Post Hoc test MANOVA in STATA


I'm trying to analyze data on pupils in a class, questioned in three consequetive years, in which there is a dichotomous variable indicating whether you are a smoker (=1) or not (=0). I ran a MANOVA to test whether the two groups differed from each other on a number of dependent variables (age, gender etc.). However, to check HOW the two groups differ from each other on each dependent variable, by just doing separate ANOVA's, I am not in the multivariate framework anymore. So, I want to test whether the two groups significantly differ on, let's say age, but then while including all the other dependent variables. How can I do this? I know I have to make some matrix and then use the manovatest command, but I do not get it to work.

Furthermore, since there are repeated measures (three years), I would like to say something about the changes over time. My main problem is that I am completely unfamiliar with the matrix commands, so a simple manovatest, test(matname) is not even working for me... Can someone help me?

With kind regards,

Dirk Pelt
Graduate Student at Utrecht University
The Netherlands