Date of Award

2005

Document Type

Thesis

Publisher

Edith Cowan University

Degree Name

Bachelor of Arts Honours

School

School of Psychology and Social Sciences

Faculty

Faculty of Computing, Health and Science

First Supervisor

Craig Speelman

Abstract

There are differing opinions as to whether skills learned in one situation can be transferred and used in new situations. Anderson's (1982, 1993) Adaptive Control of Thought theory states that complete transfer of skills from one situation to another will occur when the processes used in training are the same as those required in transfer. Logan's (1988) Instance theory posits that complete transfer will occur only if the problems used in training are identical to those used in transfer, and that partial transfer should not occur in any transfer situation. However research by Speelman and Kirsner (2001), and Speelman, Forbes and Giesen (2004) has found that in situations where the same processes and identical problems are used, complete transfer does not necessarily occur, with partial transfer sometimes occurring in these situations. There are some criticisms to the designs of Speelman and Kirsner and Speelman et al.'s experiments, therefore more research is needed to clarify some of the issues that have been raised. The current experiment investigated the skill acquisition theories of J. R. Anderson (1982, 1993) and G. D. Logan (1988), that predict that complete transfer will occur when the same problems are used in training and transfer. The study tested 61 participants on a computer task involving target problems in the form of six times table questions, and distractor problems in the form of addition or subtraction questions. While target problems were the same in training and transfer phases, distractor tasks differed between phases. It was hypothesised that changing the distractor problems would disrupt performance on the target problems, causing partial rather than complete transfer. Results confirmed the hypothesis, performance on target problems was significantly disrupted (p < .01) when the context of the task changed. The study has implications for the theories of Anderson and Logan, suggesting that such theories may benefit from amendment to allow for the effect that context can have on the transfer of skills.

Download

Share

 
COinS