A review of leveraged sample selection in variation reduction projects

Abstract

Variation reduction in critical-to-quality outputs is often the primary goal in manufacturing process improvement projects. Identifying the cause of output variation is recommended as an intermediate step in finding a low-cost sustainable solution to excessive variation. The goal of this paper is to describe and quantify the benefits of using leveraged sample selection when searching for important causes of output variation. We define leveraged sample selection as choosing parts to investigate that are extreme relative to other parts produced by the manufacturing process. In this paper, we discuss three different types of leveraged sample selection to illustrate the breadth of applicability of leveraging. We also review the existing literature and look at both planning and analysis of investigations that use leveraged sample selection. In addition, we provide a motivating example related to automotive headrests that illustrates the use of each of the three leveraged plans. Using leveraged sample selection in this context constitutes an example of the developing the discipline of Statistical Engineering.