Staggered DiD (CSDID interpreatation of pre and post avg)

[Tariq Abdullah]

I got the answer.m but cant delete the post

[Tariq Abdullah]

csdid y, ivar(place) time(year) gvar(policy_year) agg(event)

Difference-in-difference with Multiple Time Periods

                                                           Number of obs = 600
Outcome model  : regression adjustment
Treatment model: none
------------------------------------------------------------------------------
             | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
     Pre_avg |  -2.667258   9.788396    -0.27   0.785    -21.85216    16.51765
    Post_avg |  -112.4219   56.29528    -2.00   0.046    -222.7586   -2.085171

 estat simple
Average Treatment Effect on Treated
------------------------------------------------------------------------------
             | Coefficient  Std. err.      z    P>|z|     [95% conf. interval]
-------------+----------------------------------------------------------------
         ATT |  -85.05845   44.41044    -1.92   0.055    -172.1013    1.984416
------------------------------------------------------------------------------

[FernandoRios]

The Pre and Post Avgs are the Aggregated Event ATTs, for the post and pre periods. They should not be considered as Pre-treatment Test.
They are different from simple, because different weights are used compared to SIMPLE

[Tariq Abdullah]

Thanks for the clarification. But, unfortunately, I'm still confused.

So, for pre-test what command I need to use ?

And, which command I need to use to check if my policy variable is signficantly impacting the outcome variable y or not ?

Is this command enough to prove that my policy has significant impact on my outcome variable y ( post-avg coefficient)

csdid y, ivar(place) time(year) gvar(policy_year) agg(event)

Difference-in-difference with Multiple Time Periods

Number of obs = 600
Outcome model : regression adjustment
Treatment model: none
------------------------------------------------------------------------------
| Coefficient Std. err. z P>|z| [95% conf. interval]
-------------+----------------------------------------------------------------
Pre_avg | -2.667258 9.788396 -0.27 0.785 -21.85216 16.51765
Post_avg | -112.4219 56.29528 -2.00 0.046 -222.7586 -2.085171

[FernandoRios]

For pretrend try estat pretrend
for postreatment yea the difference is statistically different from zero

[Tariq Abdullah]

thank you so much ! Much obliged . Also learnt a ton from this thread!

https://www.statalist.org/forums/for...d-csdid/page30