Six Sigma Article Is Confusing
I read Robert Gnibus' article "Six Sigma's Missing Link" (November 2000, p. 77) with some interest and came away disappointed. Instead of providing a primer on how to calculate the sigma levels, the article was overly truncated to the point of being confusing and of no practical use.
After reading the article, I don't think I understand the process well enough to use the method given to calculate sigma levels from discrete processes. I have been characterizing and calculating process sigma values from continuous variable process data for the past 15 years, and I have also worked with discrete process data for control charts and taught control charting, so I am not a novice in this area.
Why are the mechanics of constructing a Z table in Excel labored over since Gnibus gives the direct calculation of the sigma value using an Excel function? The relation of Table 3 to the other tables and the sample calculations was not shown, so why was Table 3 included? What function does it fill in the article? What is the significance of the column and row headings in Table 3? How does Figure 4 relate to anything in the article? What does the notation on Figure 4 mean in terms of application?
The values in the sample calculations were not properly defined, and the origin and justification for the sample calculations were not given. Why is 1.5 sigma (I assume they are sigma units) added to the result of the Excel function twice? What value on the process sigma continuum is being calculated by these sample calculations?
Perhaps this article was a useful memory jogger to someone who already understands Gnibus' methodology, but it is useless to someone unfamiliar with his approach.
'QP' Should Publish More Basic Articles
I enjoyed the November 2000 article "Six Sigma's Missing Link" by Robert J. Gnibus (p. 77). I think QP provides a great help to members by including articles that review and/or clarify important quality concepts, tools and techniques (beyond statistical process control (SPC) formulas). There are many times when QP focuses on covering the latest developments in quality, such as knowledge management and business process reengineering. These are all good topics to cover because they help keep quality practitioners abreast of the latest developments. However, there are a number of questions that I am sure many ASQ members have that go unanswered:
What is the correct number for a sample size?
Is it beneficial to change suppliers?
How can I determine if a measurement process is adequate?
What would be a confidence interval for a capability index and for percentages and rates?
How can I statistically compare the averages of two populations, such as processes, products or measurements?
Is it okay to run SPC charts and capability studies for non-normal populations? If not, what are some other options?
How can I verify/measure normality?
QP has sometimes published innovative applications of design of experiment and failure mode effects analysis, but I think it would be more helpful if QP provided more practical examples that have a widespread application. I think ASQ's general membership would benefit tremendously from coverage of these intermediate level quality concepts because even the "Statistics Roundtable" column is sometimes too advanced for the average reader.
MARIO A. VARELA, CQE, CQA, CRE and
certified quality manager
New Salary Survey Format a Hit With Reader
I just wanted to let you know that I like the new salary survey format (November 2000, p. 31). I know that, in the past, ASQ was trying to keep its members informed, but too much information is not necessarily a good thing.
The current survey provides just the right amount of information. Giving readers the ability to look up the rest of the data on the Web is a plus for those who wish to peruse it.
ISO 9001:2000 and Baldrige Should Be Kept Separate
I read Hampton Scott Tonk's article "Integrating ISO 9001:2000 and the Baldrige Criteria" (August 2000, p. 51) and believe that it made a limited contribution to understanding the similarities and differences between the present Malcolm Baldrige National Quality Award and an earlier version of the proposed ISO 9001:2000 quality standard. The article failed to present the full range of reasons and requirements for the ISO 9000:2000 group of quality standards.
Tonk's first of two perceived weaknesses for ISO 9001:2000 is that "There is no requirement or set of requirements to measure the impact of the ISO 9001:2000 system on organizational or business unit marketing or financial and profitability results." This statement ignores the ISO 9001:2000 requirement to establish organizational objectives that support quality policy. In establishing these objectives, organizations would use measurable goals, including those for marketing, processes, and continual improvement in customer satisfaction and profits. ISO 9001:2000 serves as a worldwide generic model of a minimum but complete quality system.
Tonk also cites a second major weakness in the ISO 9000 series. He states that "There is no requirement or set of requirements to measure the impact of the ISO 9000:2000 system on overall strategic planning, or vice versa." It must be pointed out that by using the term "ISO 9000 system," Tonk is invoking all three of the new ISO 9000:2000 quality standards:
ISO 9000:2000--Quality Management Systems--Fundamentals and Vocab-ulary.
ISO 9001:2000--Quality Management Systems--Requirements.
ISO 9004:2000--Quality Management Systems--Guidelines for Perform-ance Improvements.
The documents are meant to be used together, and they fully support strategic planning. It should be noted that the ISO 9000 Fundamentals and Vocabulary standard defines "quality planning" as "that part of quality management focused on setting quality objectives and specifying necessary operational processes and related resources to fulfill the quality objectives."
As a certified lead auditor, my experience indicates that it requires a substantial effort by smaller organizations to achieve a certifiable ISO 9001 quality system. To further increase quality requirements by combining the Baldrige criteria with ISO 9001:2000 will unnecessarily complicate, jeopardize and add cost to the present American ISO 9000 quality effort. The two standards should be kept separate for now.
EDGAR W. DAWES
Edgar Dawes makes some valid points, but he misunderstood the five purposes of my article. I tried to:
1. Influence the revision of ISO 9001:1994. In this purpose I failed, because the article was published too late relative to the ISO 9001:2000 revision timetable.
2. Show the differences and the similarities between ISO 9000 and the Baldrige criteria.
3. Demonstrate what a success total quality management has become. It has influenced the Baldrige criteria and the 2000 revision of ISO 9001.
4. Show those elements lacking in the Baldrige criteria vis-a-vis ISO 9000 and those elements lacking in ISO 9000 vis-a-vis the Baldrige criteria.
5. Demonstrate that an integrated ISO 9001:2000 based management system that incorporates the Baldrige criteria could be constructed.
I believe my article accomplished goals two through five, and Dawes' letter doesn't challenge this contention.
As to Dawes' specific comments here are my responses:
1. I did use the CD2 version of ISO 9001:2000 because of the speed with which the revision process proceeded relative to the time delay between the submission of my article and its publication date.
2. ISO 9001: 2000 doesn't purport to be anything but a quality focused management system. Thus it does not deal specifically, in the way MBNQA does, with strategic planning, financial management, profitability management or marketing management.
3. The article didn't discuss ISO 9004:2000 because that was beyond its scope. For the same reason it did not discuss QS-9000, ISO/TS 16949, TE 9000, TL 9000, AS 9000 or any other ISO 9000 derivative. However, it can be demonstrated that the Baldridge criteria can fit neatly into the clauses of those standards. I chose ISO 9001:2000 because it is, as Dawes states, the fundamental, generic standard to which organizations may register.
4. Dawes correctly makes the point that the table doesn't list all the clauses for MBNQA or ISO 9001:2000. There are clauses in both that have no equivalent in the other, and the article states that. Why are there such clauses with no equivalents in the other system? Because the systems are different. It's my opinion that quality systems will become increasingly integrated due to customer pressure, and they will become hybrid, customized, modular systems.
HAMPTON SCOTT TONK
A Lack of Enthusiasm In the Quality Field?
I have been in the quality industry for 18 years and have not noticed a cloud hovering over any businesses here in Australia. The cloud of quality J.E. Frazee ("Lack of Enthusiasm in Quality Field," October 2000, p. 14) referred to in her letter includes Quality System Requirements, known as QS, and the International Organization for Standardization, known as ISO.
My perception is that the majority of companies don't like the excessive documentation and paperwork usually involved with formal quality systems. But there are ways to avoid this dilemma: Write lean and simple operating procedures and work instructions, and reduce the quantity of quality records to a minimum by specifying what records need to be kept.
Quality systems should not be viewed as "another thing to do," but rather be welcomed because they provide a document system. This is virtually the starting point of the journey to achieve the ultimate goal of excellence in quality.
I do not believe that any companies think the QS-9000 and ISO 9000 systems are a joke. The example Frazee gave about a company that completely failed an internal audit, yet passed its certification requirements a few months later, could have happened under the following circumstances: If the company felt positive about quality management systems, the situation could have been that the internal audit findings were used to initiate changes in the system to make it more effective.
It is naive to believe that any external auditors were able to identify the major faults in a system because nobody does that better than those working full-time in the company. Assertiveness and enthusiasm about continually improving your quality system are the keys to success in the quality journey.
I appreciate your advice, but my systems are streamlined. Sure, many quality folks reading this rebuttal might think that my systems are too complicated or demanding. However, the companies that I work with have extremely lean procedures: 49 pages for being QS-9000/TE, ISO 9001 and soon ISO 14001 registered. Perhaps you got lucky and work for a company where quality is the number one priority.
I am not suggesting that I believe that quality is "another thing to do," but I am quoting exact statements from companies I have worked with. Companies do not make statements such as, "We got his letter that threatens us, and if we don't comply with this quality system stuff, we won't get any business. Can you help?" It's exciting until companies realize that a competitor who hasn't made any significant progress toward a quality system is still getting on the bid list.
As you state, "It is naive to believe that any external auditors were able to identify the major faults in a system because nobody does that better than those working full-time in the company." I agree, but isn't it customary for third-party auditors to review the internal audit results and actually check the implementation of those corrective actions? I do feel positive toward quality management systems, but let's get real--those internal audit findings that you are talking about "used to initiate changes." Yeah, those are called Band-Aids. It is very hard to be enthusiastic when that's all that senior management wants.
Most companies just want to get through the registration process no matter what. I applaud you if you work for a company that actually believes in the quality management process. I believe that until the hypocrisy changes and management actually sees benefits, there will be this cloud.
Clinton Township, MI
Misuse of Word 'Calibration' by Author
The article "Using Corrective Action To Make Matters Worse" (Wallace Davis III, October 2000, p. 56) discusses scheduled calibration as an example of tampering. I agree with the point, but take issue with the words. Davis used the word "calibration" in the popular sense, but one that is incorrect from metrology's viewpoint.
ANSI/NCSL Z540-1-1994 defines "calibration" as the work of verifying the performance specification of a measuring instrument. It's a documented method of determining the actual performance of the instrument in response to known inputs. The article refers to a case study of the required monthly functional check of a pressure switch--that is a calibration. This type of calibration, performance verification, should be done at regular intervals and is a foundation of a sound measurement system. Adjustment of the instrument should only be performed when the need is indicated by the data. Adjustment at any other time is tampering with the process.
I believe that the word "adjustment" would be more appropriate in this article instead of "calibration." Then it would be clear the inappropriate adjustment of a stable instrument is the real culprit.
In the case described, the valve should be left alone because its operational range is stable and (presumably) within the specified limits. The procedure that requires tampering with a stable process should be reviewed. I assume that the range of 43 to 50 psi shown in Figure 5 is within the specification limits, from the statements that the procedure requires adjustment if it's outside the limits, and that there aren't any unscheduled adjustments in the data.
GRAEME C. PAYNE, CQT, CQE
Philosophically I am in complete agreement with everything Payne states. Unfortunately, my experience leads me to believe there is a disconnect between what most people understand the word "calibration" to mean and what ANSI/NCSL Z540-1-1994 states.
I used the popular definition of the term "calibration" in my article and then went on to criticize the concept. I did urge the company in the pressure switch example to modify its procedures to call the functional checks "calibration" and perform adjustment of instrumentation only when out of tolerance. The change was not made because it required Nuclear Regulatory Commission (NRC) approval, and the NRC inspectors held to the classical belief that calibration required adjustment of the instrument back to the setpoint value if the measured condition differed.
WALLACE DAVIS III