[R-sig-ME] Approximate standard error for variance component with lmer

[Douglas Bates]

2010/3/1 Julien Martin <julien.martin2 at usherbrooke.ca>:
> Hi
> I am wondering if it is possible to compute approximate standard error
> for variance component as proposed by Fischer et al 2004 using lmer.
>
> Thanks
>
> Julien
>
> Fischer, TM., Gilmour, AR. and Werf, JHVD. (2004). Computing approximate
> standard errors for genetic parameters derived from random regression
> models fitted by average information REML. Genetics Selection Evolution
> 36, 363-369.

Part of the point of the profiling of the parameters in mixed-effects
models that I describe in chapters 1 and 2 of the book I am writing
(chapter drafts are available at
http://lme4.R-forge.R-project.org/book/) is to show that the estimates
of variance components are not symmetrically distributed - in fact
they are quite asymmetric.  In such cases there isn't much point in
computing a standard error.  You could do so from the profiles but I
don't see the point.