[prev in list] [next in list] [prev in thread] [next in thread] 

List:       r-sig-mixed-models
Subject:    Re: [R-sig-ME] How can parameter estimates be extracted from a LMEM
From:       Christopher Philipson <christopher.philipson () ieu ! uzh ! ch>
Date:       2011-07-17 10:31:33
Message-ID: B20666DD-B5B1-47ED-9C5E-00040DAAF458 () ieu ! uzh ! ch
[Download RAW message or body]

Hi Karen,

I understand that you can use the intervals function 

for example

intervals(model1, which="fixed")

this will still provide differences so you would have to relevel your factors, and re \
run the model to get all the parameter estimates.

If this is incorrect (anyone?) then feel free to correct me.

Cheers
Chris


On 15 Jul 2011, at 23:00, Lisa Stevens wrote:

> Dear All,
> 
> I was wondering if anyone knows a way to reliably extract parameter
> estimates from a LMEM where VarIdent and corAR1 are necessary.
> 
> Here is the full model I am currently working with:
> 
> model1<-lme(asin(max_cover/100)~1+provenance*lifeform*year_trt*N*P*K,method="ML",random=~1|site/plot,weights=varIdent(form=~1|lifeform),
>  na.action=na.exclude,
> correlation=corAR1(form=~as.numeric(year_trt)|site/plot), data=dat1)
> 
> The two lifeforms I am working with are graminoids and forbs, and the
> experiment is a full factorial fertilization experiment that has been
> conducted at 26 sites. Max_cover represents 16 species and their response to
> fertilization. Because I am hypotheses testing, I have simplified the model
> using LRT and I have found a significant four-way interaction and these
> trends follow the results shown in the raw data. I used REML and LRT to
> test  VarIdent and the autocorrelation with treatment year and both are
> necessary.
> 
> Is there a method for calculating parameter estimates from such a complex
> model for a designed experiment? If there isn't, does anyone have any
> suggestions on methods or approaches I could use to get around this issue?
> 
> Thank you very much for any help or advice in advance.
> 
> Best wishes,
> 
> Karen
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-sig-mixed-models@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models

Dr. Christopher Philipson
Institute of Evolutionary Biology and Environmental Studies,
Winterthuerstrasse 190,
CH-8057,
Zurich
Tel:  00 41 44 635 61 29
Fax: 00 41 44 635 57 11
christopher.philipson@ieu.uzh.ch

_______________________________________________
R-sig-mixed-models@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models


[prev in list] [next in list] [prev in thread] [next in thread]