Rote learning as a function of similarity structure and task paradigm

Abstract

The purpose of this study was to investigate rote learning as a function of similarity structure and task paradigm. Similarity structure was vie-wed from the standpoint of psychological proximity, and it was scaled according to multidimensional techniques. A general review of the literature indicated that item structure was to be considered from at least two viewpoints. These were: (a) item structure as ordered by the serial nature of the rote task, and (b) internal structure as it underlies the components which comprise the learning task. It was hypothesized that: (a) the optimum conditions for rote learning exist when the structure implied in the serial nature of the task coincides with the similarity structure which underlies the task items to be learned; (b) no significant differences should develop when these two types of structure are placed in opposition to each other; and (c) the most inefficient conditions for rote learning occur when benefits of neither type of structure are available. The similarity structure generated to investigate these hypotheses utilized certain inferences from circumplex theory. The structure implied in the serial nature of the rote tasks was varied utilizing modifications of single and double function paired-associates paradigms. The present study was conducted in two phases. In the first phase eight prototypes of dot patterns were constructed following certain random procedures. These prototypes were linked together in combinations of four to produce eight new stimulus patterns. Similarity judgments of these patterns were obtained using the complete method of triads. Preliminary analyses of Phase One data revealed that: (a) the matrix of proximities closely resembled the correlation matrix in circumplex analysis; and (b) the psychological space implied in the similarity judgments largely conformed to the original preconstructed circumplicial structure underlying the physical stimuli. In the second phase, two item orders were generated for. a paired-associates task. These orders represented maximum adjacent item similarity, and minimum adjacent item similarity. Phase Two data were analyzed by a 2 x 2 factorial design and a test for multiple comparisons. The results of these analyses were interpreted as supporting all experimental hypotheses. Two theoretical viewpoints, inhibitory-facilitory hypotheses and discrimination theory, were examined as possible explanations for the experimental results. Difficulties arose in application of the former theories, but it appeared that the latter hypothesis was more appropriate for the data. It was concluded that an extension of the discrimination hypothesis in terms of discrimination-distances best accounts for the results of the present study.