Abstract
A procedure is proposed for use in evaluating the effect of various control chart procedures. This consists of breaking the power curve associated with a particular control chart technique into four component parts. These are:
γ1 — probability of undercorrecting
γ2 — probability of overcorrecting, but still improving
γ3 — probability of overcorrecting in excess
γ4 — probability of adjusting in the wrong direction
These components will be illustrated for a number of situations including Shewhart charts, partial correction procedures, geometrically weighted moving averages, cumulative sums, and an example which includes variability associated with the adjustments and variability in the process before adjustment.
Additional information
Notes on contributors
J. Edward Jackson
Dr. J. Edward Jackson is a staff consultant with the Eastman Kodak Company. He is a fellow of ASQC.