Monitoring and storage of irrelevant messages in selective attention

A series of experiments was designed to investigate the monitoring and storage of the irrelevant message during selective attention to one of two dichotic speech messages, and thus to throw light on the nature and level of selective “filtering” (Broadbent, 1958).

Abstract

A series of experiments was designed to investigate the monitoring and storage of the irrelevant message during selective attention to one of two dichotic speech messages, and thus to throw light on the nature and level of selective "filtering" (Broadbent, 1958). The experiments extend Cherry's finding (1953) that Ss noticed if the rejected message was identical to the selected one but a few seconds out of step. It was shown that (a) the comparison between the messages must be made at a central level, and required the identification of words and meaning rather than a simple analysis of the sounds, since Ss noticed the identity of the messages even when they were in different voices or different languages; (b) their identity is noticed at a 5-sec interval when a selected message of coherent prose leads the rejected one in time, but at 1 or 2 sec only when the order is reversed. The first result supports Treisman's suggestion (1960) that the filter acts by attenuating rather than blocking irrelevant signals. The second may give a measure of the different storage times for selected and rejected material. The critical delay interval for recognition also varied with the verbal content of the selected message: it decreased both when the information content of the words was high, and when the order of the items was crucial to the recognition. The relation of these to other findings on immediate memory is discussed.