What’s the cleanest, most pythonic way to run a regression only on non-missing data and use clustered standard errors? Imagine I have a Pandas dataframe all_data. Clunky method that works (make a dataframe without missing data): I can make a new dataframe without the missing data, make the model, and fit the model: This feels a bit clunky (esp. when I’m doing it all over the place with different right hand side variables.) And I have to make sure that my stats formula matches the dataframe variables. But is there a way to make it work using the missing argument?