rreg v. robreg

[Tracy Lam]

Hi all,
What is the difference between rreg and robreg? I'd like to conduct analyses that handles outliers using robust regression, but it is not clear to me which is the most accepted command. If it is robreg, what are your suggestions for the estimator? Is there an accepted one?

[Richard Williams]

I've never heard anybody say anything good about rreg. See Nick Cox's scathing critique at

http://www.stata.com/statalist/archi.../msg00416.html

Instead, check out mmregress, described at

http://www.stata-journal.com/sjpdf.h...iclenum=st0173

I would also check out robreg. The help file for it has a long list of references. Online you can find a discussion at

http://www.stata.com/meeting/germany...sug12_jann.pdf

Incidentally, these routines would all be down my list of things to do with outliers. First I would check for coding errors and possible model mis-specification. Outliers may contain useful information; just trying to make them go away can be a bad idea.