The confint() method allows for simultaneous confidence intervals via a multiplier bootstrap when setting joint = True. It would be nice to be able to be able to plot simultaneous confidence intervals.
To Do
Store an object self._conf_int_simult in the Feols / DiD classes that is a empty np.array by default
Add an argument joint to the following methods: Feols.coefplot, Feols.iplot, TWFE.iplot, DID2S.iplot.
If joint = True, use simultaneous CIs for plotting the confidence intervals.
If simultaneous CIs are not computed, run confint(..., joint = True) and assign the result to self._conf_int_simult
Currently, I do not think that we can compute joint CIs for the LPDID.iplot, we'd have to compute a vcovSUR variance covariance.matrix of the stacked regressions #428.
Context
The
confint()
method allows for simultaneous confidence intervals via a multiplier bootstrap when settingjoint = True
. It would be nice to be able to be able to plot simultaneous confidence intervals.To Do
self._conf_int_simult
in the Feols / DiD classes that is a empty np.array by defaultjoint
to the following methods:Feols.coefplot
,Feols.iplot
,TWFE.iplot
,DID2S.iplot
.joint = True
, use simultaneous CIs for plotting the confidence intervals.confint(..., joint = True)
and assign the result toself._conf_int_simult
LPDID.iplot
, we'd have to compute a vcovSUR variance covariance.matrix of the stacked regressions #428.