Hi,

I need a little bit of help!

I am working on a multivariate MCMC model, I basically want to model the relationship between a trait and fitness, with fitness modelled with a hurdle model. The residual variance of the binomial part of the fitness has to be fixed (fix=3 in the R list of prior), so I'd like to fix its residual covariances with the two other response variables.

I cannot find anywhere how to indicate that in the model..

My first thought was to do:

in the prior R = list(R1 = list(V=diag(3), fix=3, nu = 3) , R2 = list(V=diag(2), nu = 2))

in the model rcov = ~ idh(trait):units + us(at.level(trait, 1:2)):units,

But R then tell the block-diagonal R structures are not yet implemented.

Do you know anyway to do that?

Thanks a lot!