17: Differences-in-Differences

Question 1

DiD strategy

Answer 1

taking the difference of outcomes from the post-treatment period from the pre-treatment period

Question 2

parallel trends assumption

Answer 2

absent of treatment, difference in mean outcomes in the second period would be the same as the observed difference in the initial period

absent of treatment, treatment and control would have experienced the same changes in mean outcomes

Question 3

implementing DiD with regressions

Answer 3

dummy variable for treatment, dummy variable for time period, interaction term

point estimate for interaction term shows the mean difference in the change of outcome between treatment and control