xi3 and postgr3 are very useful commands. These two command work
together very well.
But, has anyone tried graph options in postgr3? It does not seem to be
implemented as in the previous version postgr which worked with xi2.
I would very much like to see Stata Corp extend xi3 and postgr commands
to xin and postgrn, where n reflects any desired level of interactions
(for most people, third order interactions would be enough, but in
experimental designs, the possibility of a few more should be
considered). We have now been through 2 external parties developing
similarly useful commands (i.e., desmat and xi3). Conceptually, I
prefer xi3 and postgr3 because they work well together to present both
statistics and graphical interpretations. Please make this wish part of
Stata 9.
