Diff in diff and event study with paired controls

[Mike Br]

Dear statalist members,
Assume that I have panel dataset with countries A B C D E. I want to run both a diff in diff and an event study (e.g. using "eventdd") on this dataset with paired controls. For concreteness, lets assume that treated group is composed of A and B and broad control group is C D E. However, I want controls for A to be C and D, and controls for B to be D and E. Can anyone help me how I can implement this estimation? I will be very grateful for any hints!

[Carlo Lazzaro]

Mike:
welcome to this forum.
See -help didregress-.

[Mike Br]

Thank you Carlo very much. I have played with didregress for some time now and I think I cannot figure this out well. One option I am wondering about is whether I should use two group-variables in the group() option to do this? The treatment is introduced at "Country" level, but I want to make two "ControlGroups" (C D and D E), so should I simply specify group(Country ControlGroups) ? I tried it on my dataset but the results are very similar to what I've gotten with group(Country), so I guess this is a wrong attempt.