search

xuyiqing / fect

fixed effect counterfactual estimators
Other
53 stars 20 forks source link

Clarifying the relationship between gsynth and fect #8

Closed bfifield closed 1 year ago

bfifield commented 3 years ago

Thanks very much for creating this suite of tools for panel causal inference - it's been very helpful for our own analyses already. Could you clarify more the relationship between the IFE and MC models in fect, versus the IFE and MC models implemented in gsynth? It's tough to tease out the differences between these implementations, and I'd rather rely on fect if they're equivalent to the gsynth implementations given the additional diagnostic tools built into fect. Thank you!

xuyiqing commented 3 years ago

Hi Ben,

They're not that different. In fact, we plan to merge them in the near future. FEct can accommodate treatment reversal while gsynth cannot. In the case when there's no reversal, FEct is equivalent to gsynth with EM = TRUE (so a bit slower). MC in the two packages are the same.

On Fri, Dec 11, 2020 at 9:41 AM Ben Fifield notifications@github.com wrote:

Thanks very much for creating this suite of tools for panel causal inference - it's been very helpful for our own analyses already. Could you clarify more the relationship between the IFE and MC models in fect, versus the IFE and MC models implemented in gsynth? It's tough to tease out the differences between these implementations, and I'd rather rely on fect if they're equivalent to the gsynth implementations given the additional diagnostic tools built into fect. Thank you!

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/xuyiqing/fect/issues/8, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB2PKGAJVXXM7LWFWXGOO6TSUJKV7ANCNFSM4UW7OCTA .

-- Yiqing Xu

Assistant Professor Department of Political Science Stanford University http://yiqingxu.org/

bfifield commented 3 years ago

Perfect - thanks, Yiqing! This is really helpful.