Context has an important influence on performance in a variety of tasks. In the present experiment, the context of interest was the number of consecutive trials under identical search conditions. The authors investigated how individuals learn to benefit from one form of contextual cues, the time course of such benefit, and the effects of contextual manipulations on general learning, feature learning, and automatic process development. These issues were examined using a visual search task in which investigators could manipulate both consistency and learning context. The results suggest that the manipulation of context influenced feature learning; that is, at least 10 consecutive trials were required before optimal scanning strategies could be developed and/or instituted. However, the training context manipulation did not affect the acquisition of an automatic attention response in a consistent task or the acquisition of a general skill for a varied task. Implications for task and system design and the development of training programs are discussed.
Abstract