[R-sig-ME] LMM covariance structure

[Kingsford Jones]

On Wed, Mar 4, 2009 at 10:15 PM, Pietro Ravani <pravani at ucalgary.ca> wrote:
> Dear list
> I need confirmation or correction on this: how can the structure of
> the R vcv matrix of a LMM have non negative covariances if the
> residuals are independent given the random effects (and the co-
> variates)?

Hi Pietro,

Are you suggesting that i) error covariance within a given level of a
random effect should be negative, and ii) residuals are independent?
Neither is true.

> Covariance structures such as compound symmetry, AR 1,
> Toeplitz are specified by the user or are the marginal type of vcv
> implied/induced by the random effects.

In nlme such structures are added via the correlation argument.  See
?corClasses.

hope it helps,
Kingsford Jones


> Thank you
> Pietro
>
> ---------------------------------------------------------------------------------
> Dr. Pietro Ravani, MSc (Biostatistics), MD, FNCPI
> Associate Professor - Division of Nephrology
> Departments of Medicine & Community Health Sciences
> Faculty of Medicine - University of Calgary
> Foothills Medical Centre - AB Health Services - CHR
> 1403 - 29th Street N.W. (room C210N)
> Calgary, AB, Canada, T2N 2T9
> T 403-944-8168; F 403-944-2876; E pravani at ucalgary.ca
> ---------------------------------------------------------------------------------
> P Please consider before printing
>
>
>        [[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-mixed-models at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models
>