search

synth-inference / synthdid

Synthetic difference in differences
https://synth-inference.github.io/synthdid
BSD 3-Clause "New" or "Revised" License
262 stars 98 forks source link

Examining Treatment Heterogeneity #88

Open ngupt123 opened 2 years ago

ngupt123 commented 2 years ago

If I suspect treatment may be heterogeneous (i.e the treatment may vary between men and women), I'm not sure how I'd examine it here. In normal DiD, I've typically seen a triple interaction being used to examine heterogeneity, but I don't know how to do something similar using SDiD.

davidahirshberg commented 2 years ago

A natural way to do this would be to do your analysis conditional on the feature you expect heterogeneity in. In your men/women example, estimate a treatment effect using SDiD separately on the men and women in your dataset. If you buy the independence-between-units assumption our standard errors are based on, the standard error for the difference should be [standard-error-for-men^2 + standard-error-for-women^2]^1/2. On Apr 19, 2022, 8:36 PM -0400, ngupt123 @.***>, wrote:

If I suspect treatment may be heterogeneous (i.e the treatment may vary between men and women), I'm not sure how I'd examine it here. In normal DiD, I've typically seen a triple interaction being used to examine heterogeneity, but I don't know how to do something similar using SDiD.

— Reply to this email directly, view it on GitHubhttps://github.com/synth-inference/synthdid/issues/88, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AANV7XDO3EFCNBVB6QZVNKDVF5GP7ANCNFSM5T2N7RXA. You are receiving this because you are subscribed to this thread.Message ID: @.***>