On the Specification of Operator or Occupational Workload with Performance-Measurement Methods

Five system-output or performance-measurement methods have been described in the literature for use in operator or occupational workload specifications: laboratory, analytic, synthetic, simulation, and operational-system methods. A review and analysis of these methods indicates that laboratory methods, where appropriate, are the methods of choice, with the synthetic-work technique especially well suited to examinations of general workload questions. Analytic and synthetic methods appear to yield reasonable results, but both rest on relatively fragile data bases; with correction of this deficiency and further research on time-sharing behavior or function interlacing, these methods should prove to be quite helpful, especially in systems designs and workload allocations. Simulation methods have the potential of providing quite useful information on operator workload, but simulators have not generally been employed for this purpose, and some of the difficulties implicit in their use are discussed. Operational-system methods, except for some possible safety limitations, can be used on virtually any workload-specification problem suitable for investigation in a simulator, but the problems of data recording can be substantial, and often there is little agreement on what should be measured as criteria of good performance. The need for reliable, valid, quantitative criteria to reflect system performance is stressed, and a potentially useful paired-comparisons scaling procedure is described.