Sometimes examination of control chart data reveals information
that affects other points that are not necessarily outside
the control limits. For example, if we had found that the temporary
operator working when sample 23 was obtained was
actually working during the entire two-hour period in which
samples 21–24 were obtained, then we should discard all four
samples, even if only sample 21 exceeded the control limits, on
the grounds that this inexperienced operator probably had
some adverse influence on the fraction nonconforming during
the entire period.
Before we conclude that the process is in control at this
level, we could examine the remaining 28 samples for runs
and other nonrandom patterns. The largest run is one of
length 5 above the center line, and there are no obvious patterns
present in the data. There is no strong evidence of anything
other than a random pattern of variation about the center
line.
We conclude that the process is in control at the level p =
0.2150 and that the revised control limits should be adopted for
monitoring current production. However, we note that
although the process is in control, the fraction nonconforming
is much too high. That is, the process is operating in a stable
manner, and no unusual operator-controllable problems are
present. It is unlikely that the process quality can be improved
by action at the workforce level. The nonconforming cans produced
are management controllable because an intervention
by management in the process will be required to improve