[R] Extracting correlation in a nlme model

41 views

Skip to first unread message

Kenneth Roy Cabrera Torres

unread,

May 20, 2009, 12:31:45 AM5/20/09

to r-help

Hi R users:

Is there a function to obtain the correlation within groups
from this very simple lme model?

> modeloMx1
Linear mixed-effects model fit by REML
Data: barrag
Log-restricted-likelihood: -70.92739
Fixed: fza_tension ~ 1
(Intercept)
90.86667

Random effects:
Formula: ~1 | molde
(Intercept) Residual
StdDev: 2.610052 2.412176

Number of Observations: 30
Number of Groups: 3

I want to obtain \rho = \sigma_b^2 / (\sigma_b^2 + \sigma^2)

I know that I obtain \sigma_b^2 and \sigma^2 with

> VarCorr(modeloMx1)

molde = pdLogChol(1)
Variance StdDev
(Intercept) 6.812374 2.610052
Residual 5.818593 2.412176

But, I want to know if I can obtain
\rho = 6.8123/(6.8123 + 5.8185) = 0.53934 straightforward.

Thank you for you help.

Kenneth

______________________________________________
R-h...@r-project.org 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.

Christine A.

unread,

May 20, 2009, 5:34:59 AM5/20/09

to r-h...@r-project.org

Hello,

for calculation of intraclass correlation coefficient for a simple Random
Intercept model, the function ICC1.lme() in the package "psychometric" can
be applied.

Best regards,
Christine Adrion

-----
Christine Adrion, Dipl.-Stat., MPH

Ludwig-Maximilians-Universitaet Muenchen
IBE – Institut für Medizinische Informations-
verarbeitung, Biometrie und Epidemiologie
Marchioninistr. 15
81377 Muenchen
GERMANY

Tel.: +49 (0)89 7095-4483
eMail: adr...@ibe.med.uni-muenchen.de
Internet: http://ibe.web.med.uni-muenchen.de
--
View this message in context: http://www.nabble.com/Extracting-correlation-in-a-nlme-model-tp23628854p23632075.html
Sent from the R help mailing list archive at Nabble.com.