories

• mechanism of learning from story: think about what was the implicit model of the world that this is an anomoly of then we can make our models better so our process of learning from stories is itself analogus with how we learn from studies