S2E25: Options for Small Samples
Quantitude
00:00
Basian Approaches to Small Sample Sizes
The idea behind baising approaches is that you're going to bring some prior information about aspects of your model. Mc nich has some really nice work that shows that if you use a basian approach without sufficiently informatov priors, that it's actually worse than frequents. "You're better off doubling down on good old school maximum likelihood"
Transcript
Play full episode
