Re: st: is gllamm appropriate? is it necessary?

[Michael Norman Mitchell <Michael.Norman.Mitchell@gmail.com>]

Dear Jessi

  You have articulated your problem very well, but I think that we 
would need to know just a little bit more to give you a good answer.

  1. In a couple of sentences, what is your research question.
  2. What role does "county" play? (Is it just a control variable? Is 
it a clustering variable? Is it a potential variable you want to 
interact with the other variables?)
  3. How do you want to treat the "cause of death" variable. Do you 
want to model all 9 levels simultaneously? If so, each "cause" would be 
modeled against a "reference cause". Or do you want to model each cause 
of death against all others (e.g. cause 1 vs. all others. cause 2 vs. 
all others, etc.)

  I think that, based on this information, people might have thoughts 
on the statistical model that would answer your question, which then 
would indicate what commands would be appropriate for answering your 
research question(s).

Michael N. Mitchell
See the Stata tidbit of the week at...
http://www.MichaelNormanMitchell.com

On 2010-03-13 11.23 AM, Jessica Bishop-Royse wrote:
> Hello Statalisters!  Hope you're having a wonderful weekend.
>
> I have a small question regarding multilevel modeling.
>
> I am working with a dataset with over 300,000 observations for two years of
> data.  I am trying to model a 9 category nominal dependent variable (cause
> of death), my independent variables (there are 14, which are continuous and
> dichotmous) as well as 6 county level indicators.
>
> My question this:  is this best done with gllamm?  Does anyone have any
> other suggestions for a more appropriate model?  Or perhaps a suggestion as
> to how to speed this up?
>
> Any help or suggestions are much appreciated!
>
> Thanks so much, in advance.
>
> Jessi Bishop-Royse
>
>    
*
*   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/