search

lsun20 / EventStudyInteract

MIT License
34 stars 12 forks source link

Coefficient matrix is unlabeled #1

Closed NickCH-K closed 3 years ago

NickCH-K commented 3 years ago

The package seems to work great! And the documentation is clearer than aggregate.fixest in the R fixest package.

One note: the e(b) and e(V) matrices that are returned do not have the proper coefficient names for the relative-time indicators. It's there for the covariates, but not for the various g_s (as per the documentation). This makes it pretty tricky to pull results out to make, say, a graph after the fact.

This should be fixable with a matrix colnames or matrix rownames before the ereturn.

lsun20 commented 3 years ago

Thank you very much for your interest! I am aware of the missing labels. I tried to label e(b) and e(V) initially but decided not to because of version difference in reghdfe that may or may not report the constant. Instead, I extract the relevant estimates in e(b_interact) and just updated the code to extract e(V_interact) as well. Sorry about the confusion from the documentation! The documentation is updated, also with an example of coefplot.