Using Deming's Funnel Experiment to Demonstrate Effects of Violating Assumptions Underlying Shewhart's Control Charts

Abstract

Deming's funnel experiment is used to demonstrate the effect of blind use of Shewhart's and R charts for process data that violate at least one of the assumptions underlying their correct application. The funnel experiment is a useful illustration for quality trainers to demonstrate the need to check the assumptions behind Shewhart charts. Simple graphical methods of checking the assumptions are given. How to correctly apply Shewhart charts to the funnel experiment data is discussed and an application is used to illustrate a solution. This article also outlines how the funnel experiment could be used for training in the correct use of statistical process control charts.

Key Words:

• Assumptions
• Autocorrelation
• Funnel experiment
• Heterogeneity
• Independence
• Shewhart charts
• Statistical process control