No estimates returned when other exog vars have missing values

[ozak]

Hi Jeff,

I was trying out your command on Stata 14 and running a naive regression I do not get any results back, just .. It seems the command breaks if the dataset is not identical to the full sample. E.g. an observation with a missing value in some exogenous variable, which usually would be dropped, causes this issue. It seems one has to drop all observations that have any missing values that will not be predicted out of sample. To see this, in your test code include replace weight = . in 1 before line 10 and you won't get results back.

[jshrader]

Hi Ozak,

Thanks for opening this issue. I will check it out and see if a quick fix can be made. Otherwise, it sounds like simply dropping all non-missing values will do the trick.