Simultaneous vs sequential discriminations of Markov-generated stimuli

This experiment required Ss to make same-different classifications of Markov-generated histoforms that were distortions of three different prototypes. Task memory requirements were varied by presenting stimuli either simultaneously or sequentially for comparison, and the effects of shifts in memory requirements were also assessed. Classification performance in all cases was found to be superior under the simultaneous condition. Postshift performance was strictly a function of the mode of stimulus presentation following the shift, rather than the memory requirements present in the initial trials. These results, coupled with the observed relationships between performance and a measure of individual stimulus variability under the two memory conditions, strongly suggest that prototype encoding has limited relevance to performance in the same-different classification task used here.