Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

RE: st: gllamm dropping variables


From   "Bontempo, Daniel E" <deb193@ku.edu>
To   "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu>
Subject   RE: st: gllamm dropping variables
Date   Fri, 22 Mar 2013 14:39:37 +0000

Hi Stan. 

The issue with xtlogit is that it collapses all non zero values into 1 category. I lose the "ordered logistic" aspect, or the information the levels between 0-1 would have contributed in the gllamm model with link(logit) and family(binomial).

In reading Stata Journal about the use of logit with binominal to model fractions (st0147), I understand glm had to be rewritten to allow this. I think gllamm may lack whatever was done to glm, and this causes it to drop everything.


-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Stas Kolenikov
Sent: Thursday, March 21, 2013 10:16 AM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: gllamm dropping variables

Does this lead to a perfect prediction? At any rate, for random effect logistic models, I would use -xtlogit-, as it is faster, and probably more stable, than -gllamm-.

-- Stas Kolenikov, PhD, PStat (SSC)
-- Senior Survey Statistician, Abt SRBI
-- Opinions stated in this email are mine only, and do not reflect the position of my employer
-- http://stas.kolenikov.name



On Thu, Mar 21, 2013 at 10:02 AM, Bontempo, Daniel E <deb193@ku.edu> wrote:
> I have some dummy codes for levels 2 and 3 of a 3-level group variable. With xtmixed, I can use i.group, but with gllamm I have to make dummy codes.
>
> The codes work fine when I run a linear model, but cause the variables 
> to drop when I use link(logit)
>
> This runs fine:
> gllamm regPart ae le, i(number) adapt
>
> This won't start:
> gllamm regPart  ae le, i(number) adapt family(binomial) link(logit) 
> variables have been dropped, can't continue r(198);
>
> I assume the variables are dropped for collinearity, but do notunderstand why.
>
> BTW, using glm seems to have no problem:
> glm regPart ae le, link(logit) family(binomial) cluster(number) eform
>
>
> If I use a continuous IV with gllamm it will also run. Why are my dummy codes being dropped?
>
>
>
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/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/faqs/resources/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/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index