st: trinomial confidence interval

[Mike Lacy <Michael.Lacy@colostate.edu>]

> Date: Fri, 10 Dec 2004 12:52:00 -0000
> From: "Paschalides,C" <C.Paschalides@lse.ac.uk>
> Subject: st: trinomial confidence interval
>
> Dear Statalisters: I have a trinomial distributions with possible events 
> A, B and C. I have obtained obtained results from a number of samples (or 
> if this complicates the situation much more, just consider one sample); I 
> have the marginal confidence interval for the proportion for each of the 
> events but I would like a joint confidence interval, say 95%, for the 
> actual proportions and plot the interval.
I happened to encounter this problem recently when a colleague wanted 
simultaneous c.i.s for a serious of multinomial response variables.  My 
idea was to run
-mlogit y-, i.e., a no covariate multinominal response model with y as the 
response.  Then, use
-predict se1, stdp outcome(1)- to get the standard error of
the predicted response probability for the first category, second, 
etc.  While this seemed reasonable to me (at least asymptotically), the 
standard errors I got were wildly different (order of magnitude) from what 
got calculating conventional binomial st. err. for each outcome. Any 
thoughts what might be wrong (or right <g>) here?

Regards,

=-=-=-=-=-=-=-=-=-=-=-=-=
Mike Lacy
Fort Collins CO USA
(970) 491-6721 office





*
*   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/