we are currently writing custom functions for the sandwich s.e. calculations, that is not needed, should update the calculations to use the sandwich library.
also, after confirming with Judith, for these sandwich estimator calculations, we don't need to use the clustered version, regular version is sufficient, should simplify things quite a bit.
related:
minh brought up an issue where including center effects is set to true, the confidence set calculation gives errors. After investigation, i think it is closely related to the sandwich s.e. calculation while taking into account of the clustering. Right now we are using manual functions for this. After changing to use the sandwich library, make sure this issue is also resolved.
we are currently writing custom functions for the sandwich s.e. calculations, that is not needed, should update the calculations to use the sandwich library.
also, after confirming with Judith, for these sandwich estimator calculations, we don't need to use the clustered version, regular version is sufficient, should simplify things quite a bit.
related:
minh brought up an issue where including center effects is set to true, the confidence set calculation gives errors. After investigation, i think it is closely related to the sandwich s.e. calculation while taking into account of the clustering. Right now we are using manual functions for this. After changing to use the sandwich library, make sure this issue is also resolved.