Use independent draws from a normal distribution to model a main effect or interaction. Typically used with variables other than age or time, such as region or ethnicity, where there is no natural ordering.
Details
Argument s
controls the size of errors. Smaller values
for s
tend to give more tightly clustered estimates.
Mathematical details
where
Parameter s
is provided by the user.
See also
NFix()
Similar toN()
but standard deviation parameter is supplied rather than estimated from datapriors Overview of priors implemented in bage
set_prior()
Specify prior for intercept, main effect, or interaction