DID only post treatment observations

[Juliana Cunha]

Dear All, Please, could anyone suggest a way to do the following?
I have a sample with firms that appear in the pre-treatment period and in the post-treatment; and there are firms that appear only in the post-treatment period. I use DID and for the parallel trend assumption to hold, I dropped the firms that appear only post treatment. But now, I wanted to run some robustness checks using this subsample of firms that were initially excluded (only have post treatment). How could I do it? Thank you!

[Frode Andre]

You may find your answer in this discussion: https://www.statalist.org/forums/for...alanced-sample

[George Ford]

Absent any pre-treatment values for the excluded units, I can't see how you can do anything in the DID framework.

You might just use the post treatment period and do a means difference between the groups and compare them. You could say the excluded ones have the same spread. But it's a biased measure of the treatment effect, so not sure what that gets you.