Screening of important variables

To determine which variables are important to control, it is absolutely necessary to use strategies which allow all variables to be varied simultaneously over the set of experiments. In such experiments it is equally important that the variation of each variable should be uncorrelated to the variation of other variables, i.e. that two or more variables should not be changed in the same direction over the set of experiments. 

This can be avoided H factorial designs or fractional factorial designs are used for designing screening experiments. Such experimental designs, and some other types of designs which can be used for screening experiments, will be discussed in Chapters 5-8. 

---