Date of Award

2007

Degree Type

Thesis

Degree Name

Bachelor of Arts Honours

School

School of Psychology and Social Sciences

Faculty

Faculty of Education & Arts

First Advisor

Craig Speelman

Abstract

Skill transfer is a fundamental feature in the domain of skill acquisition, however different theories present conflicting ideas regarding prediction of transfer. Anderson's (1982) Adaptive Control of Thought theory posits that the amount of transfer is proportional to the number of shared productions. Logan's (1988) Instance theory in contrast, posits that complete transfer will only occur on tasks which have been experienced before. However, work by Speelman and Kirsner (1997), Speelman, Forbes and Giesen (2004) and Johnson (2005) have produced results that counter the implicit assumptions of these theories. More specifically a disruption from the predicted learning curve was found in situations where both Logan and Anderson would predict complete transfer. Such findings have implications on the theoretical nature of skill acquisition. However these findings involved data that was averaged over many trials, and as such a more specific trial by trial account of the underlying nature of this disruption is not known. The present study examined the role of skill acquisition and skill transfer in relation to both Anderson's (1982) Adaptive Control of Thought theory and Logan's (1988) instance theory. The study involved 59 participants being presented with multiplication problems from the six times table, followed by a distractor task in the form of an addition or subtraction question. The training phase consisted of 12 blocks of trials, with six trials per block. The multiplication problems remained constant in both the training and transfer phases, while participants receiving addition problems in training then received subtraction questions in transfer and vice versa. Only reaction times for the multiplication problems were analysed and the results showed that there was a significant disruption when the participants encountered the transfer phase. Further analysis on a trial by trial basis indicated that in the three trials following the contextual change, performance was significantly slower than the first trial of the transfer phase, discounting a fleeting 'surprise' effect. A power function fitted to the reaction times of the transfer trials following the context change supported the gradual return of performance to pre-transfer levels. The results contradicted both Anderson's and Logan's predictions, and suggest that with practise, a mental set is developed which can then be disrupted through contextual change.

Share

 
COinS