[R-sig-ME] VarIdent, extracting delta value

[Peter Dalgaard]

Afshartous, David wrote:
> All, 
> 
> When using varIdent to specify different residual error variance per strata,
> one obtains the ratio between the standard deviations of the strata (see P&B
> p.211), and this is included in the summary output.  However, perhaps I'm
> missing something extremely basic but I don't see where this is within
> str((summary(m1)) for a given model and hence cannot extract it. A
> workaround is below, but if anyone knows the direct method please advise.
> 
> Cheers,
> David
> 
> 
> Library("nlme")
>  m1 = lme(distance ~ age, random = ~ 1 | Subject, weights = varIdent(form =
> ~ 1 | Sex), data = Orthodont)
> summary(m1)
> ## shows that delta = 0.4533690
> ##
> Variance function:
>  Structure: Different standard deviations per stratum
>  Formula: ~1 | Sex 
>  Parameter estimates:
>      Male    Female
> 1.0000000 0.4533690
> 
> ## 
> str(summary(m1)) 
> 
> ## workaround:
>  1/attr(m1$modelStruct[[2]], "weights")[100]
>    Female 
> 0.4533690 
>

I think the canonical way is

> exp(c(summary(m1)$modelStruct$varStruct))
[1] 0.4533690

or

> coef(m1$modelStruct$varStruct, unconstrained=FALSE, allCoef=TRUE)
     Male    Female
1.0000000 0.4533690

I'm not really sure that there is not a neater way of drilling down to

m1$modelStruct$varStruct

or summary(m1)$...


-- 
   O__  ---- Peter Dalgaard             Øster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark      Ph:  (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)              FAX: (+45) 35327907