ABSTRACT
The last chapter dealt with the basics of two-level factorial designs, including design geometry, main effect and interaction estimation, significance testing, and checking for significance if there are no repeat runs. On the surface, that might seem to be all you need to know in order to charge ahead and run one of these designs. But, in reality, there are a number of issues that must still be addressed in almost any practical application of twolevel factorial designs. Some of these issues are: (a) Will the number of runs from the full 2k
design give the accuracy needed? (b) How do you obtain a linear prediction equation from the effects and interactions that were calculated, and how do you check for the adequacy of that model? (c) What do you do with known background or nuisance variables that might influence the results but are not really of interest, like batch of raw material? (d) What can be done when the order of experimentation is difficult or impossible to randomize? These very important questions will be dealt with in this chapter.
