ABSTRACT
The over-all objective of a factorial experiment may be to get a general picture of how the response variable is affected by changes in the different factors or to find the over-all combination of factor levels which gives a maximum value of the response variable. The experiments suffer from several serious defects and this chapter explores that it is often preferable to run what is called a complete factorial experiment, in which one or more trials is made at every possible combination of factor levels. Such an experiment is often simply called a factorial experiment. The most serious defect of one-at-a-time experiments is that they are unable to detect interactions between factors, whereas complete factorial experiments can. The major advantage of complete factorial experiments is that they are the most efficient way of estimating main effects even if no interaction is present.
