st: Re: test for unordered multinomial responce of mathed paired samples

["Joseph Coveney" <jcoveney@bigplanet.com>]

Shoryoku Hino wrote:
 
This may be a beginner's question. 
I am struggling with a matched-pair data. 
I would like to ask you which test is appropriate when a response could take
one of three or more categories(not ordinal) and 
how to do it using stata. 
If a response variable were dichotomous, I know that McNemar test would be
appropriate for matched-pair data. 
Creating dummy variables followed by McNemar test may be one of the
solutions. But I think there may be better ways.

--------------------------------------------------------------------------------

If your response is multinomial and not ordered categorical, then consider
either -symmetry-, or modeling the data using -mlogit-.  If neither of these is
appealing, then try typing

    findit multinomial

and seeing whether something among the results fits your needs.

Joseph Coveney


*
*   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/