ABSTRACT
This chapter compares the two-factor design with the other tests, using the Key to Statistical Tables. The two-factor design falls in the two-or-more-factor column. The two-factor repeated-measures design could be considered to be a one-factor related-samples design; the one factor would be the two or more experimental treatments while the second "subjects" factor would merely indicate that the design is a related-samples design. It is worth considering formally the assumptions for use of the two-factor repeated-measures ANOVA, without interaction. These are similar to those for the completely randomized design. The two-factor ANOVA makes the same assumptions of normality and homoscedasticity as the one-factor completely randomized design; these can cause problems for testing the data from rating scales. A repeated-measures design is preferable to a completely randomized design for sensory testing because then the treatment means are more likely to vary as the result of differences between treatments rather than differences between subjects.
