st: RE: RE: allpossible & xtmixed

["Feiveson, Alan H. (JSC-SK311)" <alan.h.feiveson@nasa.gov>]

It's even worse than that with -xtmixed. Should this list be just for all possible combinations of variables in the fixed part only, or should it also include all possible predictors in each level of the random part, with the various possible covariance structures? (Of course most of these would never converge). In many applications of -xtmixed- one would want to consider random slopes for at least some of the predictors in the fixed part.


Al Feiveson

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Nick Cox
Sent: Friday, May 20, 2011 8:37 AM
To: 'statalist@hsphsun2.harvard.edu'
Subject: st: RE: allpossible & xtmixed

The short answer is No. 

Here is a longer answer. 

1. Please remember to explain where user-written programs you refer to come from. Why is this apparently so difficult to understand and to implement? Not paying attention makes your questions unnecessarily difficult for others to understand. You oblige people, if your question sounds interesting to them, to do some work to find out what on earth -allpossible- is. It is much easier just to explain and make your question more friendly. 

Please remember: Statalist is a discussion list, not a help line. Postings that are hard to understand are antisocial. 

2. The help for -allpossible- (SSC) starts by explaining 

"-allpossible- by default (1) computes all possible models fitted by modelcmd to response depvar and subsets of up to 6 predictors from varlist and (2) tabulates various
summary statistics for each model fitted.  Alternatively, (1') the maximum number of predictors fitted may be specified as a number less than 6 by the npmax() option."

As 26 > 6, there is no scope for do what you want with -allpossible-. Whether -allpossible- understands -xtmixed- is immaterial. 

3. -tuples- from SSC gives you ways of cycling over subsets. 

4. With 26 predictors, you have 2^26 - 1 = 67108863 possible models, just on the basis of whether predictors are in or out (without reference to interactions). The combinatorics here are against you. 

Nick 
n.j.cox@durham.ac.uk 

Dennis Kramer

Does --allpossible-- support --xtmixed--? I am trying to run an
xtmixed model with the possible combination of 26 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/

*
*   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/