Alberto Osella <arosella@irccsdebellis.it> used -mi estimate: sqreg- and
received an error:
> mi estimate, noisily dftable: sqreg BMI Etarecl Num_nucleo i.Sesso ///
> i.GrammiAlc_G_C i.NewStatCiv i.Professione1 i.scalacat, q(.47 .83) reps(400)
>
> (running sqreg on m=1)
> factor variables and time-series operators not allowed
> an error occurred when mi estimate executed sqreg on m=1
> r(101);
And Neil Shephard <nshephard@gmail.com> replied:
> You can't use the new way of specifying factor variables with -mi-
> instead you should use the old way of specigyin them. Try...
> xi : mi estimate, noisily dftable: sqreg BMI Etarecl Num_nucleo ///
> i.Sesso i.GrammiAlc_G_C i.NewStatCiv i.Professione1 ///
> i.scalacat, q(.47 .83) reps(400)
(snip)
> So for the time being if you're using -mi- or -sw-/-stepwise- you have
> to use the old way of specifying factor variables.
Alberto received an error because -sqreg- does not support factor variables.
Alberto should not use -xi- instead, though, because -xi- is not supported by
-mi estimate-. (Factor variables actually are supported by -mi estimate-.)
If Alberto would like to use -mi estimate: sqreg- with categorical predictors,
therefore, neither -xi- nor factor-variable syntax is appropriate. Alberto may
instead use -mi passive: generate- to create indicator variables for "Sesso,"
"scalacat", and the other categorical predictors. He may then include the
indicator variables in -mi estimate: sqreg- and omit both -xi- and the
factor-variable operator "i.".
-- Wes
weddings@stata.com
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
