Skip to content
Prev 13985 / 20628 Next

Standard errors of group-wise random effect intercepts

Hi all,

I might be searching for something that doesn't exist -- but is there a way
to obtain group-specific standard errors for random effect intercept
estimates?

I have hierarchical data grouped by "site," for which I've generated unique
intercept coefficients. For example:

$random$Site
          (Intercept)
ab      -9.574204
am     -9.149834
ay      -2.238734
br        5.073831
...

Is there a way to extract some sort of confidence interval on these values?
I have attempted using VarCorr(), but am having trouble getting it to
return a standard error beyond that of the standard error across *all* site
intercept estimates.

If it helps, my model is specified as:

lme1 <- lme(Biomass ~ Basal.area*Latitude - Latitude -1,
            random = ~1|Site, method="REML")

?Many kind thanks,
Jacob?