xtmixed vs mixed

[paulvonhippel]

We are rerunning old code that uses the xtmixed command. Should we replace xtmixed with mixed, or are they the same now?

Another question: I've heard conflicting information about whether xtmixed or mixed can be used with cluster-robust standard errors. I see vce(robust) and vce(cluster) in the help for xtmixed and mixed, and we don't get an error message when we run those options. But the results don't always look as I expect, and as I said I've heard rumors....

When we use xmixed ..., vce(cluster...), we are not using random effects, so one thing we could do is see if we get the same results with xtreg..., vce(cluster...)

[Stephen Jenkins]

What is the "conflicting information" and what are the "rumors" you've heard? It's hard to assess the substance of such remarks without concrete evidence put before us. Also it would be helpful to state precisely what version you're using for your "old code".

In version 14, help mixed returns, in part:

    SE/Robust
      vce(vcetype)           vcetype may be oim, robust, or cluster clustvar; types other
                               than oim may not be combined with dfmethod()

In version 14, and in versions 13.1 and 12.1, help xtmixed returns, in part:

    SE/Robust
      vce(vcetype)           vcetype may be oim, robust, or cluster clustvar

xtmixed was available in version 11, but the help file does not refer to a SE/Robust option.

The help xtmixed for version 13.1 states in part:

    xtmixed has been renamed to mixed.  xtmixed continues to work but, as of Stata 13, is no longer an  official part of Stata.  This is the original help file, which we will no longer update, so some links  may no longer work.

You can look at the Stata 14 Mixed Effects manual online at http://www.stata.com/manuals14/me.pdf