Intraclass correlation coeficient

[Hazel Paola Quesada Leitòn]

Hello 

 I'm fitting a spatial bayesian hierarchical model using: 

inla(muerte ~ 1+edadcalcu +sexo.x + Tasa + IEPres+ f(cont1, model="iid",hyper = prec.prior) + f(cont, model="besag", graph="CR.adj",scale.model = T,hyper = prec.priorE) ,family = "binomial", data = datos,quantiles = c(0.1, 0.9),control.predictor=list(compute=TRUE),verbose=T,control.compute = list(dic = TRUE, waic = TRUE)) 

And I want to calculate the Intraclass correlation coeficient but I don't know how to get the residual variance or how to get the ICC. My question is if it is possible?

[Helpdesk (Haavard Rue)]

from what I understand, the data are conditional independent, hence your
ICC need to be defined on the latent field?? Then I don't know...

There is some deviance residuals available at

> r=inla(y~1,data=data.frame(y=1:10), control.compute=list(residuals=T))
> r$residuals$deviance.residual
[1] -1.7120450217 -1.3463430593 -0.9868108942 -0.6438683710 -
0.3673579296
[6] 0.3673579296 0.6438683710 0.9868108942 1.3463430593
1.7120450217


if you have the most recent testing-version

> --
> You received this message because you are subscribed to the Google
> Groups "R-inla discussion group" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to r-inla-discussion...@googlegroups.com.
> To view this discussion on the web, visit
> https://groups.google.com/d/msgid/r-inla-discussion-group/08932782-2926-4784-bbb2-a3ed2e339057n%40googlegroups.com
> .

--
Håvard Rue
he...@r-inla.org