Transfer in multiple-sign and transposition problems: implications for Osgood's transfer and retroaction surface

Abstract

Exploration or the nature of transfer in complex learning tasks and investigation of the role of stimuli and responses in relation to Osgood's transfer and retroaction surface were the purposes of this study. Two experiments, involving manipulation of specific stimulus-response combinations in multiple-sign and transposition problems, were performed to ascertain the conditions producing positive or negative transfer. Experiment One involved the investigation of various S-R combinations within "matching" and "non-matching" tasks. Three different sets of nonsense shapes served as stimuli, with "matching" and "non-matching" concepts as responses. During Task A, each experimental group received a different S-R combination or problem, while the control group was administered the House-Tree-Person (HTP) projective test. Task B consisted or the same S-R combination for all groups: the "non-matching" problem, utilizing identical stimuli with identical responses. After completing the transfer experiment, subjects in the experimental groups were tested tor concepts of square, triangle and circle. Two samples of each figure were presented by the experimenter to determine whether the ability to verbalize geometrical form concepts (form class recognition) is related to the learning of multiple-sign problems. The experiment's primary findings were that no negative transfer resulted despite the obtaining of large amounts of positive transfer, and that the ability to formulate simple geometric form concepts was not related to success at solving multiple-sign problems. It was concluded that Osgood's transfer surface had not been satisfactorily tested by the experiment. However, these results did form the foundation for a second experiment utilizing problem-solving tasks in examining S-R combinations in transfer. Experiment Two involved the investigation of various s-R combinations (including stimulus predifferentiation) with "oddity," "smallest-size," and "largest-size" problems. Three different sets of nonsense shapes served as stimuli, with "oddity," "smallest-size" am "largest-size" concepts as responses. During Task A, the HTP projective test was administered to the control group, while each experimental group received a different S-R combination. Task B consisted of the same S-R combination for all groups: the "largest-size" problem, utilizing identical stimuli with identical "largest size" concepts. The principal findings of this experiment were fivefold: 1. No negative transfer resulted from any of the various S-R conditions. 2. "Largest-size" and "oddity" concepts act like neutral responses. 3. "Largest-size" and "smallest-size" responses appear to be similar, functionally. 4. No differential transfer effects were obtained between identical and similar stimulus conditions. 5. Stimulus predifferentiation training failed to produce any significant transfer. Since the responses fell on the neutral-similar part of the response dimension of the transfer surface, it was concluded that none of the concepts involved in the multiple-sign and transposition tasks were "opposite" in terms of the transfer surface. This latter generalization, however, depends upon the validity or the Osgood transfer surface.