Examination of Workload Measures with Subjective Task Clusters

Abstract

Multiple measures of operator workload may fail to agree or dissociate for a given task. This study proposes a new method to examine this dissociation for two categories of workload measures: Subjective ratings and performance-based secondary tasks. Eighteen tasks of differential processing resource demand were performed by subjects and rated according to workload similarity. Additive clustering analysis of the workload ratings produced overlapping task clusters. Three properties–performance, effort, and input complexity-explained the cluster solution. Dissociation was found when tasks perceived as similar in workload did not possess the same properties.