Disruption of proofreading by irrelevant speech: Effects of attention, arousal or memory?

Abstract A series of five experiments examined the effects of irrelevant speech on proofreading and memory. Four of the experiments used a proofreading task and showed that the deleterious effects of irrelevant speech: (1) depend on the speech being meaningful, (2) are only present when the burden on short‐term memory is low and (3) are manifested in a lower detection rate for non‐contextual as opposed to contextual errors. Neither the spatial location of the speech (either in terms of spatial dispersion of sources or spatial movement of a single source) nor the intensity of the speech (in a range bounded by 50 dB(A) and 70 dB(A)) had any effect on proofreading. Late selection models of attention are favoured by the results in preference to models having arousal, short‐term memory or early selection in attention as their basis. A final experiment showed serial recall for visual lists to be impaired by the presence of any speech‐like sound (including reversed speech and speech in an unfamiliar language) which suggests a set of phenomena qualitatively different from those associated with proofreading. Throughout the article the practical consequences of the findings are emphasized.