Causal effects are comparisons of what did happen with what would have happened if people had received different treatments. Randomized treatment assignment has reduced this problem to the minor technical problem of drawing an inference about a finite population of people on the basis of a probability sample from that population. Expressed differently, if we […]