nicholas 🧨
@nicholas
this one tweet boiled the blood of revealed preference stans
1 reply
0 recast
5 reactions
rafa
@rafa
I remember learning this the hard way. I tried to “die on a hill” saying some data was right, and took two weeks, but the data was indeed wrong. Data is best when it reveals anecdote blind spots, and quantifies the impact of the anecdote. But the anecdote is truly the best source of “reality” if sourced right.
0 reply
0 recast
0 reaction
