[R-sig-ME] ICC for mixed models with modeled residuals

[Simon Harmel]

Hello All,

I was wondering if for models like below, there is a standard way to
calculate the ICC given the extra modeling of residuals?

Any possible `R` package capable of calculating ICC for such models?

Thank you,
Simon

 lme(yij ~ X1 + X2 ...,
         random = ~1| subject,
         data = data,
         subset = DV == "DV1",
         correlation = corSymm(~1|subject),
         weights = varComb(varIdent(form = ~ 1 | X1),
                                          varPower(form = ~ X2)))


  glmmTMB(yij ~ X1 + X2 + ... + (1|subject),
               family = beta_family(), data = subset(data, DV =="DV2"),
               dispformula = ~X1 + X2)