It is standard practice to use joint charts in process control, one designed to detect shifts in the mean and the other to detect changes in the variance of the process. In this paper, we propose the use of a single chart to control both mean and variance. Based on the noncentral chi square statistic, the single chart is faster in detecting shifts in the mean and increases in variance than its competitor, the joint "Xbar" and R charts. The noncentral chi square statistic can also be used with the EWMA procedure, particularly in the detection of small mean shifts, accompanied or not by slight increases in variance.
noncentral chi square chart; joint "Xbar" and R charts; average run length
