Probability, Statistics, and Reliability for Engineers and Scientists

coefficient is significantly different from zero using the “T” test. The chapter also contains sections concerning the validity of the second-order prediction equation and the idea of quadratic optimization where the analyst may be interested in finding a minimum, maximum, or a saddle point, on the response surface generated by their second-order model. The authors conclude the chapter with a detailed flowchart of the sequential approach for design optimization. In conclusion, the authors’ goal of “to elegantly implement common statistical tools and the new ones to improve the quality of the process that is used to manufacture a product or render a service” was moderately successful. Use of the word “elegantly” may not be appropriate here, but one cannot fault the authors’ use of common statistical tests and graphical analyses. As to whether they have convinced the reader that they have given him/her the knowledge “to improve the quality of the process that is used to manufacture a product or render a service” is unclear. Certainly, the authors have attempted to present a fair amount of design of experiment material in this text, but the approach is not a strong one. The use of step-by-step approaches is commendable, but it is probably just as important to be sure the analysis is done correctly. Looking for significant effects through a series of t (not “T”) tests is best done on an individual basis, and not when you want to look at the overall picture (then use ANOVA). While intention is good, I cannot recommend this book as a serious text on designed experiments, but it could be a nice refresher book for someone already exposed to the material through a better text—assuming that some of the problems outlined in this review are addressed.