st: Lincom Vs Nlcom to test interaction effects

[Vincenzo Carrieri <vincenzo.carrieri@gmail.com>]

Dear Statalisters,
Thanks to all the persons which have me helped on the post Nlcom Vs
Lincom (thanks in particular to Marteen for his precious comment).
I explain better my situation:
My dependent variable is an ordered variable of satisfaction with 6
ordered categories, thus I estimate an ordered probit model. My
interest is to test whether the effect of height on such variable is
increasing or decreasing significantly with age. Given that even a
polynomial specification of age (up to squared) in the model seem to
be required, I think I have to interact  height with age and age
squared.
 I know that with ordered categories there are many possible effects,
but I think that It could be useful test an interaction effect on the
latent dependent variable. If this is right, how to do this on stata?
Maarten suggest to plot some graphs to have a better idea...which
graph can be useful in this context? If it's age squared which
complicates too much the things, I could remove it from the model,
without loosing too much in the model.

Any hint is appreciated.
Many thanks to everybody,

Vincenzo
*
*   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/