I am trying to run a difference-in-difference fixed-effects model to estimate the effect of the "discontinuation" of a treatment. The challenge here is that the treatment was discontinued from all treated units. My question is: How can I estimate the effect of this discontinuation? Is there anyway I can use the (initial) control group that has not been subject to the treatment nor the discontinuation?
In normal cases, when I estimate the effect of a treatment, not the "discontinuation" of a treatment, I use the code below kindly provided by Jeff Wooldridge.
Code:
xtreg outcome i.policy i.year, fe vce(cluster id)
Best regards,
Amira
0 Response to Difference-in-difference fixed effects for discontinued treatment
Post a Comment