st: Post Hoc Cross Tabulation Tests

[Christopher Salas-Wright <christopher.wright.2@bc.edu>]

I am running an analysis examining the difference in prevalence for
dichotomous (yes/no) substance abuse variables across a nominal
variable with 5 categories in a nationally representative dataset.

My goal is to identify significant differences overall and to identify
significant differences between nominal categories (e.g. substance
abuse rates are significantly different between White and African
American youth, but not African American and Hispanic).  If my
substance abuse variables were continuous then I would use an ANOVA
and use multiple comparison tests to identify category differences
with syntax like:

“oneway dv iv, t bon sid sch”

Currently, however, I am using the following syntax to test for
overall differences:

“svy: tab substance iv, column ci per”

That gives me point estimates, confidence intervals, chi2, and an F
statistic.  All of these are helpful, but is there a simple way that I
can do the equivalent of the post hoc ANOVA comparison tests with my
cross- tabulation?  Or is there another way of testing overall
difference and individual category differences using my dichotomous
and nominal variables?

*
*   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/