In the Stata Journal (3:3) Carlin et al. present a very useful set of tools
for analyzing multiple imputed data sets. However, there doesn't seem to be
a way to use (or program the equivalent of) the adjust command. I wonder if
anyone has suggestions as to how this might be done. Specifically, I am
interested in predicted means and appropriate CIs after adjusting the
analyses to specific education and ethnic characteristics.
