Youden Address: Enhancing the Bottom-Line Impact of Statistical Methods In Vol. 15, No.2 Winter 1995 Statistics Division Newsletter
Abstract: The purpose of this paper is to stimulate interest within the statistics community to address the root causes for statistical methods not reaching their potential bottom-line impact. Evidence of the tremendous potential impact of statistical methods, and also of their lack of widespread achievement of this potential, is briefly reviewed. Some of the key root causes for this situation are identified as: overemphasis on the tools with a lack of understanding of how statistical concepts and methods apply at the strategic and managerial levels of an organization; attempts to jump to the most sophisticated methods without dealing with the required basic building blocks; and the general lack of understanding within the statistics community of the scientific method, at least as practiced by other scientists. Specific suggestions are given for addressing each of these issues. Implementing these suggestions will require understanding of basic frameworks from the field of organizational effectiveness, and of the manner in which the scientific method is actually applied in other sciences. Recent research by George Box is used to illustrate how mathematics is necessary, but not sufficient, to effectively apply the scientific method utilizing statistical methods.
Keywords: Statistical process control - Robust design - Inventory - Managerial systems - Scientific method - Bayesian - Statistical thinking - Systems thinking - Response surface methodology - Interactions - Control charts