I ran a field experiment across 100 cities. The cities belong to two types: Type_A(50 cities) and Type_B(the other 50 cities). From January 1 to February 30(The Pre-Treatment Period), I only collected data on my variable of interest y. From March 1 to April 30 (The Post-Treatment Period), I simultaneously applied two treatments on two different sets of cities: I randomly chose 10 Type_A cities and 10 Type_B cities for Treatment_1, and another 10 Type_A cities and another 10 Type_B cities for Treatment_2. I plan to use DiD (Difference-in-Difference) for modeling. How does the specification look like? I am confused because there are several things going on together: different types of cities, multiple treatments, before and after periods, etc.

In addition, if I want to implement the specification in Stata, how to do that? More importantly, how to compute the marginal effect of the treatments in Stata?

Thanks a bunch in advance.

  • $\begingroup$ What's the measure of your time component? Weeks? Were your treatments implemented at the same time? $\endgroup$ – Andy May 4 '14 at 22:40
  • $\begingroup$ The measure of time is day, and yes, the treatments were implemented at the same time (started on March 1). Thanks. $\endgroup$ – user44968 May 5 '14 at 16:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.