Design for Six Sigma Statistics: 59 Tools for Diagnosing and Solving Problems in DFSS Initiatives

Citation

Narasimhan, K. (2009), "Design for Six Sigma Statistics: 59 Tools for Diagnosing and Solving Problems in DFSS Initiatives", Managing Service Quality: An International Journal, Vol. 19 No. 6, pp. 747-748. https://doi.org/10.1108/09604520911005107

Andrew Sleeper, general manager of Successful Statistics, LLC, is a licensed professional engineer and a Design for Six Sigma (DFSS) expert; he has worked in that field for 22 years in various capacities. His clients include Anheuser‐Busch, and Ingersoll‐Rand.

The book comprises 11 chapters and an Appendix containing 14 statistical tables. Chapters are titled according to the results to be attained using the tools, for example, Chapter 2 is about “Visualizing data”, and Chapter 6 is about “measuring Process Capability”.

Chapter 1 introduces the terminology of DFSS and the 59 tools and their purpose are tabulated. The power of DFSS tools is illustrated with an example of a developing a new fuel injector.

The next four chapters deal with ways to visualize and estimate the “voice of a process” (that is, the random behaviour of a process). Visualising data is the topic of Chapter 2, which discusses the use of time series graphs, distribution graphs, scatter plots and multivariate graphs and ends with a list of guidelines for ensuring graphical integrity. Chapter 3 introduces the concepts of probability and random variables. Chapter 4 presents tools to estimate characteristics of populations based on an analysis of samples and provides tips for selecting the most appropriate tool for different distributions such as normal, exponential, binomial and Poisson. Chapter 5, on Assessing Measurement Systems, covers methods and procedures such as control charts, variable gage repeatability and reproducibility studies to assess measurement system precision and biases, when using a go or a no‐go gage.

Chapter 6 focuses on measuring process capability to evaluate the variation of a process to determine if the process is capable of meeting the targets and tolerances specified by a customer. First, guidance is provided for selecting the most appropriate control charts for variables and attributes and interpreting them for signs of instability. This is followed by a discussion on some practical issues faced by Six Sigma practitioners in its application.

The next three chapters deal successively with detecting changes in the behaviour of processes producing normally distributed continuous data, processes producing discrete data and process producing non‐normal (for example, exponential and Weibull) distributions. Chapter 7 also introduces procedures for hypothesis tests and explains the two forms of errors that can occur. In Chapter 9, the author emphasises that the “goodness‐of‐fit” tests only prove badness‐of‐fit and not the former.

Chapter 10 describes several strategies for planning, conducting, and analysing efficient experiments when two or more factors (or inputs to a process) are involved. The first section explains the major concepts in the design and analysis of simple but efficient experiments with five examples. The next section explains the experimental terminology and provides a clear procedure for maximising the efficiency of experiments. The following two sections provide successively more information on design and analysis of two‐level and three‐level experiments. The final section shows how to improve robustness with experiments.

The final chapter presents tools to determine and select CTQ (critical to quality) characteristics and tolerances before starting tolerance design, predicting the effects of tolerances in linear and non‐linear systems, and optimising system variation.

The text is well supported by 415 figures, 103 tables, 215 examples, and 62 “Boxed texts”: How to use Minitab/Excel (36) and “Learn more about” (26) that explains concepts and techniques in greater detail. As the author points out, this book addresses the needs of practitioners and explains when each of the tools is appropriate and what to do if the tool is in appropriate.