The use of a multidimensional metric for the subjective evaluation and prediction of workload has been proposed by a number of researchers, including McCracken and Aldrich (1984) in their attentional demand methodology. This model appears well suited for the prediction, using network simulation, of levels of attentional demand and of task conflicts placed upon operators. This paper examines the development of multidimensional task rating scales and task conflict matrices and introduces a new process of development of these workload prediction tools. It also presents two studies in which the method is applied, and the resulting equal-interval rating scales and task conflict matrices. These studies indicate that while these tools may be useful within a specific operational environment, their utility as a general purpose predictor of workload is questionable.