Statalist


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

Re: st: nonlinear growth curve in xtmixed


From   "josemaria" <[email protected]>
To   <[email protected]>
Subject   Re: st: nonlinear growth curve in xtmixed
Date   Fri, 12 Dec 2008 21:29:27 -0200

Dear Stas and Maarten:
thank you for your kindness. But it seems that neither -xtmelogit- nor -xtmepoisson- nor gllamm do apply, because the response variable is not binary or count. For example, for -xtmelogit- the output is odds ratio. Rabe-Hesketh presents growth curve only in linear and quadratic polynomial forms. I am interested in continuous variables, as kg, or m, or bmi (body mass index), that moves through time according to a logistic curve, but not a curve of probability from 0 to 1. Singer & Willett use an iterative maximum likelihood procedure proposed by Pinheiro & Bates (Mixed-Effects Models in S and S-Plus,Springer, 2000, pages 337-347); the response variable in Pinheiro & Bates example is the trunk circumference (in mm) of trees, and time is day. Each tree is measured seven times.
José Maria
By the way, I am not able to write anything in Mata!

Jose Maria Pacheco de Souza, Professor Titular (aposentado)
Departamento de Epidemiologia/Faculdade de Saude Publica, USP
Av. Dr. Arnaldo, 715
01246-904  -  S. Paulo/SP - Brasil
fones (11)3061-7747; (11)3768-8612
fax (11)3714-2403
www.fsp.usp.br/~jmpsouza
----- Original Message ----- From: "Stas Kolenikov" <[email protected]>
To: <[email protected]>
Sent: Friday, December 12, 2008 6:54 PM
Subject: Re: st: nonlinear growth curve in xtmixed


As of version 10, Stata provides -xtmelogit- and -xtmepoisson- for
logistic and Poisson models, respectively. I believe Singer & Willett
only talk about logistic models there, right? You can say those are
nonlinear models since the mean functions are nonlinear in x's; to me
those are extensions of generalized linear model, and hence there is
at least some linear flavor to them. PROC NLMIXED can fit some crazy
models with direct integration of the random effects (as far as I
know), and full functionality of that is not available in Stata (at
least in the easily available commands; you can write anything you
want in Mata!)

On 12/12/08, josemaria <[email protected]> wrote:
Dear Statalisters:
 Judith D. Singer and John B. Willet, in their book "Applied Longitudinal
Analysis", Oxford University Press, 2003, present, in pages 225-232,
analysis for nonlinear growth curve using the PROC NLMIXED of SAS, and
explain that, at the time of writing the book, Stata did not have yet the
same capability.
 This is also informed at the end of the page: "Stata Textbook Examples
 Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence
 by Judith D. Singer and John B. Willett
 Chapter 6: Modeling Discontinuous and Nonlinear Change",
 site
http://www.ats.ucla.edu/stat/stata/examples/alda/chapter6 :
  "Table 6.6 on page 231.

* Skipped these for now. Stata does not yet do nonlinear mixed models." Is
it possible now to do such analysis in Stata?
 Than you for any orientation.
 José Maria

 Jose Maria Pacheco de Souza, Professor Titular (aposentado)
 Departamento de Epidemiologia/Faculdade de Saude Publica, USP
 Av. Dr. Arnaldo, 715
 01246-904  -  S. Paulo/SP - Brasil
 fones (11)3061-7747; (11)3768-8612
 fax (11)3714-2403
 www.fsp.usp.br/~jmpsouza
 *
 *   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/



--
Stas Kolenikov, also found at http://stas.kolenikov.name
Small print: I use this email account for mailing lists only.

*
*   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/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index