experimental studies have excellent properties when it comes to eliminating bias, but they may be difficult or impossible to perform. Observational studies have knotty problems with bias, but the data are relatively easy to obtain. Can one obtain some of the advantages of each? A class of observational studies known as quasi-experimental studies (also known as natural experiments) are used by empirical economists to estimate a causal relationship. These studies identify situations in which outside circumstances stances in effect randomly assign people to control and treatment groups. The difference between and experiment and a quasi-experiment is that an experiment explicitly randomized people into a treatment or control group. whereas a quasi-experiment uses observational data but relies on circumstances outside of he researcher's control that naturally lead to random assignment.