© 2004 Cayman Business
Systems
Rev:
Pre-G3 - Rendered Thursday,
February 12, 2004
There is a wide range of non-random patterns that require action. When
the presence of a special
cause is suspected, the following actions should be taken (subject to local
instructions).
1. CHECK
Check that all calculations
and plots have been accurately completed, including those for control limits and means. When using variable
charts, check
that the pair (x bar, and R bar) are consistent. When satisfied that the data is accurate, act immediately.
2. INVESTIGATE
Investigate the process
operation to determine the cause.
Use tools such as:
Brainstorming
Cause and Effect
Pareto Analysis
Your investigation should cover issues such
as:
The method and tools for measurement
The staff involved (to identify any training needs
Time series, such as staff changes on particular days
of the week
Changes in material
Machine wear and maintenance
Mixed samples from different people or machines
Incorrect data, mistakenly or otherwise
Changes in the environment (humidity etc.)