unit -invariant variable

[aroopsind]

Hi,

I am building out a model where one of my covariates, elevation, is fixed and does not vary with time as the outcome and some other covariates (e.g. pop?).

I created a long structure data format and for the column with the elevation value it repeats for each year there is an observation. I, however, get the following error:
"Error in gsynth.default(formula = NULL, data = data, Y = Yname, D = Dname, : Variable "elevation" is unit-invariant. Try to remove it. Timing stopped at: 10.85 0.32 13.47"

Any guidance on the best way to include these time-invariate covariates would be greatly appreciated.

Anand

[xuyiqing]

Hi Anand, Like in fixed effects, gsynth does not take in time-invariant covariates because they have already been "control for" with the additive and interactive fixed effects. Best, Yiqing

…

On Tue, Mar 24, 2020 at 9:38 AM Anand Roopsind ***@***.***> wrote: Hi, I am building out a model where one of my covariates, elevation, is fixed and does not vary with time as the outcome and some other covariates (e.g. pop?). I created a long structure data format and for the column with the elevation value it repeats for each year there is an observation. I, however, get the following error: "Error in gsynth.default(formula = NULL, data = data, Y = Yname, D = Dname, : Variable "elevation" is unit-invariant. Try to remove it. Timing stopped at: 10.85 0.32 13.47" Any guidance on the best way to include these time-invariate covariates would be greatly appreciated. Anand — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#40>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB2PKGC4DMFGL6K2OG4HXV3RJDO2BANCNFSM4LSZ7LEQ> .

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

[aroopsind]

Thanks for that clarification Yiping. I went back through the gsynth publication and supplemental - is there an example of the use of the additive and interactive fixed effects model to control for these time-invariant covariates and then the application of gsynth.

In one analysis I am exploring with gsynth, as a first step I used propensity score matching to identify the pool of control units based on time-invariant covariates. Curious how different these would be based on the IFE. Thanks again for taking the time to respond.

Cheers
Anand

[xuyiqing]

Trimming based on time-invariant covariates before running gsynth/IFE is certainly allowed and may improve overlap and speed. I think it's a good idea if you have a lot of data.

…

On Sun, Apr 5, 2020 at 10:05 AM Anand Roopsind ***@***.***> wrote: Thanks for that clarification Yiping. I went back through the gsynth publication and supplemental - is there an example of the use of the additive and interactive fixed effects model to control for these time-invariant covariates and then the application of gsynth. In one analysis I am exploring with gsynth, as a first step I used propensity score matching to identify the pool of control units based on time-invariant covariates. Curious how different these would be based on the IFE. Thanks again for taking the time to respond. Cheers Anand — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#40 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB2PKGHVQK4EA5QVI56W4XDRLC24DANCNFSM4LSZ7LEQ> .

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

[jtorcasso]

Hello, I was wondering if the package allowed for including observed time-invariant covariates, as discussed in Remark 3 of the paper.

[xuyiqing]

They are not compatible with gsynth, however, another package we worked on, bpCausal, can incorporate them.

…

On Thu, Dec 16, 2021 at 10:36 AM jtorcasso ***@***.***> wrote: Hello, I was wondering if the package allowed for including observed time-invariant covariates, as discussed in Remark 3 of the paper. — Reply to this email directly, view it on GitHub <#40 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB2PKGFA7HTHWDUURZF2FQDURIWUNANCNFSM4LSZ7LEQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>. You are receiving this because you commented.Message ID: ***@***.***>

-- Yiqing Xu Assistant Professor Department of Political Science Stanford University https://yiqingxu.org/

[jtorcasso]

Thanks! Will check it out.