ABSTRACT
Although the factorial and fractional factorial designs that we have explored in the past chapters are considered the “workhorses” of modern empirical investigations, there is still a lingering question: Are these designs efficient? Remember that, to be efficient, an experiment must get the required information at the least expenditure of resources. The factorials certainly get the required information in the form of single effects and the interactions and do so at the least expenditure of resources, but what if there is a nonlinear relationship between the responses we are measuring and the factors we are changing in the design? In that case we could be misled. A classic example is shown in Figure 6.1. Here the levels of the experiment have been set at two extremes, and a straight line between them shows that the slope is zero.
