There're Hofstadter-esque laws for selection effects and correlation≠causation where people overestimate their ability to learn from samples even when they factor these things in