What is sensible for your agents should be sensible for yourself

摘要

I wish to discuss a question which I have long found perplexing. Why do some researchers develop Bayesian theoretical models of human behavior, and then use frequentist techniques for their data analysis? Long ago in Poirier (1988) I remarked: “Simply put, researchers should demand of themselves what they ask of the actors in their models…. I can only encourage such colleagues to be as demanding in analyzing their own behavior as they are of the actors in the parables they espouse.”