I have question (a couple maybe) about the Univariate GLM procedure. I
have run a simply two-way ANOVA, but the cell sizes are unequal (i.e., I
have an unbalanced design). When I run the basic procedure (and ask for
descriptive statistics) I get weighted an unweighted marginal means.
This means are fairly different. My first question is, to which of these
means do the statistical tests for the main effects in the ANOVA Summary
table correspond to: weighted or unweighted. My second question is: how
do I get an ANOVA summary table for the "other" set of means given that
the default is either weighted or unweighted? And third, what role does
the Type of Sums of Squares within SPSS play? I have always used Type
III. Thank you in advance.
I may be missing something as your posting wasn't quite clear. To begin, do
you have weights? Are those weights turned on, i.e., was a Weight by
statement executed prior to the GLM statement? Your message suggests maybe
yes and maybe no.