Re: st: Why does Stata drop an additional category?

[Richard Williams <Richard.A.Williams.5@nd.edu>]

At 09:46 AM 2/17/2004 -0500, Sarah A. Mustillo wrote:
> I think I am going to have to disagree with Rich on this one and say that 
> perhaps the additional category is being dropped because of the 
> interaction term.
I think Sarah is correct in saying I originally got it wrong. :)  But, one 
other thing I will note with regards to possible discrepancies between SPSS 
and Stata:  In Stata, the command


> xi: logistic unemployed  i.religion i.educ5 i.cath*e5c
results in both e5c and _IcatXe5c_1 being included in the model, even 
though e5c is not specified separately, i.e. Stata is including the lower 
order terms.  But, in SPSS, a similar notation would result in only the 
interaction term being included and not the main effect of e5c.  So, that 
may explain why the two programs seem to be behaving differently.  So, my 
latest guess is that if you instead did

gen cathe5c = cath*e5c

and used that instead of i.cath*.e5c, spss and Stata would say the same 
things in more or less the same way.  If the original questionner wants to 
email me his spss output I'll take a look at it.  Of course, why there is 
both an educ5 (treated as categorical) and an ec5 (treated as continuous) 
is still unclear to me.

*
*   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/