[ADMB Users] Values for random effects

[H. Skaug]

Hi

Since U is a categorical variable (factor) with 7 levels. I suspect
that you want a random intercept, so that you should have used:

....  random = ~ 1, group="U", ....

This will give you one single variance parameter.

Does this help, or do I miss your point?

Hans

On Fri, Nov 12, 2010 at 8:08 PM, Schweizer, Peter E.
<schweizerpe at ornl.gov> wrote:
> Hello admb users,
>
>
>
> I just recently  ‘discovered’ glmm.admb and  would be interested to learn if
> admb in R can provide a  single value for random effects, similar to a
> ‘working correlation’ in SAS when using GEE.
>
>
>
> As of now I have
>
> M1<-glmm.admb( Y ~  V1 + V2 + V3 +V4 +V5, random = ~ U, group="U", data=A1,
> family="nbinom")
>
> In this case, U has 7 groups and in the model summary for above  model M1 I
> am getting (7) values for the diagonal matrix of the random effects. What I
> would like to obtain is a value that expresses the random effect for all
> groups in ONE single value, similar to a ‘working correlation’ in SAS when
> using GEE.
>
>
>
> Does this sound reasonable?
>
>
>
> Kind regards,
>
>
>
>
>
> Pete
>
> _______________________________________________
> Users mailing list
> Users at admb-project.org
> http://lists.admb-project.org/mailman/listinfo/users
>
>