Randomization and complete data on the randomized experimental units are needed for cause-and-effect inference. What happens when that logic is broken?
Randomization and complete data on the randomized experimental units are needed for cause-and-effect inference. What happens when that logic is broken?