Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: Alternative to using -suest- with -xtmixed-


From   "Ploutz-Snyder, Robert (JSC-SK)[USRA]" <[email protected]>
To   "[email protected]" <[email protected]>
Subject   st: Alternative to using -suest- with -xtmixed-
Date   Tue, 29 Mar 2011 16:01:41 -0500

Hello Statalisters..

I'm running  Stata (IC) 11.1 on a Windows 7 machine in 64 bit processing.
 

I have the need to get an "omnibus" test for beta coefficients across three different -xtmixed- models.  I know that -suest- won't work because -predict- following -xtmixed- does not allow the -store- option. I've seen a couple of posts on the statalist archives about this but I hadn't seen a resolution posted that would work for my situation.

In a nut shell, the three models I want to run are:

xtmixed y1 covariate1 c.a##i.session ||id:
xtmixed y2 covariate2 c.a##i.session ||id:
xtmixed y3 covariate3 c.a##i.session ||id:

Each model has a common interaction term, and each model has a co-variate that is unique to each model.  One of the terms in the interaction ("a") is continuously scaled; the second term ("session") is a factor variable with three levels.

I want to get a test for the interaction term(s) ACROSS dependant variables.  Since it is not currently possible to use  -suest- with -xtmixed-, is it legitimate to run these three models using -reg- without accounting for the clustering (because -suest- needs it that way), store each model's estimates, then run -suest- WITH THE -cluster- option?


Thanks in advance for your thoughts.


p.s. I can't use -sureg- in this instance because my data are longitudinal ("session" has three levels representing three repeated time periods when data were collected), and -sureg- does not allow vce(cluster id) and doesn't allow  -robust- to adjust the SE's either. It would be great if Stata could tweak -sureg- so that it could incorporate clustering, or better yet, develop an -xtsureg- if that's possible?  

p.p.s. It would be nice if Stata could tweak -xtmixed_postestimate- so that it could accommodate -suest-  too!!  



*
*   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/


© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index