Does CSDID store weight (dripw)?

[Alex Weng]

Hi everyone,

As we know that it's a tradition to show covariates balance before and after weighting.

I am using CSDID with dripw option to balance my treatment and control groups and I wanna show that dripw improves covariates balance. Does anyone know if CSDID stores weight? or if there are any other ways to show that dripw improves covariates balance?

Thank you so much,
Alex


For example, if we are using the CSDID example:

use https://friosavila.github.io/playing...rdid/mpdta.dta, clear
csdid lemp lpop , ivar(countyreal) time(year) gvar(first_treat) method(dripw)

tab first_treat
tab year

*here we generate treat and control groups to show balance for the 2004 treated group before weighting:

gen treat2004=1 if first_treat==2004
replace treat2004=0 if first_treat==0
ttest lpop if year<2004,by(treat2004)

*How do I show that the difference in lpop between treated and control groups improve after using dripw?

[FernandoRios]

No ir doesn’t
The reason is that the same control can receive different weights when compared to different cohorts and even different times
when you use notyet as control is even more complicated
at best you can do the exercise for each attgt using drdid and go from there
( it stores the logit info in e())

[Alex Weng]

thank you!