A Comparison of dispersion quality control charts

The problem of detecting shifts in the process variability has not received as much attention as that of detecting shifts in the process mean of continuous production processes, even though it is important in the context of quality control. We examine and compare shewhart s (1931) and Page s (1963) procedures for detecting shifts in variability based on the sample range (R-charts), with procedures based on the sample standard deviation (S-charts). The underlying process control variables are assumed to be normally distributed. We also compare Cumulative Sum (CUSUM) procedures using sample ranges and sample standard deviations. We shall show by average run length (ARL) comparisons that procedures based on the sample standard deviation detect shifts from target value more quickly than procedures based on the sample range.