16
STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL Volume 12, No. 3 _ Spring, 1992 Not to be outdone by the Presidential primaries all around us, ASQC is hav- ing elections too. In late February you probably got your annual ballot from Headquarters. In early March you prob- ably received a postcard explaining that the label you got with your ballot was missing critical information—namely your Section and Division numbers, which you needed to write onto your ballot to assure its validity. Tell me—what did you do when you got your ballot? Did you get out your magnifying glass and scan your label for the missing data? Or perhaps you keep good records, so you can look up your ASQC numbers when- ever you need them. Maybe you've memorized them. What I did was to get confused. I phoned Headquarters and got the needed info over the phone. Given the state of things, that was nice. But think of the cost, in time, confusion, and (heaven forbid) irritation, not to men- tion the cost of 85,000 postcards. Let's do a calculation. Suppose each member was confused for one minute. That's one minute times 85,000 members, for a total of 85,000 minutes, or about 1,417 hours. Dividing by an average of 168 hours of working time per month, the result is about 8.4 months of working time lost to confusion over the ballot. Supposing each member was confused for two minutes, the result is about 16.9 months of working time lost. Now suppose each member was confused for two min- utes and made a two-minute phone call to Headquarters. That's 16.9 months lost to initial confu- sion, plus 33.7 months on the phone! (Remember, there's two people involved in each phone call; it's your time plus the operator's time.) The grand total for this modest scenario is 510,000 minutes, or 8,500 hours, or 50.6 working months, or 4.2 years—spent on wonder- ing about the ballot. My point is not to make fun of Headquarters. The point is that in large-scale processes, errors that occur anywhere will be multiplied a thousandfold. The eco- nomic leverage of rapid manufacturing carries with it the downside risk of producing many more defects ..--much faster. Corrective action after the fact is not a solution—the damage has already occurred. "Putting preventive sys- ■■•1.11.• CHAIR'S MESSAGE by Conrad Fung terns in place" by tacking new procedures onto old to patch up quality loopholes, is not a solution either. This has the side effect of attaching burdensome administra- tion to value-adding work. It's time to revisit two classic papers on this subject. The first is by Tim Fuller, "Eliminating Complexity from Work: Improving Productivity by Enhancing Quality" (National Productivity Review, 1985). In this paper, Tim showed how a major part of work time is spent on finding and fixing errors—undoing what should not have been done in the first place. Two sur- veys described in the paper found that only 35.0% and 42.6% of available work time, respectively, was spent on value-adding ("real") work. Also in this paper, Tim proposed a concept of Complexity to help understand non-real work, namely the extra process steps that are needed to handle errors. Complexity, for example, could include a process to check whether parts needed for an assembly have all arrived; processes for writing down which parts are missing; for storing incomplete assem- blies until the parts do arrive; for shipping a partial assembly to the customer while they wait for the parts; for getting a supervisor's signature to release such an assembly; for following up with the customer when the parts do arrive; etc., etc. The best way to deal with errors is not to have them. Another paper worth re-reading is "When Murphy Speaks—Listen" by George Box (Quality Progress, October 1989). This paper makes the point that Murphy's Law (that whatever can go wrong, will go wrong) is good news, rather than bad—because if we take note of what has gone wrong, and improve the sys- tem so it doesn't happen again, the system will continu- ously improve. An amusing incident with this paper is that George's name was mistakenly left off of the title page when it went to press. But that just proves the point. The Annual Quality Congress in Nashville is almost upon us. Please come to Statistics Division's Annual Business Meeting at the Opryland Hotel on Monday, May 18, at 5:30-7:00 pm. We'll review the State of the Division, the year's activities and tactical plans, and ask your input on what we should do for the future. The exact room will be posted at the hotel. We'll pro- vide hors d'oeuvres and refreshments.

•1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

Page 1: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

STATISTICS DIVISION

NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

Volume 12, No. 3 _ Spring, 1992

Not to be outdone by the Presidential primaries all around us, ASQC is hav- ing elections too. In late February you probably got your annual ballot from Headquarters. In early March you prob- ably received a postcard explaining that the label you got with your ballot was missing critical information—namely your Section and Division numbers, which you needed to write onto your ballot to assure its validity. Tell me—what did you do when you got your ballot?

Did you get out your magnifying glass and scan your label for the missing data? Or perhaps you keep good records, so you can look up your ASQC numbers when-ever you need them. Maybe you've memorized them. What I did was to get confused. I phoned Headquarters and got the needed info over the phone. Given the state of things, that was nice. But think of the cost, in time, confusion, and (heaven forbid) irritation, not to men-tion the cost of 85,000 postcards.

Let's do a calculation. Suppose each member was confused for one minute. That's one minute times 85,000 members, for a total of 85,000 minutes, or about 1,417 hours. Dividing by an average of 168 hours of working time per month, the result is about 8.4 months of working time lost to confusion over the ballot. Supposing each member was confused for two minutes, the result is about 16.9 months of working time lost. Now suppose each member was confused for two min-utes and made a two-minute phone call to Headquarters. That's 16.9 months lost to initial confu-sion, plus 33.7 months on the phone! (Remember, there's two people involved in each phone call; it's your time plus the operator's time.) The grand total for this modest scenario is 510,000 minutes, or 8,500 hours, or 50.6 working months, or 4.2 years—spent on wonder-ing about the ballot.

My point is not to make fun of Headquarters. The point is that in large-scale processes, errors that occur anywhere will be multiplied a thousandfold. The eco-nomic leverage of rapid manufacturing carries with it the downside risk of producing many more defects

..--much faster. Corrective action after the fact is not a solution—the

damage has already occurred. "Putting preventive sys-

■■•■•■•1.11.•

CHAIR'S MESSAGE by Conrad Fung

terns in place" by tacking new procedures onto old to patch up quality loopholes, is not a solution either. This has the side effect of attaching burdensome administra-tion to value-adding work.

It's time to revisit two classic papers on this subject. The first is by Tim Fuller, "Eliminating Complexity from Work: Improving Productivity by Enhancing Quality" (National Productivity Review, 1985). In this paper, Tim showed how a major part of work time is spent on finding and fixing errors—undoing what should not have been done in the first place. Two sur-veys described in the paper found that only 35.0% and 42.6% of available work time, respectively, was spent on value-adding ("real") work. Also in this paper, Tim proposed a concept of Complexity to help understand non-real work, namely the extra process steps that are needed to handle errors. Complexity, for example, could include a process to check whether parts needed for an assembly have all arrived; processes for writing down which parts are missing; for storing incomplete assem-blies until the parts do arrive; for shipping a partial assembly to the customer while they wait for the parts; for getting a supervisor's signature to release such an assembly; for following up with the customer when the parts do arrive; etc., etc. The best way to deal with errors is not to have them.

Another paper worth re-reading is "When Murphy Speaks—Listen" by George Box (Quality Progress, October 1989). This paper makes the point that Murphy's Law (that whatever can go wrong, will go wrong) is good news, rather than bad—because if we take note of what has gone wrong, and improve the sys-tem so it doesn't happen again, the system will continu-ously improve. An amusing incident with this paper is that George's name was mistakenly left off of the title page when it went to press. But that just proves the point.

The Annual Quality Congress in Nashville is almost upon us. Please come to Statistics Division's Annual Business Meeting at the Opryland Hotel on Monday, May 18, at 5:30-7:00 pm. We'll review the State of the Division, the year's activities and tactical plans, and ask your input on what we should do for the future. The exact room will be posted at the hotel. We'll pro-vide hors d'oeuvres and refreshments.

Page 2: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

LETTER FROM THE EDITOR ASQC STATISTICS DIVISION OFFICERS

Chair Conrad A. Fung Center for Quality and Productivity Improvement

University of Wisconsin WARF/1109B 610 Walnut Street Madison, WI 53705 (608) 263-2652

Chair-Elect Joseph G. Voelkel Rochester Inst. of Technology Center for Quality and Applied Statistics 1 Lomb Memorial Dr., Bldg.18 P.O. Box 9887 Rochester, NY 14623-0887 (716) 475-2231

Secretary Galen C. Britz 3M Center 220-9W St. Paul, MN 55144 (612) 736-6499

Treasurer Richard A. Lewis Monsanto Chemical Company 800 N. Lindbergh Boulevard St. Louis, MO 63167 (314) 694-7735

EDITOR'S CORNER Deadline for the Summer Issue is May 30, 1992. Send all letters EXCEPT FOR CHANGE OF ADDRESS to: Nancy Baxter Thomas J. Lipton Co. 800 Sylvan Avenue C-210 Englewood Cliffs, NJ 07632

Communications regarding change of address should be sent to ASQC. This will change the address for all publications you receive from ASQC including the newsletter. You can handle this by phone (414) 272-8575 or (800) 248-1946.

Dear Readers: The verbatim comments from the

membership survey have been sum-marized (see article elsewhere in this issue). Some respondents liked the technical content of the newsletter while others found it too technical. We have been trying to address this topic by printing two columns: Basic Tools and Mini-Paper. What do you think ?

The membership survey high-lighted some topics that readers may want to learn more about. Many respondents were not famil-iar with affinity diagrams. This technique is presented in the Basic

Dear Nancy, I am concerned about the contin-

uing decline in the pass rate on the Certified Quality Engineer Exam. It is currently (I am told) about 25%. The Cleveland Section and other sections, offer many courses to help members become more proficient in the body of knowledge required for their chosen careers, and to become certified, i.e., recognized by their peers as experts in their field. The Certification process is a combina-tion of experience, education, and examination. The examination pro-cess appears to be the greatest obstacle. Many people take the examination and do not pass it, and can't understand why. Under pres-sure to become certified, many take a few courses and expect to become engineers. Many have only moder-ate, if any at all, knowledge of the mathematics necessary to become quality engineers. I am not talking about sophisticated statistical con-cepts, such as ANOVA or DOE, either classical or Taguchi. I am speaking herein of the basics, such as central tendency, dispersion, or

Tools column. A version of the Malcolm

Baldrige Award article was print& in the Summer 1990 issue. The art cle has been updated and expanded since that time. Maureen Heaphy, the author, has been a member of the Board of Examiner and assisted in the development of the 1992 cri-teria. The article will be of primary interest to those readers whose companies are applying for the Award. Other companies and orga-nizations are using the criteria to review and guide their quality pro-grams

Nancy

the shape of distributions, and how these concepts might be used to determine a slightly more complex idea such as Cpk. The symbolism used for such ideas: mu (.t), sigma (a), and Cpk, etc. cause alarm in those not accustomed to using mathematical symbols.

What can be done about it? Let's consider what it is we are asking people to do. I will only consider one = --concept, mu (p.) or central tendency, because of limited space. The per-son considering this concept and its application to probability distribu-tions and process capability—to name only two of its uses—must bring with him/her a storehouse of "intellectual baggage" to enable him/her to understand the concepts and practical applications which he/she is certain to encounter. Consider how mu (p.) might be taught by "knowledgeable" teach-ers. Mu (p..) is an average or center of the "population", represented on a graph; most of the time we don't know mu (4), because the population

Continued on page 3

LETTERS TO THE EDITOR

MISSION • Promote Statistical Thinking for quality and productivity improve-ment. • Serve ASQC, business and industry, academia, and govern-ment as a resource for effective use of statistical methods for quality and produc-tivity improvement. • Provide a focal point within ASQC for problem-driven develop-ment and effective use of new statistical methods. • Support the growth and development of Division members.

PAGE 2

STATISTICS DIVISION NEWSLE

SPRING, 1992

Page 3: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

Continued on page 4

Letters Continued from page 2 is too large. We then have to esti-mate this population parameter with a statistic based on a sample from the larger population, which is probably not the same as mu 40, or maybe it is, we'll call it x-double bar (1) or x-bar (I). We can calculate with varying degrees of confidence that it will probably fall within some range. This is admittedly an over-simplification meant only as an example of what might be neces-sary to know when learning a con-cept such as mu (u). What is really meant by all of this talk about mu (g)? What is the underlying "bag of knowledge" that we expect learners to bring with them in order to understand what we mean? Should we be satisfied if the learner can apply mu (p) with little understand-ing of its meaning? Whatever it is that we expect, we do not seem to be getting it across enough of the time; just look at the state of American Business.

As teachers, our assumption is often that students come prepared to learn new concepts. Much of the time they don't. It is crucial that the 'reaching of such concepts be done in a manner which enables those learning the concepts to understand them in detail, so they might use them to solve problems in the future. It seems that it is necessary to reteach basic mathematics each time we teach some new or more complex statistical concept. The sad part is that this mathematics edu-cation has already been done before. What is the cost to industry and 'society to train people in the basics (AGAIN)? As Deming says, "...it is unknown and unknowable". What we do know is that the manufactur-er or service organization must pass some of the cost of retraining to the customer. I know of companies reteaching reading, writing, and mathematics to their employees in order to make them more produc-tive within their organization and more able to cope within society. What is wrong? There seems to be enough blame to go around. But it is pointless to blame anyone. The situation must be improved, that's

11. How do we do it? We as educa-

tors of statistical concepts are charged with the task of passing on knowledge about statistics and the application of statistics to the quali-ty function. It has been my experi-ence that teachers have knowledge, but is having knowledge enough? Most "teacher types", especially those trained in the sciences, believe that they can stand in front of a class and impart knowledge by some magical process. That is instructing, not teaching. These teachers labor under two miscon-ceptions. One, that anyone with knowledge can teach. Two, if people don't understand what the teacher is trying to teach them, then the students are stupid or lazy. We hear, "That's the way I learned! Why can't they?" There seems to be some form of Ego/Knowledge prob-lem on the part of such teacher types. They seem genuinely sur-prised and irritated when students give them poor evaluations, because they (the students) did not think that they learned anything. The teachers then blame the students for their (the teacher's) inability to pass on knowledge.

Many teachers teach as if all peo-ple learn the same way, i.e., both theoretical and practical concepts are taught as if they are equivalent. When teaching probability and statistics, we must keep in mind that there is a great deal of differ-ence between theoretical and practi-cal thinkers. Most people are one or the other and rarely some are both types (read about theories of left and right brained processes). According to Deming and Shewhart, we have to first plan what it is that we are going to teach, and then fig-ure how we are going to do it. Can we hold the process of learning in any different regard than we do a process in the manufacturing or ser-vice industries? I think not! If we do, then we deserve what we get. If we want problem solvers, we must train them to be problem solvers. The way we are now teaching peo-ple is not working. Ask most employers. It is my guess, based on numerous conversations and discus-sions, that there is not a knowledge-able person in the field of quality, with a mathiscience/ engineering background, who has not asked, why is it that so many people enter-

ing the field have such inadequate backgrounds? I submit that part of the reason for this is, that there is at work, what I call an "ignorance cascade". The problem comes about, because we are in a hurry to impart knowledge at a quickened pace. We hurry through our work as teachers and sometimes we leave out impor-tant facts. New teachers are not completely prepared because of inadequate training. They pass knowledge to the next group of learners, but they leave out some additional detail or details. A result is that subsequent learners are less capable than those that preceded them. If such a problem were to continue through many genera-tions, it would not take long for peo-ple to be inadequately taught or trained. Look around, it seems plau-sible to me. I think this effect is cas-cading from the lowest levels of edu-cation to the highest levels.

The quick fix, Americans are great at quick fixes, is obviously remediation. But, there has to be some kind of long term aim. Our aim should be to help people to learn how to think with statistics as well as to think for themselves, to enable them to use methods that solve problems. Methods and theo-ries on how to accomplish this abound. My personal recommenda-tion is to apply the statistical meth-ods of W. Edwards Deming to this very important and difficult endeav-or. Our cultural and economic future depends upon choosing the best approach to improving our edu-cational system and the way we teach people at every level.

Wm. Mill) Orgel Cleveland Section-ASQC

Education Chairman

SPRING, 1992

STATISTICS DIVISION NEWSLETTER PAGE 3

Page 4: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

46TH ANNUAL QUALITY CONGRESS NASHVILLE, TENNESSEE• May 18-20, 1992* STATISTICS DIVISION SESSIONS

If you are unable to attend the AQC but would like copies of any of these papers, please contact the Newsletter Editor, Nancy Baxter, T.J. Lipton Co., 800 Sylvan Ave., Englewood Cliffs, NJ, 07632; (201) 894-4011, FAX: (201) 894-7860.

MONDAY, MAY 18 1:45-3:15 PM Session 1.1.8

"CuSum Charting for Management Applications" Tsong-how Chang, Associate Professor, University of Wisconsin-Milwaukee, Milwaukee, WI

A method of charting cumulative statistics for production and opera-tions management control will be introduced. Chart construction and interpretation are described and compared with conventional control charts.

"Accrediting a Machine for a Lifetime of Quality" Dave Middleton, Quality Assurance Engineer, Eastman Kodak, Rochester, NY

Letters to the Editor Continued from 3 Dear Ms. Baxter,

I have recently received 10 copies of your Winter issue of the Newsletter of the Statistics Division of the ASQC. This is a very welcome gesture; the copies sent have been circulated among our Committee members and among member of our associated Quality Forum.

One item that seemed particular-ly interesting was the article by William Lawton. Our Committee is seeking articles like this for submis-sion to the journal of the British Institute of Quality Assurance with which we have recently established a working relationship. Lawton's article seems exactly the kind of thing that we would like to see pub-lished in that journal.

Yours sincerely, P.R. Fisk, Secretary

Royal Statistical Society Committee on Quality

Improvement Edinburgh, Scotland

Designed experiments can be used to accredit machines for production, evaluating a lifetime of operating conditions better, faster, and cheap-er than would a traditional accredi-tation.

"Poisson Green M&M's -Attributes Simplified" Bill L. Wortman, Principal/Owner, Quality Council of Indiana, West Terre Haute, IN

A floor level review of key points in attribute sampling and charting will be presented. Principles are highlighted using bags of M&M's and historical data.

3:45-5:15 PM Session 1.2.8

"The Evolution of a System for Teaching EVOP" Thomas B. Barker, Associate Professor, Rochester Institute of Technology, Rochester NY

"A Management Procedure for Design of Experiments" Lai K. Chan, Professor, The

The Statistics Division is spon-soring a tutorial,"A Practical Introduction to Experimental Design", prior to the AQC on Sunday, May 17 in Nashville, Tennessee.

Experimental design is one of the most important tools for improving the quality of industrial products and processes. It is of strategic importance that any company aspir-ing to world-class quality status master these powerful methods.

This tutorial will provide a quick and practical overview of the key concepts of design of experiments, two-level and fractional factorials, and graphical methods for data analysis. You will also be shown how to use "A Practical Aid for Experimenters," a set of tables developed by the instructor which greatly simplifies the task of designing and analyzing experi-ments.

This tutorial requires little prior knowledge of statistics beyond a flair of numbers and an under-

University of Manitoba, Winnipeg, Manitoba, Canada; Eng Hui Gan, Senior Statisticiai Intel Technology Sdn. Bhd., Penang, Malaysia; and Tak K. Mak, Associate Professor, Concordia University, Montreal, Quebec, Canada

Proposes "DOE-7", a 7-step sys-tematic management procedure for design of experiments, and describes four variation-reduction / on-target DOE techniques.

TUESDAY, MAY 19 8:00-9:30 AM Session 2.1.8

"Self-Organizing Experimental Design for Quality" Huaming Zhai, Statistics & Quality Consultant, University of Wisconsin, Madison, WI

"A Responsible Critique/ Enhancement of Taguchi's Method" John S. Ramberg, Professor of

Continued on page 5

standing of graphs. Soren Bisgaard is an assistant

professor at the Center for Quality and Productivity Improvement and the Department of Industrial Engineering, University of Wisconsin-Madison. He holds a master of science degree and a doc-torate in statistics from the University of Wisconsin-Madison. He has received the Shewell and Brumbaugh Awards from the American Society for Quality Control and Ellis Ott Award.

Course material will include lec-ture notes and a copy of "A Practical Aid for Experimenters." The tutori-al begins at 8:30 a.m. and concludes at 5:00 p.m. Registration fee is $210 per person and includes materials, luncheon and breaks. To register for this tutorial contact the Conference and Exhibits Department at ASQC, (414) 272-8575. All phone registra-tions must include the number of a major credit card.

ACQ Pre-Conference Tutorial

STATISTICS DIVISION NEWSLETTER PAGE 4 SPRING, 1992

Page 5: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

46th Congress Continued from page 4

Systems & Industrial Engineering, University of Arizona, Tucson, AZ; and Joseph J. Pignatiello, Jr., Depart-ment of Industrial Engineering, Texas A&M University, College Station, TX

10:00-11:30 AM Session 2.2.8

"Simultaneous Engineering and Statistics" Cristobal Samaniego, Senior Statistician, Cadillac Motor Car, Detroit, MI

"Concurrent Engineering of a Service Order Process" James P. Pennell, Member of Technical Staff, AT&T Bell Laboratories, Vienna, VA; Daniel P. Petrozzo, Systems Engineering Consultant and John C. Stepper, Member of Technical Staff, AT&T Bell Laboratories, Middletown, NJ.

Presents a case study of a contin-uing effort to apply the concepts of

,__concurrent engineering while re- •ngineering a business process and support system.

"Process Optimization for Multi-Attributed Product" David Stevens, Senior Computing Engineer, Bell Helicopter Textron, Fort Worth, TX; and R.C. Baker, Professor of Information Systems and Manage. ment Sciences, University of Texas at Arlington, Arlington, TX

Describes a mathematical proce-dure that combines techniques from response surface methodology, opti-mization, quality control, and deci-sion theory in order to solve process optimization problems.

1:45-3:15 PM Session 2.3.8

Panel Discussion "Report of the Deming Study Group LIP Deming Study Group of Greater Detroit

In Detroit, MI, a group of statisti-ions and practitioners meets to dis-

cuss, with Dr. W.E. Deming's guid-

ance, topics that will clarify and increase the body of knowledge and philosophy of continual improve-ment. For the past year, the focus of the Study Group has been the appli-cation of Profound Knowledge to the educational system.; the aim was the "Joy of Learning Maintained." The panel will report the results of these discussions. Ample time will be allo-cated for questions and participant interaction.

3:45-5:15 PM Session 2.4.8

"C.I. Lewis' Theory of Knowledge Related to Quality Improvement" Mike Lovitt, Consultant, Provost Consulting Group, Inc., Austin, TX

Discusses C.I. Lewis' theory of knowledge, its influence on both Shewhart and Deming, and what it meant to organizations pursuing quality improvement.

"Quality Topics in the Business School Curriculum" David P. Doane, Professor, Oakland University, Rochester, MI

Sampling methods are used to assess the extent of quality topic cov-erage in business textbooks and classes, and to make comparisons over time.

"Content and Methodology for a Course in Experimental Design" Don Benbow, Professor of Mathematics, Marshalltown Com-munity College, Marshalltown, IA; and Roger Berger, Professor of Industrial & Manufacturing Science Engineering, Iowa State University, Ames, IA

Discusses the content and methodology for a course in experi-mental design for practicing engi-neers and technicians.

WEDNESDAY, MAY 20 8:00-9:30 AM Session 3.1.8

"A Simple Method for Estimating Process Spread" Richard M. Goodson, Quality Assurance Manager, Johnson Controls, Watertown, WI

A graphical technique for esti-mating process spread for both nor-

mally distributed and non-normally, distributed processes using the Weibull distribution.

"Comparing Sigma Estimates for Unstable Processes" Jennifer Mil-11;ns, Statistician; Kwan Lee, Senior Statistician; and Arved Harding, Practicing Statistician, Eastman Chemical Company, Kingsport, TN

Processes with trends, oscilla-tions, and sudden upsets are simu-lated. Estimates of sigma based on the moving range and sample stan-dard deviation are calculated and compared.

"White-Collar Statistical Process Control" Earl M. Mustonen, Director of Quality Operations; Ron Asher, Program Management Specialist; and Michael Puzzuoli, Manager of Quality Programs, General Dynamics Land Systems, Warren, MI

When effectively applied to administrative processes, SPC tech-niques can provide the foundation for better process measurement, management, and continuous improvement.

9:45-11:15 AM Session 3.2.8

"Using Statistical Tools to Meet the Challenge" Michael Mazu, Senior Statistician, Alcoa-Warrick Operations, New-burgh, IN

Illustrates how statistics can be used to select the best process-oper-ating conditions.

"Use of Probability Plots in Analyzing Designed Experi-ments" Robert F. Brewer, SPC Consultant, Industrial Design & Engineering, Lakeville, PA

Multiple probability plots provide information on the shape of the dis-tributions, which provides impor-tant clues for process improvement. This is further aided by designed experiments which provide the con-ditions under which the data were obtained.

Continued on page 6

SPRING, 1992 STATISTICS DIVISION NEWSLETTER PAGE 5

Page 6: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

The Statistics Division is seeking nominations for the William G. Hunter Award presented annually in order to encourage the creative application and development of sta-tistical thinking and methods for quality improvement.

Presented in memory of the Division's founding chairman, the award recognizes that person whose

46th Congress Continued from page 5

"The Role of Pure Error on Normal Probability Plots" Pat Whitcomb, President, STAT-EASE, Inc., Minneapolis,MN; and Kinley Larntz, Professor of Statistics, University of Minnesota, St. Paul, MN

Shows that augmenting the prob-ability plots of two level factorial effects with points representing pure error (estimated from replicate experiments) aids in model selection.

Session 3.2.9

"Misleading Signals: X-bar & R Charts vs EWMA Charts" Ralph C. St. John, Professor, and Daniel J. Bragg, Associate Professor, Bowling Green State University, Bowling Green, OH

"Multivariate Moving Average Control Charts" Thomas Margavio, Assistant Professor, Southwest Missouri State University, Springfield, MO; and Michael D. Conerly, Associate Professor, University of Alabama, Tuscaloosa, AL

Peter J. Jacobs Statistical Consulting Department 3M Center, Building 224-45-19 St. Paul, MN 55144 (612) 7334257

Completed forms should be returned to Pete no later than July 1, 1992.

Membership Survey Analysis, Continued:

Surprises, Verbatim Comments, Reflections

Surprises The first surprise was that the

survey firm we hired said that, of the 501 people they phoned, 500 responded! The firm was amazed at this percentage of callers staying on the line: they had never seen any-thing like it. Thank you, Stat Division members!

The second surprise was that a full 45% of the respondents called themselves managers (Figure 1, Winter Newsletter).

Verbatim. Comments The verbatim comments we

received from certain questions pro-vide useful information to help us focus on our customers.

The three broad questions we asked were: 1) Give two examples about what

you like about the Stat Division. (Zero or one answer OK)

Newsletter

39%* Don't know

21 Keeps you up to date

9 Newsletter articles

7 (14 other answers ...)

* Read "39% of respondents said they liked the newsletter."

2) Give two examples about what you dislike about the Stat Division. (Zero or one answer OK)

Nothing

44% Don't know

17 Articles too technical

6 (3 other answers ...)

3) Other issues to pass on to the Stat Division leadership. (Zero or one answer OK.)

Nothing

77% Good job

2 More applied/basic material 3 Nothing offered in my area 1 of the country

Survey good idea 1 Would like local refresher 1

course for CQE Correspondence courses 1 More affordable conferen-

ces/courses

Continued on page 7

Nominations Sought for 1992 William G. Hunter Award

William G. Hunter actions most closely mirror Bill Hunter's strengths, which were as a communicator, a consultant, an educator (especially for practition-ers), an innovator, an integrator (of statistics with other disciplines), and an implementor (who obtained results).

The award is presented annually at the Fall Technical Conference. The first such award was presented in 1988 to William J. Hill, Director of the University of Wisconsin's Center for Quality and Productivity Improvement. The award was pre-sented in 1989 to Rudolf G. Kittlitz, Research Associate for Du Pont Fibers and in 1990 to Ronald D. Snee, Senior Consultant for Joiner and Associates. The most recent award was presented in 1991 to Gerald J. Hahn, Manager of the Management Science and Statistics Program at the General Electric Company's Research and Development Center.

Nomination forms for this year's award can be obtained from any of the Statistics Division officers at the Annual Quality Congress in Nashville on May 18-20. Just stop by the Statistics Division booth dur-ing the day or the hospitality suite in the evening. Nomination forms can also be obtained directly from the William G. Hunter Award Committtee Chair:

STATISTICS DIVISION NEWSLETTER SPRING, 1992 PAGE 6

Page 7: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

Of the 20% of the surveyed mem-bers who are CQE-certified, 72% listed such certification as being "very" or "extremely' important, and another 19% as "somewhat" important. The 80% who are not CQE-certified placed less value in certification: 36% in the "very" or "extremely" category, and another 33% in the "somewhat" category.

This is how the certified mem-bers studied (more than one answer OK).

Section-sponsored re-fresher courses

60%

Other refresher courses 25 Books 22 Self-Study 12 Old tests 4 College courses 4 (3 other answers ...)

----- The extent to which members use . variety of statistical/quality tools, and whether they should use the tools more, was presented in the Winter Newsletter. For those who thought they should be using the tools more (along with the %'s who said they should), here are the major reasons why. We only include those tools here that at least 40% of the respondents said they should more.

Cause & Effect Diagrams (69% said they should use more)

Problem-Solving Tool 13%* Identify Root Causes 13 Process Understanding 8 Useful/Effective 7 Group/Team Tool 7 (7 other answers ...)

* Read "Of those who said they should use the tool more, 13% said it was because it is a problem-solv-ing tool."

Control Charts (46%)

Better process control/ 22 understanding

Increase # of variables 11 charted

Indicate on-going quality 10 Spot problems earlier 6 Improve quality 3 (many other answers ...) Two-level factorials (45%) Better process understand- 13 ing

Solve problems 9 Info. on importance of 7 different factors

(many other answers ...)

Flow Charts (42%)

Visualize process 43 Material flow 14 Standardize process 8 Problem analysis 7 (7 other answers ...) Taguchi Experiments (42%)

Solve problems

15 Info. on importance of

9

different factors Increase productivity

6 (many other answers ...)

Reflections

Overall, the objectives of the sur-vey were met. However, in retro-spect, we would do a few things dif-ferently.

Most importantly, we would have run a trial survey first. This would Likely have generated the following upgrades to the survey:

*Get more information on respondents who said they were managers. (How would we do this? Members who are man-agers: any ideas? Write to the editor: we need your input!)

*Directly query the members on aspects of the Newsletter. For example:

*What percent read what parts of the newsletter? *What about the technical con-tent of the articles (some of the respondents liked it, some said they were too technical)? *What else would you like to see included? What would you liked changed or dropped?

'Respondents occasionally con-

fused the Stat Division with ASQC as a whole. The survey could be modified to make this distinction clearer.

At the AQC in May, we will decide whether the first two or three items above are important enough to warrant a follow-up sur-vey. If so, it would likely be done by mail (this one was done by tele-phone).

GRANTS FOR BEGINNING RESEARCHERS OFFERED

Through a generous donation from the Statistics Division of the American Society for Quality Control, six grants are available to pay for registration, lodging and food for beginning researchers to attend the 1992 Gordon Research Conference on Statistics in Chemistry and Chemical Engineering (usual cost is $390). Grant recipients will be responsible for their own travel expenses. These grants are available to individuals who have received their highest degree since May 1989 or who are still in school and who are involved in research in statistics, chemistry or chemical engineering. Applications should consist of a let-ter of interest, a resume, a letter of recommendation from a major pro-fessor, research supervisor, or col-league familiar with the applicant's research record, and an abstract of research to be presented in the con-ference poster session.

For consideration at the first decision point, applications must be received by the conference Vice Chair by May 15, 1992. Please send application (4 copies) for these grants to Roger W. Hoerl, Conference Vice Chair, Scott Paper Company, Scott Plaza II, Phila., PA 19113, with a copy to the conference chair, Richard D. De Veaux, Dept. of Civil Eng'g and OR, Princeton University, Princeton, NJ 08544. Applicants must also apply directly to the conference itself by contact-ing Dr. A.M. Cruickshank, Gordon Research Center, University of Rhode Island, Kingston, RI 02881; (401) 783- 4011 (E-mail:bcp101@ uriaccitnet).

Survey Results Continued from page 6

Since CQE certification may be --important to the members, we

asked several questions on the cer-tification and exam.

SPRING, 1992 STATISTICS DIVISION NEWSLETTER PAGE 7

Page 8: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

BASIC TOOLS

YOU WANT ME TO DO WHAT !!!! (Or Making Sense Out of Chaos Using Affinity Diagrams)

Jim Bossert, Eastman Kodak Company, Rochester, NY

The affinity diagram is one of the most widely utilized of the Seven New Management Tools. It can be used to effectively display ideas generated from a brainstorming session. Many of you may be famil-iar with the technique but did not know what it was called, how it was developed, or the approach to be used in its generation.

The affinity diagram process was developed in Japan by a Japanese anthropologist, Jiro Kawakita. Looking for understanding of the societies that he was studying, he took the ideas he had written on pieces of paper, spread them on the table and looked for natural group-ings. He discovered this process allowed him to sort through all kinds of seemingly unrelated data and identify fundamental themes for each grouping. These themes did not always follow traditional lines of thought and so, he concluded that they enhanced breakthrough think-ing. Kawakita refined the method, named it KJ, and registered the KJ Method as a trademark.

How does one conduct an affinity diagram session ? First, a team has to be formed. There are different points of view on the size of a team. Some authors state that the "ulti-mate" team size is between 4 and 8 people. Others like to use a larger representation of an organization, say 20 to 30 people. Choose the size of the team you are comfortable managing. The key is to select as diverse a group as possible; a multi-disciplined and multifunctional team. This approach insures that you can challenge traditional view-points and yet still be open-minded to new perspectives. A good way to assemble a team is to list all of the job functions in the organization and choose those that are critical for them to be part of the team. This usually identifies 6 to 8 people. The remainder of the list can generally be split between those individuals you would really like to have but are not essential and those that you want to be familiar with what you

are doing. This latter group can usually be put on an "information-only" list.

After your team has been orga-nized, determine the issue, topic, product, or product line to be addressed by the group. This part of the process is critical_ Everyone needs to understand the topic and why it was chosen. This will allow the group to focus on brainstorm-ing. One school of thought suggests that the issue statement should be as vague as possible. The reasoning here is to minimize the chance of "funnelling" ideas into the way they have always been done. This is not without some risk. Too vague a statement can dilute the process by having people address too much; never answering the question. Another school of thought suggests being as specific as possible so that the critical issue is addressed. The danger is that no new approaches will come out of this process. This can be offset by having as large a team as possible.

Now that the topic has been iden-tified, the brainstorming can begin. There are as many ways to brain-storm. Make sure that all ideas are recorded on a flip chart. It's a good idea to have a second recorder pre-sent to record the same ideas on 3x5 cards or a sticky pad. Use of cards or pads depends on where the ses-sion is held and how the team will organize the data. If the room has large table(s), cards work well. If there is limited room available, use sticky pads so that the walls can be used for sorting. A team member should not record because it is diffi-cult to record and contribute at the same time. When brainstorming is completed, the cards/pads should be spread out by the team members on the table or walls in a random order.

The team is now ready to sort. Sorting should be done in silence. By imposing a rule of silence there is less chance of influencing an indi-viduals' perspective. Silence also forces the creative side of the brain

to take over and gets team members to look at data from a different viewpoint. When sorting cards/pads, team members should be looking for instinctive groupings/categories. Sorting tends to be a relatively quick process.

Once sorting is completed, the challenge begins; defining the cate-gories on header cards. The team needs to decide the most concise descriptor for each category. Sometimes, this can be one of the cards already generated, although this rarely happens. The team will have to come to a consensus to the category name. Usually, half of the category names are self-evident. Naming the other categories will take time because there may be dis-agreement over what to call the cat-egory. However, if the process is not rushed, excellent category names will be generated.

The final step is laying out the affinity diagram. There are twc methods. The first way is to lay the cards/pads out and place the catego-ry descriptor/header on the top of the cards. Aptly named, parent and children. The other method is to form a tree diagram with the head-er/descriptor on the left side of all the sub-categories. This is similar to taking an organization chart and turning it sideways. This method is used most frequently with QFD (Quality Function Deployment) matrices.

Affinity diagrams can be a pow-erful tool in the organization of data. It allows grouping of ideas into categories, thereby making sense out of chaos. Once you have experienced the process, you will discover its merits and benefits.

References Brassard, Michael, The Memory

Jogger Plus+rm, Goal/QPC, Methuen, MA, 1989.

Mizuno, Shigeru. Management for Quality Improvement, Producti-vity Press, Campridge, MA, 1988.

PAGE 8

STATISTICS DIVISION NEWSLETTER SPRING, 1992

Page 9: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

GORDON CONFERENCE

1992 GORDON RESEARCH CONFERENCE ON STATISTICS IN CHEMISTRY AND CHEMICAL ENGINEERING

July 27-July 31, 1992• New Hampton School, New Hampton, NH Chair: Richard D. De Veaux, Princeton University Vice Chair: Roger W. Hoerl, Scott Paper Company

1. Speaker: John F. MacGregor, McMaster University, Ontario Canada Topic: Multivariate Process Control Discussant: Svante Weld, Umea University, Sweden Moderator: Karen Kafadar, National Cancer Institute

2. Speaker: Adrian F.M. Smith, Imperial College, London, UK Topic: Bayesian Methods for Reliability Discussant: Bruce Hoadley, Bellcore Moderator: Edward I. George, University of Chicago

3. Speaker: Ingram Olkin, Stanford University Topic: Meta Analysis in Science Discussant: Donald B. Rubin, Harvard University Moderator: Donna F. Stroup, Centers for Disease Control

4. Speaker: Lyle H. Ungar, University of Pennsylvania Topic: Neural Networks for Chemical Process Control Discussant: James M. Minor, Du Pont Moderator: Roger W. Hoerl, Scott Paper Company

5. Speaker: Trevor J. Hastie, AT&T Bell Laboratories Topic: Generalized-Generalized Additive Modules Discussant: Anthony C. Atkinson, London School of Economics, UK Moderator: Diane E. Duffy, Bellcore

6. Speaker: Philip K. Hopke, Clarkson University Topic: Multivariate Methods for Environmental Chemical Analysis Discussant: Nouna Kettaneh-Wold, MDS Inc. Moderator: Age Smilde, University Center for Pharmacy, Groningen, Netherlands

7. Speaker: V. V. Fedorov, Academy of Sciences, Moscow, Russia Topic: Moving Regression: Analysis and Design of Experiments Discussant: Richard D. De Veaux, Princeton University Moderator: James M. Lucas, Consultant

K. Speaker: John C. Bailar, McGill University, Quebec, Canada Topic: New Directions in Risk Assessment Discussant: J. Michael Steele, University of Pennsylvania Moderator: Karen J. Hyver, Baxter Health Care

9. Speaker: Andreas Buja and Deborah F. Swayne, Bellcore Topic: Interactive Statistical Graphics using the X windows System Discussant: Randall D. Tobias, SAS Institute Moderator: Karen J. Bandeen-Roche, Johns-Hopkins University For applications to attend the conference contact: Dr.

A. Cruickshank, Gordon Research Conferences, University of Rhode Island, Kingston, RI 02881; (401) 783-4011. ([email protected]). To submit abstracts for poster session papers, contact: Roger W. Hoerl, Scott Paper Company, Scott Plaza II, Philadelphia, PA 19113. For further information on other aspects of the conference, contact: Richard D. De Veaux, Department of Civil Engineering and Operations Research, ACE-43 Engineering Quadrangle, Princeton Univeristy, Princeton, NJ 08544; (609) 258-6494 (deveaux@ frosty.princeton,.edu).

STATISTICS DIVISION MEMBERS OPPORTUNITY TO UPGRADE STATUS TO SENIOR MEMBER

If you are currently an ASQC Member who meets the following eligibility requirements you should apply for Senior Member status. Eligible Members must:

1. Be at least 30 years of age, and actively involved in the quality profession for at least 10 years. Graduation in an approved engineering science, mathematics, or statistics curriculum is considered the equivalent of 4 years of professional experience. and,

2. Qualify under one or more of the following: a. be responsible for important engineering or inspec-tion work involving quality control for at least 2 years. b. be an instructor of quality control, engineering or statistical methodologies as applied to quality control for at least 2 years; should be capable of teaching a variety of courses in the quality field. c. be a professional engineer or member of a technical society of national status in any country for which the qualifications require a standing equivalent to that required for a senior member of ASQC.

All eligible Members interested in upgrading their membership status should contact the Statistics Division Examining Chair, Bob Perry, at Grand Metropolitan Technology Center, 330 University Avenue S.E., Minneapolis, MN 55414, or at (612) 330-8916.

;PRING, 1992 STATISTICS DIVISION NEWSLETTER PAGE 9

Page 10: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

THIS QUARTER IN TECHNOMETRICS

THIS QUARTER IN JOT TABLE OF CONTENTS • Vol. 24, No. 2, April 1992

ARTICLES Statistical Control Charts Based on a Geometric Distribution Frank Kaminsky, James C. Benneyan, Robert D. Davis,

and Richard J. Burke

An Explanation and Criticism of Minute Accumulating Analysis Michael Hamada

Double Sampling X Charts J.J. Daudin

Multivariate Control Charts for Individual Observations Nola D. Tracy, John C. Young, and Robert L. Mason

DEPARTMENTS Computer Programs

A Goodness-of-Fit Test Based on P-P Probability Plots F.F. Gan and KJ. Koehler

Technical Aids Sample Sizes for Two-Sample Tests on Means Lloyd S. Nelson

Book Reviews Wetherill, G.B. and Brown, D.W., Statistical Process Control: Theory and Practice John S. Dowd

Lochner, R.H. and Matar, J.E., Designing for Quality Lynne B. Hare

Shina, G., Concurrent Engineering and Design for Manufacture of Electronic Products. Michael Pecht

Tierney, L., LISP-STAT: An Object-Oriented Environment for Statistical Computing and Dynamic Graphics Rick L. Edgeman

Fellers, G., SPC for Practitioners: Special Cases and Continuous Processes Richard W. Kulp

Wilson, E.B., An Introduction to Scientific Research Robert L. Perry

STATISTICS DIVISION ANNUAL BUSINESS MEETING

Statistics Division's Annual Business Meeting will be held at 5:30-7:00 pm on Monday, May 18,1992 at the Opryland Hotel in Nashville, in conjunction with the 1992 Annual Quality Congress. The exact room will be posted at the hotel. The program for the meeting will be: • Meet the officers • State of the Division: membership and financial status • Vision, Mission, Strategies, and Tactical Plans • Annual Quality Congress and Fall Technical Conference • Statistics Division Operating Manual / Infrastructure

Membership survey • Plans for Education • Open forum for Stat Division members: formal and informal

feedback

All Statistics Division members are invited. Hors d'oeuvres and refreshments will be provided at the meeting.

TAGUCHI'S PARAMETER DESIGN: A PANEL

DISCUSSION It is more than a decade since

Genichi Taguchi's ideas on quality improVement were introduced in the U.S. His parameter design approach (also known as robust design) for reducing variation in products and processes has generat-ed a great deal of interest among both quality practitioners and statisticians. The statistical tech-niques used by Taguchi to imple-ment parameter design have been the subject of much debate, howev-er, and there has been considerable research aimed at integrating the parameter design principles with well-established statistical tech-niques. On the other hand, Taguchi and his colleagues feel that these research efforts by statisticians are misguided and reflect a lack of understanding of the engineering principles underlying Taguchi's methodology.

While there have been several -papers and books explaining, reviewing, or criticizing Taguchi's ideas, most of these have not ade-quately captured the diverse views on the topic. In particular, the views of the so-called Taguchi school have not been well represent-ed in statistical journals. Therefore, a panel discussion by a group of leading researchers and practition-ers appeared to be the best way to provide readers with a balanced and up-to-date overview of: a) the importance and usefulness of the principles underlying parameter design, b) Taguchi's methodology for implementing them, and c) the various research efforts aimed at developing alternative methods.

The panel discussion is the lead article in the May 1992 issue of Technometrics. The participants are: Bovas Abraham and Jock MacKay (Univ. of Waterloo), George Box (Univ. of Wisconsin, Madison), Raghu N. Kacker (Nat. Inst. of Stds. and Tech.), Thomas J. Lorenzen (General Motors Res. Labs), James M. Lucas (Du Pont), Raymond H.

Continued on page 11

PAGE 10

STATISTICS DIVISION NEWSLEilER

SPRING, 1992

Page 11: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

Technometrics Continued from page 10

Myers and G. Geoffrey Vining (VII and Univ. of Florida), John A. Nelder (Imperial College, London), Madhav S. Phadke (Phadke Associates), Jerome Sacks and William J. Welch (Nat. Inst. of Stat. Sciences and Univ. of Waterloo), Anne C. Shoemaker and Kwok L. Tsui (AT&T Bell Labs and Georgia Inst. of Tech.), Shin Taguchi (American Supplier Inst.), and C.F. Jeff Wu (Univ. of Waterloo). The discussion focuses on both the role of parameter design and of Taguchi's methodology for imple-menting it. The topics discussed include the importance of variation reduction, the use of noise factors, the role of interactions, selection of quality characteristics, signal-to-noise (SN) ratios, experimental strategy, dynamic parameter design problems, and applications. The dis-cussion also provides an up-to-date overview of recent research on alternative methods of design and analysis and includes an extensive bibliography.

This is different from the usual sort of panel discussion. First, it vas not feasible to assemble all par-dcipants in a common location. Second, the proceedings of a "free-for-all" oral discussion may have added to existing confusion rather than shed new light on the issues. For these reasons, comments were solicited from participants on a number of topics, and a panel dis-cussion was created from their com-ments. Panelist provided comments on topics in which they have worked or had practical experience. Their comments were organized into sec-tions to give readers a balanced pic-ture of the different views on each topic. Panelists had some opportu-nity to read and respond to the com-ments of others. All comments were edited extensively to make the over-all discussion and the individual sections flow smoothly and to remove tangential material and excessive overlap. The discussion has been deliberately kept at a con-ceptual level and, for the most part, readers have been referred to other sources for relevant technical

It is not the goal of this panel dis-

cussion to try to resolve any of the differences that currently exist on the issues. It is merely intended as a forum for a technical discussion of the diverse views so that readers have a better basis for reaching their own conclusions. Readers should also find the up-to-date overview of recent research efforts and the extensive bibliography use-ful in obtaining a further under-standing of the issues.

Vijayan N. Nair, Editor

REGION 12 ANNOUNCEMENT QA DAY AT ARGONNE

NATIONAL LABORATORY STATISTICS DIVISION

PAPER PRESENTATIONS The Statistics Division sponsored

three talks at QA Day at Argonne Laboratory.

"Six Sigma Statistics" presented by Bill Smith, Vice President and Senior QA Manager, Motorola, Inc.

"Developing Exponential Sampling Plans" presented by

Jim Baker, Quality Specialist, 3M Company and Bob Karm, Supplier QA

Engineer, Rank Video Services America.

"Optimizing Torque Output Levels Using Statistics" presented by

Bill Gartland, Senior Quality Engineer, Furnas Electric.

Jed Heyes, Region 12 Councilor, moderatored these presentations. The conference was held April 2, 1992 in Argonne, Illinois.

17th Annual Summer -Institute of

Applied Statistics The Role of Statistics

and Statisticians in TQM June 17-19, 1992 at Brigham Young

University, Provo, Utah

The conference will center on some fundamental concepts of TQM and the critical role of statistical science. It will focus on many of the tools and techniques that statisti-cians and other quality practition-ers must learn to become effective catalysts in the improvement of quality in all sectors of society.

Dr. Thomas J. Boardman will be the principal discussion leader. Dr. Boardman is a Professor of Statistics at Colorado State University (on leave), a Fellow of both the ASA and ASQC, and a con-sultant with Boardman Associates. He has enjoyed a close association with W. Edwards Deming and has mad extensive application of Deming's philosophies in industry, academia, and government.

Dr. G. Rex Bryce, Professor of Statistics at BYU and Fellow of the ASA, along with several practicing statisticians who are currently con-sultants in the various aspects of TQM, will also lead discussions.

Who Should Attend? Indus-trial and academic statisticians, quality professionals, and others who are involved in quality related aspects of business, industry, or education.

The fee is $350 per person. For a complete brochure and registration, contact Shauna Kuykendall at the Department of Statistics, Brigham Young University, 230 TMCB, Provo, UT 84602, (801)378-4527, FAX:(801)378- 5722,[email protected].

SPRING, 1992 STATISTICS DIVISION NEWSLETTER PAGE 11

Page 12: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

DIVISION PUBLICATIONS Basic References in Quality Control: Statistical Techniques

The Statistics Division has prepared a series of authoritative booklets on important statistical tech-niques for quality control. The popular name of this series - the "How To" series - reflects the practical, applications- oriented approach taken throughout.

This ongoing series offers a valuable way of keeping up to date with developments in the field. These books are all of the intermediate level - recommended for those with some knowledge of the field, usually requir-ing advanced math. Volume 1: How to Analyze Data with Simple Plots

Wayne Nelson # T3501 Member Price $16.95 List Price $18.95

Volume 2: How to Perform Continuous Sampling (CSP) Kenneth S. Stephens # T3502 Member Price $18.95 List Price $20.95

Volume 3: How to Test Normality and other Distributional Assumptions, Revised Edition

Samuel S. Shapiro # T3505 Member Price $21.95 List Price $23.95

Volume 4: How to Perform Skip-Lot and Chain Sampling

Kenneth S. Stephens # T3504 Member Price $18.50 List Price $20.50

Volume 5: How to Run Mixture Experiments for Product Quality, Revised Edition

John A. Cornell # T3505 Member Price $18.95 List Price $20.95

Volume 6: How to Analyze Reliability Data Wayne Nelson # T3506 Member Price $17.50 List Price $19.50

Volume 7 : How and When to Perform Bayesian Acceptance Sampling, Revised Edition

Thomas W. Calvin # T3507 Member Price $17.95 List Price $19.95

Volume 8 : How to Apply Response Surfac Methodology, Revised Edition

John A Cornell # T3508 Member Price $21.95 List Price $23.95

Volume 9 : How to Use Regression Analysis in Quality Control, Revised Edition

Douglas C. Crocker • # T3509 Member Price $16.95 List Price $18.95

Volume 10: How to Plan an Accelerated Life Test -Some Practical Guidelines

William Q. Meeker and Gerald J. Hahn # T3510 Member Price $18.50 List Price $19.95

Volume 11: How to Perform Statistical Tolerance Analysis

Neil D. Cox # T3511 Member Price $24.50 List Price $26.95

Volume 12: How to Choose the Proper Sample Size Gary G. Brush # T3512 Member Price $28.00 List Price $30.50

Volume 13: How to Use Sequential Statistical Methods Thomas P. McWilliams # T3513 Member Price $19.95 List Price $21.95

Volume 14: How to Construct Fractional Factorial Experiments

Richard F. Guest and Robert L. Mason # T3514 Member Price $23.95 List Price $25.95

Volume 15: How to Determine Sample Size and Estimate Failure Rate in Life Testing

Eduardo C. Mourn # T3515 Member Price $21.95 List Price $23.95

COMPLETE SERIES - # T3500 Member Price $260.00 List Price $290.00

To order call ASQC Quality Press (800) 952-6587.

ASQC Statistics Division 1991-1992 Regional Councilors Regional Councilor Coordinator - Edward R. Hansen

Gillette Company, Prudential Tower Building, Boston, MA 02199 • (617) 421-7370

Region 1 Edward R Hansen Gillette Company

Prudential Tower Building Boston, MA 02199

(617) 421-7370 Region 2

Donald J. Baird Xerox Corporation

800 Phillips Road, 0200-18C Webster, NY 14580

(716) 422-8085 Region 3

Thomas Calvin (IBM Corporation) 13 Malstorme Road

Wappingers Falls, NY 12590 (914) 433-9240

Region 4 Ian B. MacNeill

Director, Statistics Laboratory Engrg. and Math. Sciences Bldg. University of Western Ontario

London, Ontario N6A 5B9 (519) 361-6618

Region 5 Joseph R. Troxell Lasalle University

Box 374 20th Street and Olney Avenue

Philadelphia, PA 19141 (215) 951-1031

Region 6 Marilyn Hwan

LSI Logic 3115 Alfred St.

Santa Clara, CA 95050 (408) 433-6362

Region 7 Thomas J. Vaden

Consultant 5765 Grand Avenue Riverside, CA 92504

(714) 382-6525 Region 8

William M. Bleau (Picker International Inc.)

1130 Stonecrest Dr. Tailmadge, OH 44278

(216) 473-2385 Region 9

Carlos W. Moreno Ultramax Corporation 650 Northland Blvd.

Cincinnati, OH 45240 (513) 825-7794

Region 10 Gregory P. Gruska

(The Third Generation, Inc.) 4439 Rolling Pine Drive

West Bloomfield, MI 48033 (313) 363-1654

Region 11 George W. Marrah

Department of Mathematics and Computer Science

James Madison University Harrisburg, VA 22807

(703) 568-6534 Region 12

Jed Heyes Rank Video Services of America

555 Huehl Road Northbrook, IL 60062

(708) 291-2227 Region 13

Richard D. Schleusener Kodak Colorado Division Melt/Coat Department

C-42, 3rd Floor Windsor, CO 80551

(303) 686-0530 Region 14

E.B. Godsey 303 Ridgebriar Drive Richardson, TX 75080

(214) 690-1744 Region 15

William H. Woodall Department of Management

Science and Statistics University of Alabama

P.O. Box 870226 Tuscaloosa, AL 35487-0226

(205) 348-6085

PAGE 12

STATISTICS DIVISION NEWSLETTER

SPRING, 1992

Page 13: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

Malcolm Baldrige National Quality Award

Maureen S. Heaphy The Transformation Network, Inc. 4439 Rolling Plne Dr. West Bloomfield, MI. 48323 313/363-1654

The Malcolm Baldrige National Quality Award (MBNQA) was signed into law

on August 20, 1987. The Award is named for Malcolm Baldrige who was Secretary of Com-merce from 1981 until his accidental death in 1987. There are currently three categories of eligibility: Manufacturing. Service and Small Business. No more than two awards will be given in each category and no award need be given if the standards of the Award are not met. Any U.S.A. for-profit business or (with some restrictions) subsidiaries may apply. The intent of the Award is to promote:

• awareness of quality and its impact on competitiveness

• understanding of the requirements for excellence in quality

• sharing of information on successful strategies and on benefits derived

The Award overseer is the National insti-tute of Standards and Technology (NISI. for-merly known as National Bureau of Standards) under the direction of the Department of Com-merce. The Administrator for the Award is the American Society for Quality Control (ASQC). The 1991 board of examiners consists of 9 judges and approximately 220 examiners (some are designated as senior examiners). Funding for the Award is through private con-tributions to the Foundation for the MBNQA.

The Award process is initiated when an applicant submits the Eligibility Determination Form. If the applicant is qualified then a written Application Package containing a two page overview of the applicant's business and re-sponses to 28 examination items is required. For the Manufacturing and Service category, the application report is restricted to 75 pages. For small businesses, the limit is 50 pages.

The evaluation process, Figure 1, always results in a written feedback report to the ap-plicant. Once the submission deadline date is reached, applications are assigned to examin-ers. Each of the 5 to 6 examiners reviewing an application complete an independent assess-ment. Based on the ratings at this first stage, the judges decide if the application should continue in the review process.

If the application is not selected for the second stage. a feedback report is written based on written comments from the indepen-dent assessments.

Figure 1: Evaluation Process

The second stage review consists of a senior examiner and 3 to 5 examiners discuss-ing their ratings from the first stage and reach-ing a consensus for each examination item. The judges use the consensus ratings to select which applicants will receive a site visit. If an applicant is not selected for a site visit they will receive a feedback report based on the con-sensus results.

A site visit involves 4 to 6 examiners includ-ing a senior examiner. After a site visit, the applicant receives a feedback report based on all of the information that has been reviewed. For the final contenders this adds up to about 400 hours of evaluation each.

A panel of judges reviews all of the evalu-ation reports and makes a recommendation to the overseers of the Award, the National Insti-tute of Standards and Technology. The Insti-tute presents the Judges' recommendation to the Secretary of Commerce for the final Award decision. The Awards are then announced and

*PRING, 1992

STATISTICS DIVISION NEWSLETTER PAGE 13

Page 14: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

presented in the last quarter of the year. In the past the Awards were presented by the Presi-dent or Vice-President of the Unites States in special ceremonies in Washington, D. C.

The feedback reports for all of the appli-cants are sent after the Award winners have been announced.

The Board of Examiners are selected ac-cording to their expertise, experience and peer recognition. They do not represent their com-panies or organizations. The training for the examiners is a 3 day class based on case studies. The focus is on reducing variability among the examiners and building a cooper-ative team spirit.

Many companies and organizations are viewing the criteria as the national definition of Total Quality Management (TOM). The criteria is being used for more than just applying for the Award. Internally companies are using it to review and guide their quality initiatives. •

In 1991, there were over 250.000 copies of the guidelines distributed. The number of ap-plicants and sites visits are summarized in Table 1.

Category 1988 1989 1990 1991 Mfg 45 S 10) 23 (8) 45 (6) 38(9) Service 9 ( 2) 6 (2) 18 (3) 21(5) Sm Bus 12 ( 1) 11 (0) 34 (3) 47(5) TOTAL 66 (13) 40 (10) 97 (12) 106(19)

Table 1: Number of Applicant (and site visits)

The previous Award winners are as follows:

1991 • Solectron Corporation • Zytec Corporation • Marlow Industries

1990 •Rochester Division (IBM) • Cadillac Motor Car Company (GM) • Federal Express • Wallace Company, Inc.

1989 •Milliken & Company • Business Products and Systems

(Xerox)

1988 •Motorola • Commercial Nuclear Fuel Division

(Westinghouse) • Globe Metallurgical

The application contains seven examina-tion categories which are summarized here in an excerpt from NISI material.

1.0 Leadership The senior executives' success in creating and sustaining a quality culture.

2.0 Information and Analysis The effectiveness of information collection and analysis for quality improvement and planning.

3.0 Strategic Quality Planning The effectiveness of integrating quality re-quirements into business plans.

4.0 Human Resource Development and Management

The success of efforts to develop and realize the full potential of the work force for quality.

5.0 Management of Process Quality The effectiveness of systems and processes for assuring the quality of all operations.

6.0 Quality and Operational Results The results in quality achievement and quality improvement, demonstrated through quantita-tive measures.

7.0 Customer Focus and Satisfaction The effectiveness of systems to determine customer requirements and demonstrated success in meeting them.

These seven categories are then further defined with 28 examination items. Under each item there are areas to address, 89 in total. A sample, shown in Figure 2, is from the 1992 application guidelines. Only 1 of the 3 items for Leadership is shown.

A numeric rating is done for each of the 28 items and used to calculate the score for each of the seven categories. The relative impact of each category is shown below in Table 2.

Category. . Percent

Leadership 9

Information and Analysis 8

Strategic Quality Planning 6

Human Resource Development 15

Management of Process Quality 14

Quality and Operational Results 18. Customer Focus and Satisfaction 30

TOTAL _ 100

Table 2: Relative Impact of each Category

The key concepts in the criteria include

• customer driven quality • leadership creates environment for

quality

• continual improvement • full participation • fast responses

• design quality and prevention

• long range outlook

• management by fact

PAGE 14 STATISTICS DIVISION NEWSLE1 h,R. SPRING, 1992

Page 15: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

• partnership development • public responsibility

CATEGORY -- > 1.0 Leadership (90 pts.) The Leadership category examines senior

executives' personal leadership and involve-ment in creating and sustaining a customer focus and clear and visible quality values. Also examined is how the quality values are inte-grated into the company's management sys-tem and reflected in the manner in which the company addresses its public responsibility.

ITEM --> 1.1 Senior Executive Leadership (45 pts.)

Describe the senior executives' leader-ship, personal involvement, and visibility in de-veloping and maintaining a customer focus and an environment for quality excellence.

AREAS TO ADDRESS a. senior executives' leadership, personal

involvement, and visibility in quality related ac-tivities of the company. Include: (1) reinforcing a customer focus; (2) creating quality values and setting expectations; (3) planning and re-viewing progress toward quality and perfor-mance objectives; (4) recognizing employee contributions; and (5) communicating quality values outside the company.

b. brief summary of the company's quality values and how the values serve as a basis for consistent communication within and outside the company.

c. personal actions of senior executives to regularly demonstrate, communicate, and re-inforce the company's customer orientation and quality values through all levels of man-agement and supervision.

d. how senior executives evaluate and im-prove the effectiveness of their personal lead-ership and involvement.

Figure 2: Example of Category, Item and Area

A 3 dimensional approach is used in scor-ing the items. As applicable, each item is re-viewed for approach, deployment and results. The approach is to be a prevention based system, showing continual improvement and excellent integration. Deployment refers to the extent that the approaches are applied to all products and services. The results derived from the approach need to be sustained and show excellent levels.

The distribution of written scores, prior to site visits, is shown in Table 3 .

Scoring Range

Number of Companies

1988 1989 1990 1991

0 - 125 0 0 0 3

126 - 250 0 1 7 14

251-400 1 8 18 38

401 - 600 31 15 51 36

601 - 750 23 12 19 15

751 - 875 11 4 2 0

876-1000 0 0 0 0

Number of applications

66 40 97 106

Table 3: Distribution of Scores

Some of the characteristics of higher scoring companies are:

• Aggressive quality goals and plans

• Benchmark/response time drivers

* P ro active customer systems

• Quantitative orientation

• Major human resource investments

This is contrasted with some of the character-istics found in the lower scoring companies:

• Passive leadership • Reactive customer systems • Limited measures/benchmarks

• Quality in key product/service only

•Limited closed loop

Advice to candidates is to be concise, fac-tual and quantitative. In completing the appli-cation, pay particular attention to approach, deployment and results. Demonstrate a de-fined strategy, prevention based system and continuity. Avoid empty narrative in your an-swers.

For more information on the Award, contact:

Malcolm Baldrige National Quality Award National Institute of Stds and Technology Administrative Building Room A-537 Gaithersburg, MD 20899 301 / 975 - 2036

SPRING, 1992 STATISTICS DIVISION NEWSLETTER PAGE 15

Page 16: •1.11.• STATISTICS DIVISION NEWSLETTERasq.org/statistics/1992/03/asq-statistics-division... · 2013. 12. 16. · STATISTICS DIVISION NEWSLETTER AMERICAN SOCIETY FOR QUALITY CONTROL

STATISTICS DIVISION ANNUAL BUSINESS MEETING MONDAY, MAY 18 1992, 4■:-5...31 . 7:00 PM

ANNUAL QUALITY NASHVILLE, TENNESSEE

All members are invited to attend. This is an opportunity for our membership to give suggestions, comments, and ask questions.

Non-Profit Org. U.S. Postage

PAID Cedarburg, WI Permit No. 199

INSIDE THIS ISSUE: 46th Annual Quality Congress p. 3-6 AQC Pre-Conference Tutorial p. 4 Nominations for Hunter Award p. 6 Membership Survey p. 6-7 Basic Tools: Affinity Diagrams p. 8 "How To" Series p.12 Mini-Paper: Malcolm Baldrige p. 13-15

Quality Award

STATISTICS DIVISION AMERICAN SOCIETY FOR QUALITY CONTROL c/o Nancy Baxter T.J. Lipton 800 Sylvan Avenue C-210 Englewood Cliffs, NJ 07632

All communications regarding membership including change of address, should be sent to: American Society for Quality Control P.O. Box 3005 Milwaukee, WI 53201-3005

(414) 272-8575

The ASQC Statistics Division Newsletter is a publication of the statistics Division of the American Society for Quality Control.

All communications regarding this publi-cation, EXCLUDING CHANGE OF ADDRESS, should be addressed to:

Nancy Baxter, Editor ASQC Statistics Division Newsletter Thomas J. Lipton Company 800 Sylvan Avenue C-210 Englewood Cliffs, NJ 07632

Other communications relating to the Statistics Division of ASQC should be ad-dressed to:

Conrad A. Pung Center for Quality and Productivity Improvement

University of Wisconsin WARY/1109B 610 Walnut Street Madison, WI 53705

(608) 263-2652

ASQC STATISTICS DIVISION NEWSLETTER