This study represents a reanalysis of Heath's (1963) data using Coombs' unfolding technique. Some additional objectives of this study were to find the dimensions underlying the cognitive preferences, to test Heath's first two hypotheses in terms of individuals as well as clusters, and to compare cluster and individual analyses. The first dimension, which goes from Applications to Memory to Principles to Questioning, accounts for over half of the individuals and over two-thirds of the clusters. A two-dimensional configuration accounts for 90% of the individuals and 99% of the clusters. The ordering along the second dimension goes from Memory to Application to Questioning to Principles. The analyses of clusters and individuals were shown to give identical solutions. The following hypotheses were supported: (a) that the group would Physical Science Study, Committee (PSSC) group and the control be generally located in different regions of the same joint space, the PSSC group being more densely distributed than controls in regions near Principles and Questioning, the controls being more densely distributed than the PSSC group in regions near Memory and Applications; (b) that achievement scores for PSSC clusters would be higher in regions near Principles and Questioning than in regions near Applications and Memory; and (c) that the region-achievement relationship hypothesized in "b" would be stronger in the PSSC group than in the control group. Implications following from the analysis were discussed.