Could someone explain the math/theory behind what Gelman is saying around the 26:50 mark where he stats the 6 dishes in 1 large dish have a 1/sqrt(6) in them and how that makes it better? Not criticizing just trying to learn.
The case studies that he mentioned brougth me insights that none of my 4 statistical class profesors did (statistics 101, inference and decision theory, prediction methods and experiment design). Can you recommend some resources to dig deeper into case studies issues with double variance, aggregated data points and bias?
