'Until not much more than 20 years ago, economists frequently lamented the fact that they were limited in their empirical analyses to statistical assessments of market behavior, because controlled economic experiments were (thought to be) infeasible, unethical, or both. Much has changed in the...