Random effects standard deviation

[CyuHat]

First of all I would like to thank you for this great package. Secondly, I wanted to know if there is a way to get the random effects standard errors of the model produced by the svy2lme function. Because it is possible to access several other elements of the object, but not this.

Thank you

[tslumley]

It's not clear how to calculate a standard error for the random effects. It's not like ordinary maximum-likelihood mixed models where you can use the Fisher information. In theory there's a sandwich estimator, but it's very slow and requires fourth-order sampling probabilities, and it's not very accurate. I think you will need to use resampling. That's what the boot2lme() function is designed for.