ABSTRACT
A limited number of tasks in educational and psychological scaling have been established. The tasks can be divided into four areas: Ordering Tasks, Categorical Ratings, Similarity Judgments, and Free Clustering. Ordinal tasks involve ranking psychological objects in some way to produce dominance data, that is, one stimulus dominates another. Such data are often called nonmetric because only judgments of greater than (>) or less than (<) are required. Ordered Categories subsume many of the most frequently utilized unidimensional scaling tasks. These measures are commonly referred to as Summated Ratings, Likert Scales, or Successive Categories. Such titles, however, refer to different assumptions about the data and different analyses applied to the ordered categories data rather than to the task itself. In free clustering, the psychological objects are individually listed on slips of paper or cards. Free clustering is valuable because the underlying structure of the objects is not predetermined. Similarity judgments can be handled by CLUSTER analysis or by multidimensional scaling.
