thanks so much for this great package! I have one small question. It seems to me that standard errors are currently not clustered when computing pre-trend coefficients. Or am I missing something?
Basically, should lines 282 onwards in did_imputation look something like this (with the cluster argument)?
kylebutts
commented
4 months ago
Hi Kyle,
thanks so much for this great package! I have one small question. It seems to me that standard errors are currently not clustered when computing pre-trend coefficients. Or am I missing something?
Basically, should lines 282 onwards in did_imputation look something like this (with the cluster argument)?
pre_est <- fixest::feols( pre_formula, data[data$zz000treat == 0, ], weights = ~zz000weight, warn = FALSE, notes = FALSE, cluster=cluster_var )Thanks again!
Stefan