The Application of Additive Factors Methodology to Workload Assessment in a Dynamic System Monitoring Task

摘要

This report describes the theory that lies behind applications of Sternberg's additive factors methodology to the selective assessment of primary task workload, within the framework of a multiple resources model of human information processing. In applying this methodology, a reaction time task is performed alone and concurrently with a primary task of interest. Orthogonally, a characteristic of the RT task is varied to prolong one stage of processing. If the effect of this manipulation is greater in the presence of the primary task, than in its absence (i.e., an interaction), then the resource demands of the primary task are assumed to overlap with those imposed by the RT manipulation. If there is no interaction (additivity), then different resources are required. Conclusions of previous research efforts that have employed additive factors in dual task paradigms are summarized. A validation experiment is then reported in which a failure detection/monitoring task is employed as a primary task. Manipulations of perceptual and response load of a secondary reaction time task are performed, while subjects engage in a primary task of monitoring for, and detecting failures of an autopilot-controlled dynamic system. This primary task is assumed to place heavy demands on the earlier stages of processing.