Closed mingyuqi0914 closed 2 years ago
Hi Mingyu,
Thank you very much for your interest in our package! I recently updated the package so that e(V_iw) also returns the the estimated covariance between the coefficient estimates for time_p0, time_p, … . So “lincom (time_p0+ time_p1+ time_p2+ time_p3+ time_p4)/5.” should work for you to get the averaged effect once you update the package (via GitHub)! Take a look at “Examples / Aggregating event study estimates” in the help file of the updated package (also explained below)
Since “lincom” looks for coefficients and variance covariance matrix stored in e(b) and e(V), a workaround is that after you estimate the coefficients on time_p0, time_p1, time_p2, time_p3, and time_p4, try the following commands:
. matrix b = e(b_iw) . matrix V = e(V_iw) . ereturn post b V . lincom (time_p0+ time_p1+ time_p2+ time_p3+ time_p4)/5
It seems "ereturn post b V” can replace e(b) and e(V). The above code works on the example I included in the helpful (after updating the package). But please let me know if you run into any bug!
Thanks, Liyang
Hi Liyang,
This is very helpful! Thank you very much for your explanation!
Best, Mingyu
Hi lsun20,
I wonder if I can generate an aggregated average treatment effect after fitted eventstudyinteract? Something like option estat simple in CSDID?
Thank you! Mingyu