You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for setting up and maintaining this package!
I recently read Roodman et al. (2019) and noticed that the Stata version of boottest() includes an option to return the numerators of the bootstrapped test statistics using svmat(numer). These numerators represent the coefficient estimates from all bootstrap replications under the null hypothesis β = 0.
It appears that currently this isn't available in fwildclusterboot::boottest(), but would be a useful option. Would it be possible to consider including this or a similar feature in a future update?
Thanks again for your great work on this package!
The text was updated successfully, but these errors were encountered:
Sorry for the confusion. I'm referring to the numerator of the bootstrapped t-statistics—not the final t_boot values. Under the null hypothesis that the coefficient is 0, the numerator corresponds to the bootstrapped coefficient estimates, which is exactly what I'm looking to retrieve.
These are computed in this function as "numer". This function is then called via fast_n_reliable implementation, which itself is called by run_bootstrap. And last, run_bootstrap is the function called by the estimation APIs boottest.lm, boottest.fixest etc.
So should be doable!
Btw, in the "fast and reliable" algo from the newer MacKinnon, Nielsen, Webb paper, the numerator is computed here: link
Hi,
Thanks for setting up and maintaining this package!
I recently read Roodman et al. (2019) and noticed that the Stata version of boottest() includes an option to return the numerators of the bootstrapped test statistics using svmat(numer). These numerators represent the coefficient estimates from all bootstrap replications under the null hypothesis β = 0.
It appears that currently this isn't available in fwildclusterboot::boottest(), but would be a useful option. Would it be possible to consider including this or a similar feature in a future update?
Thanks again for your great work on this package!
The text was updated successfully, but these errors were encountered: