[R] Another Mix Model Question

[Spencer Graves]

	  To reinforce, Dimitris' excellent suggestion, I'd like to refer you 
to sect. 2.4 in Pinheiro and Bates (2000) Mixed-Effects Models in S and 
S-Plus (Springer, pp. 92-96).  Testing for parameters on a boundary of a 
parameter space, as, e.g., whether a variance component is zero, 
involves a violation of the assumptions for the stanadard asymptotic 
theory.  In this section, Pinheiro and Bates compare the output of 
simulate.mle with improved theory.  I highly recommend this section (and 
the book more generally).

	  spencer graves

Dimitris Rizopoulos wrote:

> AFAIK the COVTEST option just computes a Wald test! Since you want 
> test for variance components (which is on the boundary of the 
> parameter space), I'd suggest to use a LRT (i.e., anova.lme(model.1, 
> model.2)) and moreover consider the simulate.lme() function of the 
> "nlme" package.
> 
> 
> Best,
> Dimitris
> 
> ----
> Dimitris Rizopoulos
> Ph.D. Student
> Biostatistical Centre
> School of Public Health
> Catholic University of Leuven
> 
> Address: Kapucijnenvoer 35, Leuven, Belgium
> Tel: +32/16/336899
> Fax: +32/16/337015
> Web: http://www.med.kuleuven.be/biostat/
>      http://www.student.kuleuven.ac.be/~m0390867/dimitris.htm
> 
> 
> ----- Original Message ----- 
> From: "Alfonso M Sanchez-Lafuente" <alfonso at slafuente.net>
> To: <r-help at stat.math.ethz.ch>
> Sent: Tuesday, June 21, 2005 9:57 AM
> Subject: [R] Another Mix Model Question
> 
> 
> 
>>Hi again,
>>
>>thank you for your previous answers. Just another question, though 
>>...
>>
>>I get the following variance components after fitting a mixed model.
>>
>>
>>Groups   Name              Variance Std.Dev. Corr
>>PlantID  TreatmCtrl        0.51784  0.71961
>>         TreatmNoAccess    4.77469  2.18511  -0.063
>>         TreatmNoKeel      4.22726  2.05603   0.513  0.751
>>         TreatmNoSpur      0.45918  0.67763   0.158  0.303  0.319
>>         TreatmNoStand     3.45357  1.85838  -0.736 -0.070 -0.435 
>>0.495
>>PlantID  PollClassApis     1.12364  1.06002
>>         PollClassBombAnth 0.42769  0.65398  -0.759
>>Residual                   3.09669  1.75974
>>
>>My question is: if n random effects are included in a model, how can 
>>I
>>test the hypothesis that the variance of such effects is 0 ?
>>
>>Some sort of COVTEST option in Proc MIXED in SAS (sorry, SAS is 
>>still
>>more familar to me than R).
>>
>>-- 
>>
>>----------------------------------------------
>>Alfonso M. Sanchez-Lafuente
>>Departamento de Biologia Vegetal y Ecologia
>>Facultad de Biologia
>>Universidad de Sevilla
>>Avd. Reina Mercedes 9
>>E-41012, Sevilla, Spain
>>email: alfonso at slafuente.net / slafuente at us.es
>>
>>______________________________________________
>>R-help at stat.math.ethz.ch mailing list
>>https://stat.ethz.ch/mailman/listinfo/r-help
>>PLEASE do read the posting guide! 
>>http://www.R-project.org/posting-guide.html
>>
> 
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html

-- 
Spencer Graves, PhD
Senior Development Engineer
PDF Solutions, Inc.
333 West San Carlos Street Suite 700
San Jose, CA 95110, USA

spencer.graves at pdf.com
www.pdf.com <http://www.pdf.com>
Tel:  408-938-4420
Fax: 408-280-7915