# INLA functions (continued)

**Gianluca Baio's blog**, and kindly contributed to R-bloggers]. (You can report issue about the content on this page here)

Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.

I have polished up one of the two functions I’ve thought of implementing for INLA and it’s now available in the development version of the R package. So if you’ve got INLA installed in your R version (see how you can do it here, if you don’t), you can update it to the development version by typing

inla.upgrade(testing=TRUE)

This will add the new function inla.contrib.sd which can be used to express the uncertainty in the structured (“random”) effects of your INLA model in terms of the standard deviation, instead of the precision (which INLA gives by default).

In the help, I consider a simple example: I simulate $N=12$ Binomial trials. For each, first I simulate the sample size to be a number between 80 and 100 (subjects)

n=12

Ntrials = sample(c(80:100), size=n, replace=TRUE)

**very**simple hierarchical model where each trial represents a clustering unit (

*ie*I assume a trial-specific effect). You do this in INLA by using the command

*precision*of the structured effect $z$ is of course just as informative, it is in general (more) difficult to interpret, because it is on the scale of 1/standard deviation. On the other hand, you can’t just take take the reciprocal of the (square-rooted) summaries to obtain info about the posterior distribution of the standard deviation, because the transformation is not linear and thus it does not directly apply to the moments and quantiles.

*that*posterior marginal distributions. That’s what inla.contrib.sd does

**leave a comment**for the author, please follow the link and comment on their blog:

**Gianluca Baio's blog**.

R-bloggers.com offers

**daily e-mail updates**about R news and tutorials about learning R and many other topics. Click here if you're looking to post or find an R/data-science job.

Want to share your content on R-bloggers? click here if you have a blog, or here if you don't.