Suppose a State is considering whether to require that all public school teachers have a master's degree. Currently, only 40 percent of the teachers in the state have such a degree.

a. A researcher conducts a cross-sectional analysis that compares test scores of students in the state whose teachers have master's degree to the test scores of students in the state whose teachers do not. The researcher finds that students whose teachers have master's degrees score significantly higher on the standardized tests. Why might such a study be biased?

b. Now suppose another state conducted an experiment in which 500 students were randomly selected to be either in a class taught by a teacher with a master's degree (treatment group) or a class with a teacher without a master's degree (control group). This experiment found that the treatment group scored significantly better than the control group. How useful is this experiment in informing the state's decision about whether to have a master's degree requirement?