The sandwich effect reassessed: effects of streaming, distraction, and modality.

The degree of disruption from interleaving auditory irrelevant items within a sequence of to-be-remembered items--the sandwich effect--was examined in two experiments. Previous demonstrations of the effect have shown that the penalty for interleaving items is small and that changing irrelevant tokens is no more damaging than repeating ones (contrary to the classic changing state effect of irrelevant sound). The results of Experiment 1 suggest that these earlier results were due to the lack of tokens in the irrelevant sequence (in part, the result of using a span method). The results of Experiment 2 also show that the sandwich effect was marked and, further, that it comprised two elements, one due to the partitioning of relevant from irrelevant streams (which may be promoted by similarity of identity within sequence or pitch disparity between sequences), and the other a classic irrelevant sound effect (with effects of changing state). The results are discussed in terms of the role of perceptual organization within and between modalities in short-term memory.