[R] fix variance parameter values for lmer estimation

Douglas Bates bates at stat.wisc.edu
Fri Apr 25 19:06:40 CEST 2008

On 4/25/08, Wilfried Cools <Wilfried.Cools at ped.kuleuven.be> wrote:
> Dear list-members,

>  a model is fit with lmer, but I want to force the variance parameter values
>  to be as defined by me

>  I thought, use 'start' to specify initial values and only allow for one
>  iteration ?
>  my question is how to do that ?

>  to specify the 2x2 matrix of variance parameter values:
>  start=list(groups=array(2,-0.5,1),dim=c(2,2))

Well, the problem is that if you are going to decide what the
arguments to the function should be and how they should be interpreted
you are going to need to write the function yourself.  There is
nothing in the lmer documentation that says it will fit a model with
the variance parameter values as defined by you and there is nothing
to say that you can specify a list with arbitrary names as the value
of the start parameter and somehow lmer will magically infer what you
mean by these starting values and do the right thing.  Sorry but I
don't write mind-reading software.  You are going to have to write it

>  now I need to make sure the mean structure is estimated, while the variance
>  parameter values are kept constant ??
>  e.g. lmer(resp~pred1+pred2+(1 +
>  pred1|groups),start=list(groups=array(2,-0.5,1),dim=c(2,2)),....)
>  cheers,
>  Wilfried
>  Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm
>  ______________________________________________
>  R-help at 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.

More information about the R-help mailing list