DESIGN TOPICS II
Two kinds of designs are considered in this section. The first is nested design, an interesting, often useful variant of factorial design. The second concerns designs with natural groups of subjects, one application of nesting.
The treatment is brief. Its purpose is to call attention to the conceptual structure of these kinds of designs. This will help you understand what an article is up to when it says it used nesting. You will also have one important check on the validity of a study in which subjects are run in groups.
By understanding the nesting idea, moreover, you will be able to recognize when this kind of design might be useful for your own purposes. In that case, you should consult a text that goes into detail on the many variants. a
Factor X is said to be nested in factor G if each level of X appears at only one level of G. This differs from the crossed factors in the usual factorial design, in which each level of X appears at all levels of G.
The difference between nesting and crossing appeared in the mixed repeated measures designs of Chapter 6: Subjects is nested in the between variable, crossed with the within variable. In the (S × A) × G design, each subject appears at only one level of G, the between subjects variable. Hence S is nested in G; but S is crossed with A because each subject appears at every level of A.