Confounded Experimental Designs and Human Factors Research

Abstract

The use of confounded factorial designs has been seriously neglected in the human factors literature. A confounded factorial is constructed by systematically confounding blocking variables with one or more interactions which are believed to be statistically insignificant or inconsequential to the researcher. These designs offer the advantages of increased economy and power. These designs are especially useful when research is being conducted on military personnel and subjects must be selected from different military facilities, which may result in heterogeneous subject populations. A concrete example illustrates how confounding of this type can be used to the researcher's advantage through the tailored construction of a confounded design.