Resumen de The Result Of Not Adjusting Shewhart X Control Charts For Positively Autorrelated Data
Howard R. Clayton, Victor R. Prybutok, Martha M. Harvey
This simulation study examines the performance of the Shewhart x control chart without adjustment for autocorrelation for both traditional fixed interval charts and variable interval charts. We also vary the interval length, autocorrelation parameter,and shift in the process mean. The results suggest that failure to diagnose the autocorrelation has potentially serious implications in practica! Applications because when the chart is unadjusted for autocorrelation it gives erroneous ATS values.
Furthermore the magnitude of the difference between the ATS for unadjusted charts versus the results that are expected for non-autocorrelated data is sufficiently large that it is of practica! significance. In addition our results show that the average time to signa! is related to the sampling interval, shift magnitude, and the amount of autocorrelation.
