A cost based strategy for assessing improvements in manufacturing processes

Over the past decade there has been increasing research in the use of statistics in manufacturing, but there is continuing evidence of poor understanding in industry of even simple statistical concepts. In this paper we use a relatively straightforward application-hypothesis testing to assess the effectiveness of a process improvement-to show that standard statistical concepts are not easily applied to manufacturing measures. We develop an alternative approach that achieves the same objective as a conventional hypothesis test but uses concepts and parameters that are familiar to manufacturing engineers. This leads to simple algorithms for determining a test criterion and a sample size, and in passing illustrates that conventional significance levels may be inappropriate for manufacturing applications.