Statistics for quantity: Control and Capability
Comments on statistical control
• Having seen how x and s (or x and R) charts work, we can turn to some important
comments and cautions about statistical control in practice.
• Focus on the process rather than on the product This is perhaps the fundamental idea
in statistical process control. We might attempt to attain high quality by careful
inspection of the ﬁnished product or reviewing every outgoing invoice and expense
• Inspection of ﬁnished products can ensure good quality, but it is expensive. Perhaps
more important, ﬁnal inspection often comes tool ate:
• when something goes wrong early in a process, much bad product may be produced
before ﬁnal inspection discovers the problem. This adds to the expense, because the
bad product must then be scrapped or reworked.
• The small samples that are the basis of control charts are intended to monitor the
process at key points, not to ensure the quality of the particular it eosin the samples.
• If the process is kept in control, we know what to expect in the ﬁnished product. We
want to do it right the ﬁrst time, not inspect and ﬁx ﬁnished product. Choosing the “key
points” at which we will measure and monitor the process is important.
• The choice requires that you understand the process well enough to know where
problems are likely to arise.
• Flowcharts and cause-and effect diagrams can help. It should be clear that control
charts that monitor only the ﬁnal output are often not the best choice.
• The interpretation of control charts depends on the distinction between x-type special causes
and s-type special causes. This distinction in turn depends on how we choose the samples from
which we calculate s (or R).
• We want the variation within a sample to reﬂect only the item-to-item chance variation that
(when in control) results from many smal