Sometimes when I complain about the horrible state of the art in social science, and especially in human resource management, senior colleagues tell me that it's near impossible to convince companies to do proper large randomized trials.