Visual modelling and discovery learning as constraints on the acquisition of coordination

Scully and Newell’s (1985) model of observational learning conceptually linked the evidence from visual perception research on the nature of information picked up by observers (i.e., relative motion information), to Newell’s (1985) framework of motor learning stages (cf., coordination, control, and skill). They proposed that the first stage of learning is concerned with the assembly of a coordination pattern, and that coordination-relevant information can be picked up easily from a model due to the evolutionary sensitivity of the human visual system to relative motion information. It was predicted that augmented visual perceptual information on coordination of the model could constrain the learners’ assembly of a successful coordination solution by directing search in the perceptual-motor workspace toward an appropriate attractor. Some work by Schoenfelder-Zohdi (1993) has lent credibility to these arguments. It was found that observational learning facilitated the approximation, within some bandwidth, the relative motion of a model more successfully than a strategy of discovery learning. However, these data on the efficacy of the visual perception perspective were based on analysis of modeling by an individual subject, and there is clearly a need for further research on this approach. The aim of this experiment was to examine the predictions of the visual perception perspective with a simple, low-constraints aiming task. Specifically, we were interested in the following predictions: (i) due to observation of the model directing learner’s search of the perceptual-motor workspace, it was expected that modeling subjects would approximate the model’s relative motion more quickly than a discovery-learning group. It was expected that a modeling group would show a lower level of inter-trial variability in movement coordination compared to a discovery group, particularly early in practice; (ii) if the observers acquired an appropriate coordination pattern more quickly than discovery learners due to directed search, it was predicted that the modeling subjects would score higher than the discovery learners, particularly early in practice.