Data Analysis for Impact Evaluations – Quiz 7

1. When is the ideal time in the impact evaluation timeline to draft the pre-analysis plan?

2. Which of the following are appropriate ways to address multiple hypothesis tests? (select all that apply)

In Quiz 6 you estimated the interaction effect on treatment*total_bl. We’re going to execute a similar analysis here, with a slight twist to demonstrate how your findings can change if you change your model specification (and hence the importance of pre-specification). Generate a binary variable equal to 1 if the student scored above the average baseline score and 0 otherwise. Interact this binary variable with treatment, and regress Year 3 endline scores on treatment, the binary variable, and this interaction term. Include village pair fixed effects and don’t forget to cluster your standard errors.

3. What is the coefficient on the interaction term? Round to the nearest 0.01.

4. Which of the following statements is true?