Interpretation of the hyperparameters

[Clara Delecroix]

Hi,

Sorry for this naive question, I am new to R INLA.

I am fitting a model for several areas of interest. Each model includes fixed effects, and random effects (spatial effect using besag model, random effect using iid model and a random effect of the country using iid). I would like to interpret the proportion of variance explained by each of these effects. 

For that, I have looked at the precision of each of these effects using fit$summary.hyperpar. I calculate the variance as the reciprocal of each random effect and then calculate the proportion of the variance as the variance of one effect / sum of variances of the different random effects. Is that correct for an INLA model, or is there a better way to interpret and compare these effects? 

Thank you in advance,

Clara