Globalizing Statistical Process Control and Restoring Shewhart's Methodology

Abstract

Many of the operations management texts present statistical process control (SPC) in a similar manner. Control charts occur in three separate formats: xbar-charts and r-charts for variables, p-charts for attributes, and c-charts for defects per unit. While the types of charts are distinct, there is significant commonality in the analysis of charts. Students are taught to examine charts for evidence of assignable causes of variation versus common causes. For all of the kinds of charts, assignable causes are suggested by 1) observations outside of the control limits, 2) runs of observations above or below the chart center line, and 3) observed patterns in the observations. The presentation of statistical process control suggested here is different in two respects: 1)the varieties of control charts are presented in a more unified theoretical framework provided by ISO-9000 and ISO-9001, and 2) the suggested presentation is a partial restoration of Shewhart's original framework. Thus the suggested framework provides a global context for SOC and a pattern of uniformity in both the theory and analysis of process control charts.