Yes, it is possible to use multicomp for multi-strata designs.
It is more difficult than for single-straum designs because multicomp
doesn't work with the aovlist objects. Instead you must respecify the
design as a single stratum design and use that.
If you are working with the subplot treatments you will get the right
answer immediately. If you are using whole-plot treatments, then you will
need to use the more primitive multicomp.default and specify the denominator
mean square yourself.
There are two relevant examples in the HH package, available from CSAN for
S-Plus 8 http://csan.insightful.com.
The functions are also available for S-Plus 7 from my book's website:
Heiberger, Richard M., and Burt Holland (2004). Statistical Analysis and Data
Display: An
Intermediate Course
with Examples in S-Plus, R, and SAS, Springer–Verlag, New York. With
accompanying online files:
Springer-Verlag, New York. http://springeronline.com/0-387-40270-5
?mmc gives an example split plot design using the subplot treatment and
residual.
?mmc.mean shows how to use the multicomp.default (actually I wrapped that
inside the MMC function, but the calling sequence is similar).
Rich
|