I have unbalanced sets of data for a 3-way ANOVA.  Factor A has 37 elements,
Factor B has 6 and Factor C has 2 with 4 repetions for the first and 3 for
the second, so that there are 888 and 666 = 1554 sets of numbers.  My
attempts to perform contrast tests with the base package (2.8.1) have
resulted in errors that the vectors are not of the same length, or a default
summary.lm table for all one, two and three-way combinations (I need all the
interactions) with intercept values I cannot reconstruct.  Is there a
package that handles contrasts with unbalanced data sets, or have I missed
something obvious?

Thanks.

Dave Cooper

	[[alternative HTML version deleted]]