Date of Award


Degree Type


Degree Name

Doctor of Philosophy


School of Arts and Humanities

First Advisor

Professor Craig Speelman

Second Advisor

Dr Guillermo Campitelli


Skill acquisition theories suggest that automaticity of lower-level processes is required before the acquisition of higher-level skills can be attempted. However, there is a disparity between the theoretical expectations of skill acquisition and the empirical findings in the transfer of training research. Research has found that when a change is made to the contextual conditions in which a skill is acquired, the learned response becomes less skilled. When skill transfer occurs performance is disrupted so that reaction times are slower than observed prior to the context change. This observation has been made with several different tasks, however no research has established whether a transfer disruption is observed with automatic skills. The discrepancy between the theoretical assumptions and empirical findings suggests that aiming for automaticity in education may not be best practice. The experiments in the current thesis were designed to examine whether automaticity disrupts or enhances transfer performance. The studies were based on Lassaline and Logan’s (1993) visual numerosity task and Speelman and Parkinson’s (2012) two-step task design. The study has a particular emphasis on individual differences, and thus individual participant data are explored to determine the pervasiveness of trends observed in the group data. In experiment one it was found that experimental design might play a role in the acquisition and probability of transfer, with the experimental conditions revealing differences in disruption and acquisition of automaticity. Group results in experiment two suggest that automaticity is unaffected by context changes, however individual results revealed that some participants failed to approach automatic performance. In experiment three participants were approaching automaticity, however a large percentage of participants did not demonstrate a shift from controlled to automatic processing. Furthermore, group results suggest that performance is unaffected by context changes in transfer, yet, this observation was not reliably presented amongst individuals with many individuals demonstrating transferable skills while not attaining automaticity. Overall, the results appear to be congruent with Lassaline and Logan’s (1993) findings. According to the group data, automaticity appears to facilitate transfer, and performance continues in accordance with the power law of learning; automaticity was transferred despite novel context changes. However, individual data indicates that not all participants are behaving this way. The current results question whether automaticity should be the desired outcome in education settings as many people failed to achieve automaticity. Further research is required at an individual level that includes factors such as working memory ability and task approach to determine why some participants deviate way from group data trends, and why they may be affected differently by context changes.

Included in

Psychology Commons