Abstract
The usefulness of visually based flight simulators for training may be compromised by the phenomenon of simulator sickness. Although a significant problem, design specifications for alleviation of simulator sickness have not yet been produced. The problems researchers face in this area are multiple: (a) Simulator sickness ispolygenic, rendering experimental isolation of variables ineffective; (b) simulator sickness is polysymptomatic, which must be reflected in measurements of the human response; and (c) there arc statistical limitations including limited sample sizes, adaptation over flights, small effect sizes, and large individual differences. These problems render certain research and engineering strategies more effective than others. Our conclusions, based on analysis of a large data base (more than 2,000 exposures in military simulators) and statistical power calculations, suggest that improved simulator-design criteria can best be studied in 'field experiments' in which large sample sizes permit the relatively small effects of several different simulator-equipment features to be isolated, contrasted, and revealed.