Electronic statistics textbook banner

Statistical Advisor, Quality Control Charts

Use QUALITY CONTROL. This chapter discusses a wide selection of control charting methods.

In general, the purpose of quality control charts is to monitor the quality of an ongoing process. For example, in the manufacture of piston rings, one could take samples of piston rings and measure their diameters. One can continuously plot the means for successive samples that are taken. With the aid of statistical techniques (proper control limits), this will allow the quality control engineer to detect if and when the process mean (piston ring diameter) shifts outside acceptable limits.

There are several variations on this basic idea. Quality control charts can be produced for mean measurements (as in the example above), process variability (e.g., ranges, standard deviations), simultaneously for means and variability, or for attributes (e.g., number of defective items). Rather than plotting means, one can plot some moving average across samples or the cumulative sum of deviations from engineering specifications (CUSUM). One can also simultaneously monitor several measurements (T**2 chart)