Boris,
there are lots of ways to use different statistically calculated
numbers. I am suspicious, nonetheless, that your concept of a plant wide
goal for Cpk, either as a point estimate or as a confidence interval,
will not let you reach the larger goal you seek.
One can manipulate the math to show that an SPC chart control limits are
mathematically equivalent to a Student 't' test. the interpretation of
certain terms is not identical, exactly, but who cares? (slap your face,
Jay.)
the Cp and Cpk can be mathematically adjusted to show the math
equivalence. Without going into the details, I believe you can show that
The Cpk, CI, and 't' values are closely related.
use of point estimates of Cp and Cpk as yardsticks or standards or goals
tends to ignore the great sensitivity of the values to variations in the
standard deviation. Translation: you would be much better off to use
archival documentary data which comes from 50-100 measurements to get
decent estimates of the standard deviation. This leads to questions of
process variability and product variability.
In turn, these variability's will lead (eventually) to a question of
customer needs, as interpreted by (engineering, marketing, CEO,
whoever). Finally all this number pushing gets down to the real issues!
Not every dimension can be easily or precisely measured. Not every
dimension needs the same Cpk. So much for plant wide values.
I recommend you put your key focus on the issues that really count -
customer requirements, as interpreted. Let these determine production
requirements.
Cheers,
Jay
Boris wrote:
Hi, Do anyone there have an experience to set organizational (plant)
Cpk goals using Confidence Interval and/or hypothesis testing?
Most places use just point estimates for Cpk but in the liturature
(like classic Montgomery SPC book) confidence interval approach is
described.
I'd like to hear about such Cpk assessment system, how does it work in
practice.
Boris.
=
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at
http://jse.stat.ncsu.edu/
=
--
Jay Warner
Principal Scientist
Warner Consulting, Inc.
North Green Bay Road
Racine, WI 53404-1216
USA
Ph: (262) 634-9100
FAX: (262) 681-1133
email: [EMAIL PROTECTED]
web: http://www.a2q.com
The A2Q Method (tm) -- What do you want to improve today?
=
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at
http://jse.stat.ncsu.edu/
=
