November 28, 2022
 Quality Applications SPC Guide First Word Last Word

Michael J. Cleary, Ph.D.
mcleary@qualitydigest.com

Inspired by the narcissism displayed in the California gubernatorial recall election, Hartford Simsack is now willing to acknowledge that the person he most admires in the world, above all public and private figures, is himself. “I’m an individual, and what could be more important than that?” he assures himself as he looks in the mirror each morning.

Even in training a new team of inspectors at Greer Grate & Gate, Simsack is aware that the most important aspect of his presentation is how well he comes across. He makes eye contact with the mirror in the back of the training room and speaks over the heads of the trainees with the sure knowledge that his audience considers him No. 1 when it comes to statistical process control.

With this intense focus, it comes as no surprise that Simsack’s favorite chart is the individual moving range chart. He recommends it for a variety of applications, including batch processes, such as those used at Greer Grate & Gate for mixing paint used for wrought-iron fencing. In his current training he offers an example of an individual moving range chart to impress his audience with its importance.

This chart indicates individual values that are in control or stable. However, one of the moving ranges is out of control, eliciting a question from one of Simsack’s trainees. “Don’t worry about that one point,” he assures the participant. “The individual points are all in control, and that’s what’s most important because this is an individual moving range chart.”

Was Simsack’s response correct, or was he mixing statistics with the psychology of ego?

The answer is no. Intent on improving his own image, Simsack has forgotten that control limits for the individuals section are derived by using the moving range section of the chart:

If there is an incorrect number, the control charts will be invalid, making it impossible to know whether the individuals are in control or not. Simsack must determine the cause for the out-of-control moving range, then eliminate that moving range and calculate the control limits.

In the case of a moving range above the upper control limit, the moving range average will be larger. This would lead to an inflated value for the moving range and create wider control limits. What might follow is an out-of-control point in the individuals section of the chart, but the control limits wouldn’t detect this because they are so wide.