[R-sig-ME] Question about lme4

[Ben Bolker]

jesse cloud <cloud.jesse7 at ...> writes:

> 
> Hi, This is Jesse. I am very interested in your lme4 package. I hope to fit
> the linear mixed model using your package. But I can only get the variance
> estimates for the random effect from the output. Is that possible to get
> the asymptotic variance estimates of these variance estimates?
> 
> Thank you very much.
> 
> Best,
> Jesse
> 


  With some difficulty, yes.  I seem to have posted partial answers
to this on Rpubs twice:

http://rpubs.com/bbolker/varwald
http://rpubs.com/bbolker/waldvar

The latter is later, but it might be worth looking at both of them
(at some point I should reconcile them ...)

Keep in mind that the asymptotic variances of the variance estimates
are often very inaccurate summaries of the uncertainty ...