Measuring attentional demand in continuous dual-task performance.

Despite its intuitive appeal, the commonly held assumption that there is some general limitation on dual-task performance has been shown to be seriously flawed (Allport 1980; Navon 1984). Central to this has been the inability to measure the attentional demands of tasks, without which there is no way to determine whether their joint demands exceed the hypothetical general limit. In the absence of such a measure, dual-task interference can always be explained by the alternative possibility that specific interference has occurred. A method is described in which the attentional demands of tasks can be measured and cross-validated by the use of two scales. Two experiments are described in which a general attentional limit is found; the measurement of attentional demand is consistent across scales and can be made at a level of precision approximating that of an interval scale.