[R] anova applied to a lme object

[Berta]

Thanks José Rafael, I will try with library(ape) (at the moment I cannot 
load it).

VarCorr gives  the variance estimates for the random effect and the error 
terms. However, what I am looking for is a measure of the explained 
proportion of variance, such as it is R2 in regression models, and more 
precisely, I am looking for a measure of the explained proprotion of 
variance of each of the variables considered (continuous variables and other 
with random slope). For example, Snijders and Bosker (2003) pg 102 dedicate 
a chapter in their book to  "how much does the multilevel model explain" 
(chapter 7) and derive formulaes for R_1 and R_2  (variance in the first and 
second level respectively). Things seem to get complicated when a slope 
random effect is included in the model, as in my case.  It seems that 
package HLM provides the necessary estimates.

I will have a look at library(ape), thanks for the suggestion.

The book I mention is: Snijders, TAB and Bosker RJ (2003). Multilevel 
Analysis. An introduction to basic and advanced multilevel modeling. SAGE, 
London.

Berta




----- Original Message ----- 
From: "José Rafael Ferrer Paris" <jr_frrr at yahoo.de>
To: "Berta" <ibanez at bioef.org>
Cc: <r-help at stat.math.ethz.ch>
Sent: Wednesday, March 07, 2007 5:16 PM
Subject: Re: [R] anova applied to a lme object


> The variances of the random effects and the residual variances are given
> by the summary function. Maybe VarCorr or varcomp gives you the answer
> you are looking for:
>
> library(nlme)
> library(ape)
> ?VarCorr
> ?ape
>
> JR
> El mié, 07-03-2007 a las 13:09 +0100, Berta escribió:
>> Hi R-users,
>>
>> when carrying out a multiple regression, say lm(y~x1+x2), we can use an
>> anova of the regression with summary.aov(lm(y~x1+x2)), and afterwards
>> evaluate the relative contribution of each variable using the global Sum 
>> of
>> Sq of the regression and the Sum of Sq of the simple regression y~x1.
>>
>> Now I would like to incorporate a random effect in the model, as some 
>> data
>> correspond to the same region and others not:  mylme<- lme(y~x1+x2, 
>> random=
>> ~1|as.factor(region)). I would like to know, if possible, which is the
>> contribution of each variable to the global variability. Using 
>> anova(mylme)
>> produce an anova table (without the Sum of Sq column), but I am not sure 
>> how
>> can I derive the contribution of each variable from it, or even whether 
>> it
>> is nonsense to try, nor can I derive a measure of how much variability is
>> left unexplained.
>>
>> Sorry for the type of question, but I did not find a simple solution and
>> some researchers I work with love to have relative contributions to 
>> global
>> variability.
>>
>> Thanks a lot in advance,
>>
>> Berta
>>
>>
>>
>> >
>>
>> ______________________________________________
>> 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
>> and provide commented, minimal, self-contained, reproducible code.
> -- 
> Dipl.-Biol. JR Ferrer Paris
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~
> Laboratorio de Biología de Organismos --- Centro de Ecología
> Instituto Venezolano de Investigaciones Científicas (IVIC)
> Apdo. 21827, Caracas 1020-A
> República Bolivariana de Venezuela
>
> Tel: (+58-212) 504-1452
> Fax: (+58-212) 504-1088
>
> email: jferrer at ivic.ve
> clave-gpg: 2C260A95
>
>
>
> ___________________________________________________________
> Telefonate ohne weitere Kosten vom PC zum PC: http://messenger.yahoo.de
>
>
>
>