Including group-level covariates in multi-level model

Interfaces rstanarm
1

Hi, just a quick coding question related to RStanARM and multi-level models. Do I need to do differentiate between group- and individual-level covariates? For example, with the infamous Radon study, could I include both the indicator for floor and the county mean at the individual level, for example:

stan_lmer(formula = y ~ x + u.full + (1 | county), data = ...)

…or is there some other structure I need to use? And if so why?

Thanks,
Blake

2

I think that is fine.