Introduction to Decision Analysis (Second Edition). (Book Review).

By David C. Skinner

Probabilistic Publishing, Gainesville, Florida, 1999, xiii + 369 pp. ISBN 09647938-3-0. List: $34.95. (Hardback)

David C. Skinner has given an interesting and useful overview of decision analysis, combining business, psychology, and engineering aspects.

Decision analysis applies knowledge from three main fields. Business contributes understanding of the business context for decision-making, ideas from strategic thinking, and a framework for how decision analysis can be used by groups. From psychology, the field gets understanding of human biases that can affect probability and value judgments. And engineering is primarily where the more quantitative aspects of decision analysis are studied and developed: probability, utility theory, influence diagrams, and the value of information.

While Decision Analysis is strongest in the areas drawn from business, it covers all the topics I have listed above from the three areas and integrates those topics well. The level of detail is not enough for someone to learn the mathematical topics from this book. (He says, 'This book will help you start to learn decision analysis.") But Mr. Skinner often explains the business implications of the mathematical aspects. For example, he mentions that the computed value of information can be useful in eliminating analysis paralysis, that is, the tendency to continue to do analysis to avoid making a decision.

Mr. Skinner is very generous in sharing material that is useful for teaching decision analysis. His web site (www.tifoe.com) has several forms for various steps of decision analysis, a useful explanation of assessing probability distributions, an explanation of how to create influence diagrams, and other material. Some of this material is in the book, but some is not. I did not check every page, but he seems to have granted permission for the use of the material as long as acknowledgement is given to him.

Mr. Skinner's book has also helped me to understand how decision analysis is used. Because I have not had much experience outside of academics, I have gathered examples of uses of the material I teach, from journals, periodicals, former students, and other practitioners. I have felt confident in teaching students about the actual applications of many industrial engineering and operations research techniques: especially linear programming, simulation, regression, and statistical quality control. In particular, I always assure students that, from among all their classes, all of them will use ideas from engineering economics.

However, despite the appeal of decision analysis, I have had more trouble collecting examples of its use. While I have used some articles and had one former student tell me how he used the material, frankly I felt uncomfortable reassuring students that the material would be useful. I usually resorted to the argument that the relative lack of published cases involving decision analysis was evidence of its use; it was so valuable that companies did not want to talk about how they use decision analysis. I recognized the weakness of that argument.

Mr. Skinner was one of the founders (and served as Chair) of the Decision Analysis Affinity Group (DAAG) and his website (www.tifoe.com) also hosts material for DAAG and for its annual conference. I quote from one web page (www.tifoe.com/daaginfo.lhtml):

"The Decision Analysis Affinity Group (DAAG) is a multi-industry group of decision analysis practitioners who get together once a year to share ideas, successes, and failures. Industries represented include oil and gas, pharmaceutical, utilities, heavy manufacturing, automotive, and chemical. The annual meeting is held at a sponsoring practitioner's site. In the past five years we have had Eli Lilly, Texaco, Monsanto, and Chevron host the conference. Each year two Chairmen are selected to oversee the conference."

Throughout the book, Mr. Skinner uses examples from work he has done to illustrate his explanations. Also, the DAAG conference papers are online and give other examples of the use of decision analysis. Decision analysis is, indeed, used successfully at a number of large companies.

The book also helps me understand what some decision analysts do well, a set of attributes that I have heard called "a good desk-side manner." For example, he describes how to draw a simplified or skeleton decision tree before eliciting all the detailed information needed to draw a complete tree. He presents friendly methods for eliciting probabilities. His examples from his own work also show that he knows how to tell a decision maker or an expert that he or she is wrong. He describes his book as "the only book written by a practitioner for a practitioner."

Finally, Mr. Skinner gives me clues about why I have perceived decision analysis to be an insiders' club. As a woman in a predominately male field, I have developed a thick skin regarding slights, but decision analysis has always seemed to me to be less open than any other field in which I participate; this situation is in sharp contrast to the openness of engineering economics, where we welcome newcomers and typically elect them immediately as officers in our organizations.

The DAAG web page I quoted above describes who may become a member and attend the annual DAAG conference and contains the striking sentence: "no consultants or academics are allowed unless invited by the Chairman." I find myself thinking of Groucho Marx's statement that he didn't want to belong to any club that would have him, but instead I think that any club that doesn't want me must have secrets that I want to know. In any case, I feel better that my perception of decision analysis as an insiders' club was not mere paranoia on my part.

In conclusion, Mr. Skinner's book would not be useful as a textbook for engineering or other quantitatively oriented students. It is better as an overview than as a tool for learning. The book is very valuable as a glimpse into the seldom seen world of the actual use of decision analysis.