1 of 31

The Effects of “Girl-Friendly” Schools: Evidence from the BRIGHT School Construction Program in Burkina Faso

Kazianga, Levy, Linden, and Sloan (2013)

2 of 31

Getting started

3 of 31

Preliminary Questions

  • What is the BRIGHT program?

4 of 31

Preliminary Questions

  • What is the BRIGHT program?
  • How are schools assigned to villages?

5 of 31

Preliminary Questions

  • What is the BRIGHT program?
  • How are schools assigned to villages?
  • What about this setup makes it possible to use a regression discontinuity design?

6 of 31

Preliminary Questions

  • What is the BRIGHT program?
  • How are schools assigned to villages?
  • What about this setup makes it possible to use a regression discontinuity design?
    • Idea: the way in which these schools were introduced discontinuously makes it possible to use RD!

7 of 31

Preliminary Questions

  • What is the BRIGHT program?
  • How are schools assigned to villages?
  • What about this setup makes it possible to use a regression discontinuity design?
    • Idea: the way in which these schools were introduced discontinuously makes it possible to use RD!
  • What is the economic theory set forth in the intro? (It’s subtle)

8 of 31

Broad Identification Strategy

  • What would be the issue with comparing school attendance in villages with BRIGHT schools versus villages without?

​

9 of 31

Identification Strategy (Broadly)

  • What would be the issue with comparing school attendance in villages with BRIGHT schools versus villages without?
  • What is the identifying assumption (that allows us to infer a causal effect?)

10 of 31

Identification Strategy (Broadly)

  • What would be the issue with comparing school attendance in villages with BRIGHT schools versus villages without?
  • What is the identifying assumption (that allows us to infer a causal effect?)
    • How might this assumption fail?

11 of 31

Research Design

12 of 31

Research Design

13 of 31

Research Design

14 of 31

Testing the Strategy

15 of 31

16 of 31

17 of 31

What do we want to see in this figure?

(Hint: The goal is still to check the identifying assumption)

18 of 31

Checking other DVs

19 of 31

20 of 31

21 of 31

Treatment Effects

22 of 31

23 of 31

Question 1: what is the estimand here? (ITT/ATE/ATET/ATEU/etc.?)

Question 2: Is this sharp or fuzzy RD?

24 of 31

Gender Heterogeneity

25 of 31

26 of 31

Which column is “fuzzy RD”? What is its estimand?

27 of 31

Which column is fuzzy RD?

What are the first stage and the second stage?

28 of 31

Wrapping up

29 of 31

Discussion Questions

  • Do you find this empirical strategy convincing? Is the research question important?

30 of 31

Discussion Questions

  • Do you find this empirical strategy convincing? Is the research question important?
  • What else would you have liked to see addressed in this paper?

31 of 31

Discussion Questions

  • Do you find this empirical strategy convincing? Is the research question important?
  • What else would you have liked to see addressed in this paper?
  • Do you think that this paper has external validity?
    • What limitations on external validity does the RD strategy add (if any)?