Things get a bit more complicated when you’re dealing with experiments, in which an intervention of some kind occurs, but the same basic idea holds. In an experiment, you can only generalize results to an entire population if you selected the experimental sample randomly from that population. Further, you can only argue that some intervention…