Stimulus array and recall method as variables in audio-visual split-span memory

Performance on an audio-visual split-span task was found to be adversely affected by the concurrent operation of two factors, namely: alternation of stimulus classes between the two input messages, and the use of a recall method which entails alternate sampling of the two messages. However, these factors had no significant effect when operating singly. The present study indicated that subjects prefer the strategy of reporting visual items first when using modality recall, and that this strategy is more efficient than that of reporting auditory items first. Also, modality recall proved to be the preferred method under free recall instructions. While there were more visual errors than auditory errors, visual commissions exceeded order errors, but auditory order errors exceeded commissions. The redundancy proposal of the search hypothesis (Yntema and Trask, 1963) was tested, and was found to be inadequate as an account of audio-visual performance.