Observed, Unweighted Means

Means of Groups

Observed, unweighted means are available for full factorial designs involving effects of 2 or more categorical factors. When cell frequencies in a multi-factor ANOVA design are equal, no difference exists between weighted and unweighted means. When cell frequencies are unequal, the unweighted means are calculated without adjusting for the differences between the sub-group frequencies. These are computed by averaging the means across the levels and combinations of levels of the factors not used in the marginal means table (or plot), and then dividing by the number of means in the average.