I want yearly coefficients of treat with their confidence intervals as well as the year fixed effects. So, I did:
Code:
xi: regress index b2005i.year##treat i.id b2005i.year, vce(cluster id)
Instead, I am trying to use parmby, which is supposed to do all that in one command.
Code:
char year [omit] 5 //setting 2005 as base year parmby "xi: regress index treat i.id i.year, vce(cluster id)", by(year) saving(myparms,replace)
Is this the right way to obtain confidence intervals- upper and lower bound, using parmby?
Because with this command, I am able to generate separate dataset with labels "estimates" and "min95" and "max95". However, the values for min95 and max95 are the same! What I want instead is lower and upper bound of the confidence intervals for the yearly "treat" coefficients that I got.
Moreover, I don't get the year fixed effects in the new dataset.
What is wrong in my parmby command? What is the correct way to get CI's for yearly coefficients of treat as well as year fixed effects?
Thanks!
0 Response to Right way to generate a dataset of estimates of treatment effect by year and their CI's as well as the year FE
Post a Comment