Overtraining and stimulus change in extra-dimensional shifts

Abstract

It has been established that overtrammg facilitates discrimination-reversal learning in human Ss under a wide variety of conditions. (See Wolff, 1967, for a review of pertinent studies.) Whether the learning of an extradimensional shift (EDS) is similarly facilitated by this variable is open to question. Considering this matter in a recent article, Wolff (1967) tentatively concluded that the exact effect of overtraining on EDS may ultimately depend on both individual difference variables and procedural variations in the administration of the shift itself. More specifically, he hypothesized that the magnitude of overtraining facilitation is an inverse function of both subject age and the degree of stimulus change from original to transfer learning, Supporting this hypothesis are the following facts.(I) Nearly all studies showing strong overtraining facilitation on EDS have been performed with college Ss and llave employed an arrangement in which the shift stimuli were the same as were used in original learning~ nearly all studies showing no significant effect of overtraining on EDS, on the other hand, have been performed with children and -have used <In arrangement in which the shift stimuli were different from those used in original learning. I Thus, either age, shift arrangement, or both variables, alone or in combination, act to modify the effect of overtraining on EDS.(2) Consonant with the hypothesis that shift arrangement is an effective variable in the above studies, Walther (1962) and Sugimura (1960), in the only experiments which compared the effects of overtraining on different EDS arrangements, found that the magnitude of overtraining facilitation in adults was inversely related to the degree of stimulus change occurring between original learning and shift. This effect, however, was not statistically significant in either study.(3) Also supporting the shift-arrangement interpretation is the fact that slight (but nonsignificant) overtraining "interference" has been obtained in nearly every study with children in which the stimuli were changed from original to shift learning. I This fact supports this interpretation because in the single study with children in which the stimuli remained unchanged over the shift transition, slight (but nonsignificant) overtraining "facilitation" was observed (Iwahara & Sugimura,Although not conclusive, the above studies suggest that both age and shift arrangement may be important variables in overtraining-EDS phenomena. In the present study, an attempt was made to demonstrate more clearly the effect of1he second of these factors. In line with Wolffs (1967) hypothesis and the trend of the above reported studies, it was predicted that overtraining would produce greater facilitation in an EDS in which the ~timuli remained constant throughout the task than an EDS in which the stimuli changed over the shift transition.EXPERIMENT I SubjectsThe original sample consisted of 123 graduate education students, of which 27 were eventually eliminated-one for refusal to cooperate, one f...