Technometrics, November 1996

Publisher's note: The Complete Guide to the CRE was originally titled The Reliability Engineering bible.

Reliability Engineering Bible [The Complete Guide to the CRE], by Bryan Dodson and Dennis Nolan, Tucson: Quality Publishing, 1995, xii + 489 pp., $99.00

As the title suggests the authors ostensibly had a lofty goal in mind during the writing of this book. Although it has many good qualities, however, it does not deserve the sublimation of being called the Reliability Engineering Bible. The authors claim that this book has three valuable uses: (1) It provides a good preparation for the Certified Reliability Exam (CRE) sanctioned by the American Society for Quality (ASQ). (2) It can be used as a textbook for a two-semester college course in reliability. (3) It serves as a comprehensive reference book. I proffer, provided the engineer has experience and/or format training in reliability models, that this book does an excellent job at supporting their first and third claims but misses the mark on the second.

As stated in the preface, the authors' primary goal is to "present the entire body of knowledge for reliability engineering as defined by" the ASQ in their CRE examination booklet. That goal is definitely achieved. Not surprisingly, because of the comprehensive aim, the material is often somewhat cursory. Nonetheless, a formally trained reliability engineer will find this book a useful reference in preparing for the CRE exam. It is very well organized and with few exceptions follows the outline given by ASQ. Pursuant to this theme, Chapter 1 offers advice on exam-taking strategies as well as detailing the materials that you should bring to the exam such as calculators, snacks, pencils, references, and so forth. Chapter 2 completes the introduction with discussions of data integrity and analysis, as well as defining an array of relevant terms related to failure data.

The remainder of the book is composed of chapters that are concomitant with the various topics outlined in the CRE examination booklet (on-line versions available at the ASQ home page, Internet address: http://www.asqc.org). Chapters 3-6 cover mathematical reliability models supplemented with numerous and varied examples. Chapters 7-11 cover the "nonquantitative" aspects of reliability. Because the CRE exam heavily emphasizes the concepts in Chapter 3, "Distributions," and Chapter 4, "Prediction, Estimation, and Apportionment Models" (together composing over 40% of the book), I will take commensurate time here to discuss the quality of these chapters.

Chapter 3 begins with an introduction to discrete and continuous distributions and then covers parameter estimation (maximums likelihood estimators and method of moments only), hazard and probability plotting, and finally confidence intervals and hypothesis testing. Although I found the material on hazard and probability plotting to be of superb quality, I am disappointed with this chapter overall. In their zeal to avoid "proofs and derivations," the authors left out substance as well. Because the CRE exam heavily emphasizes the concepts covered in this chapter, to get any benefit I feel that the practitioner must have at least one course in mathematical statistics and one course in reliability, as well as a basic statistical knowledge. This undermines the authors' thesis that practitioners with "a basic knowledge of statistics will be able to pass the ASQ CRE exam after reading this text and working the exercises at the end of this chapter" (p. xi) There is no motivation or discussion as to what reliability means; just its probabilistic meaning is given. Similarly they define means and variances of random variables with no discussion as to how they relate to each other or how they relate to reliability. Moreover, because they are purporting to have written this book for reliability engineers, is seems that they would have some substantive discussion about desirable behavior and/or distinctive behavior for hazard functions, yet they do not. For example, there is no reference to the "bathtub" curve until the end of Chapter 6, page 276, and this, it seems, is made only in passing. For similar reasons, I feel that this book is not appropriate for the classroom, regardless of the students backgrounds. There are many examples in which the authors are too casual with important details. For the sake of brevity, I will only mention a few transgressions. The independence requirement is not mentioned in the context of a t statistic being proportional to the ratio of a standard normal and the square root of a chi square. The support of Poisson distribution is given as finite instead of countably infinite. More globally, they make no clear distinction between parameters and estimates of parameters and are very loose with language, using phrases like "moment estimators. . .match the accuracy of maximum likelihood estimators if the sample size is large" (p.66).

Chapter 4, on the other hand, is quite substantive and very well written. It begins with an informative discussion of reliability data acquisitions from public databases such as FARADA (Failure Rate Data) and IEEE Data. The rest of the chapter pertains to the reliability of systems- that is, series, parallel, and k-out-of-n. I found the coverage here to be superb and comparable to that of Martz and Waller (1982), a reference I always keep handy. Two other timely topics that this book treats superbly, on an introductory level, are Monte Carlo simulation and software reliability.

Chapter 5, "Reliability Testing," and chapter 6, "Maintainability and Availability," complete the quantitative portion of the book and they, although quite terse, are of high quality. Chapters 7-11 cover the topics of criticality analysis, part selection and derating, reliability management, product safety, and human factors in reliability, respectively. Each chapter is filled with useful information for the practitioner, presented clearly, and demonstrates the authors' immense industrial experiences.

Finally, the authors' last claim is that "practitioners will find this text valuable as a comprehensive reference book" (p. xii). This is mostly true in that it provides an excellent outline for timely topics important to the practitioner. Practitioners will often find, however, that they must go to the references, given at the end of each chapter, to get substantive details. Books such as those of Lawless (1982), Martz and Waller (1982), Kececioglu (1991), and Nelson (1990) are invaluable references, approved by ASQ and should be used in conjunction with a book as such as the one we are discussing.

In summary, I enjoyed reading this book and found it to be relevant and informative. It most definitely can serve as a useful reference to the practitioner, but for the reasons discussed previously, it does not warrant the name Reliability Engineering Bible [The Complete Guide to the CRE]. Although I do not recommend this book for use in the classroom, a reliability engineer will find that reading this book, in conjunction with other ASQ-approved references, is a smart strategy for passing the American Society for Quality Control's Certified Reliability [Engineer] Exam.

Mark Carpenter

Sam Houston State University

Kececioglu, D. (1991), Reliability Engineering Handbook (Vols. 1 and 2), Englewood Cliffs, NJ: Prentice-Hall.

Lawless, J.F. (1982), Statistical Models and Methods for Lifetime Data, New York: John Wiley.

Martz, H.F., and Waller, R.A. (1982), Bayesian Reliability Analysis, New York: John Wiley.

Nelson, W. (1990), Accelerated Testing, New York, John Wiley.