ABSTRACT
Chapters 10 through 12 discussed designs with random factors, with a special focus on withinsubjects designs. Chapters 13 and 14 provided a multivariate alternative for analyzing data from within-subjects designs. Chapters 15 and 16 present yet additional methods for analyzing data from designs that include one or more random factors. In general, the methods of these two chapters can be useful whenever scores are potentially correlated with one another. When might we expect scores to be correlated? One typical example is in within-subjects designs, where we obtain more than one score from each individual. Scores from the same person are usually more similar to one another than scores from different people, in which case scores will correlate with one another. Scores can also be correlated even when we obtain only a single score from each person. A typical example is when individuals are organized into clusters, such as classrooms, factories, or communities. Once again, scores within a cluster are usually more similar to one another than are scores between clusters, leading to a pattern of correlations among the scores. This chapter focuses on the first situation, where multiple scores are obtained from each individual. The next chapter presents the second situation, where scores are clustered, as in a nested design.
