Alexander Fischer

> You mean this PR? Or the pyfixest project itself? I was referring to pyfixest itself - so far, we haven't implemented a range of statistics, e.g. F-tests and adjusted...

Hi @toobaz, sorry for never really getting back to you about this! I had a thousand things going on plus was on vacation... Sorry! From my perspective, this PR is...

> Hi @s3alfisc , thanks for the update! I wasn't reviewing the PR because it is a "draft" Ouhh, yes, just noticed this now, sorry about that! > I understand...

> ... did you happen to test mixing pyfixest and other models? This of course fails... for statsmodels, I get an > AttributeError: 'OLSResults' object has no attribute '_fixef' error....

Ok, sold! So the next steps would be that - I'll remove the`pf.Stargazer` class from `pyfixest` & update the vignette (i.e. prompt users to use `from stargazer.stargazer import Stargazer` -...

I would definitely try and volunteer testing the experimental feature on a pyfixest branch if this would make it into the next formulaic release!

Will spend some time to test it this weekend. Thanks @matthewwardrop & @MarcoGorelli !

Relatedly: with hetero errors, changing the `cluster.adj` from `TRUE` to `FALSE` leads to a change in standard errors, which I believe should not be the case? Likely related to this...

Here are the dof corrections in comparison to `sandwich`: ```r library(fwildclusterboot) library(fixest) data(voters) head(voters) voters = na.omit(voters) fit = feols( proposition_vote ~ treatment + income, #weights = ~weights, data =...

I went back to our mail exchange from a very long time ago and it turns out that you've already answered my question @lrberge - I just did not get...