Hi,

I have a panel dataset and I intend to run a fixed-effect model. I am having issues with one variable, education, that should be time-invariant but is not. It is very close to being time-invariant but due to very few observations (0.07%), it is deemed as time-varying. I checked this through the command 'xttrans' How do I deal with such an issue? Do I drop those individuals or just go with education being time-varying.

Thanks for your help


Aarushi