Protocol Analysis of Aptitude Differences in Figural Analogy Problem Representation

Individual differences in performance on figural analogy tests are usually attributed to quantitative differences in processing parameters rather than to qualitative differences in the formation and use of representations. Yet aptitude-related differences in categorizing standardized figural analogy problems between high and low scorers have been documented which bear close similarities to expert/novice differences in problem representations. Protocol analysis was used to assess problem representations in an attempt to provide converging evidence of aptitude-related qualitative differences. Twenty high and low scorers on one form of the Cognitive Abilities Test (CAT) figural analogy subtest, selected from a large college population, served as subjects. Subjects solved 25 multiple-choice figural analogy problems from a CAT subtest, reviewed problems and described analogy relations, rated problem difficulty, and described their problem solving strategy. Three independent raters assessed relationships attributed to each analogy problem in terms of seven theoretically fundamental relations (rotation, reflection, number, size, shape, shading, and spatial displacement). Preliminary results suggest that formal analyses will reveal strong qualitative differences between high and low scorers. High scorers appear to represent problems in terms of well-constrained, transformational relations, while low scorers focus on looser spatial relations or on shared perceptual characteristics, and are much more idiosyncratic in their representation than are high scorers. (NB)