Experimental Impact Evaluation Design – Quiz 5

1. Which of the following statements about stratified random assignment is true? (Choose all that apply)

Suppose that you wanted to stratify random assignment of the EG baseline student list on baseline test performance. One option could be to identify the mean test score, create a binary variable for whether a student scored above the mean or below the mean, and stratify on that binary variable. Implement this treatment assignment, assigning 50% of students to treatment and 50% of students to control, and use the ‘rround’ option in gsample to ensure that the overall assignment is as close to 50/50 as possible.

2. How many students scored above the average baseline test score?

3. Conduct a balance test comparing average test scores for your treatment versus control groups. What is the p-value for this test? Round to the nearest 0.01.

Now, instead of stratifying on the binary variable for above/below mean performance, stratify on total baseline scores itself.

4. How many strata are there in this random assignment?

5. Conduct a balance test comparing average test scores for treatment versus control groups under this alternative treatment assignment. What is the p-value for this test? Round to the nearest 0.01.