Robust standard errors hausman test

[Oliver Gatland]

Hi all,

Im encountering a problem when trying to run a hausman test with robust standard errors. Ive run the random effects regression then used xtoverid but stata returns an error code reading: 2000b: operator invalid.

I was wondering if anyone could tell me where im going wrong?

Thanks!

[Carlo Lazzaro]

Oliver:
you did nothing wrong, but overlooking that -xtoverid-, being a bit old-fashioned, does not allow -fvvarlist- notation.
Just prefix your regression code with -xi:- and see if the error message creeps up again.

[Oliver Gatland]

Thanks for your response Carlo,

Now when I try stata isn't even recognising xtoverid! If I run a hausman test without using the regressions with robust standard errors and use those results would that be ok or is that something completely different?

[Carlo Lazzaro]

It depends, but I would not consider reliable going -hausman- and then invoking non-default standard errors.
Try to track down what's the matter with -xtoverid-, instead.