A test of Vitz and Todd's model of the judged complexity of straight-line figures

Abstract

Two sets of stimuli, each consisting of nine geometric patterns, were computer-generated: in one set, the patterns were squares which were partitioned randomly into rectangles; in the other, the patterns were squares which were partitioned randomly into polygons. Ten subjects gave magnitude estimates of the complexity of the stimuli. The Vitz and Todd (1971) index of complexity was not a consistently better predictor of estimated complexity than the ratio of the number of lines in each pattern to an index of symmetry. Moreover, the power function relating the complexity estimates of the rectangular partitions to the Vitz and Todd index was significantly different from the power function relating the complexity of the polygonal partitions to the Vitz and Todd index. It is concluded that the Vitz and Todd model requires major additional assumptions in order to be a candidate for a process model of pattern perception. © 1975 Psychonomic Society, Inc.