The items of concern are: 1)the PB designs that are not in the 2^(p-q) fractional factorial series have very complicated alias structure. It may be best for the standard to recommend the 2^(p-q) designs. 2) there are questions whether these designs are enhanced by adding center points; 3) Some of the examples in the current version contain factors that are not controllable; for one example, several factors are categorical; if we can find better examples, that would be helpful; 4) there has been a suggestion that it would beneficial to have a quantitative test for factor significance included , in addition to the graphical approach presently included; 5)there could be benefits to including an EXCEL worksheet to do the computations; 6)some suggestions have been made concerning inclusion of a measure of variance reduction or error reduction, associated with each factor; 7) there are other issues which have not been mentioned above.
Keywordsfoldover; fractional factorial design; half-normal plot; Plackett-Burman; ruggedness; screening design;
Item Going To Next Ballot Level