Michael Ingre, who is accustomed to working with SPSS, posted about having
trouble setting up a repeated-measures ANOVA in Stata.
Michael's description of the experimental design didn't explicitly say so,
but the behavior of Stata's -anova- upon switching the variable order of the
subjects and groups factors in the command line indicates that subjects are
nested within groups.
If that's so, then the model specification would be groups,
subjects-within-groups as the error (denominator) mean square for groups,
then time, and finally the interaction of groups and time:
-anova dv grp / sub|grp time grp*time-.
There is a description of the syntax (including what forward slashes, pipes
and stars are for) in the user's manual for the -anova- entry. There's also
a FAQ about repeated-measures anova on StataCorp's website,
www.stata.com/support/faqs/stat/anova2.html . I think that both the user's
manual and the FAQ have examples that use datasets analogous to Michael's.
It appears that Michael's dataset has at least one missing value. I think
that this topic is broached in the user's manual in the entry for -anova-.
If there is a concern over the imbalance, there are alternatives to consider
in this situation, such as -xtreg , re-.
Joseph Coveney
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
