Statistics for Quality (Short Course)
Statistics for Quality
A Comprehensive Six-Part Series
Anand Joglekar, Ph.D.
November 3, 4 and 5, 2008
Michael's at Shoreline
2960 N Shoreline Blvd
Mountain View, CA 94043
The Golden Gate Polymer Forum announces a three-day short course covering six major statistical methods. Presented in a practical manner with examples from polymeric materials, medical device and drug delivery applications, the program is customizable for registrants who wish to attend one, two or all three days.
UPDATE: Registration is closed.
November 3, 2008 (Day 1)
8:00-11:30 AM - Program 1. "Comparative Experiments - Demonstrating Change or Equivalence"
1:00- 4:30 PM - Program 2. "Screening Experiments - The Heart of DOE"
November 4, 2008 (Day 2)
8:00-11:30 AM - Program 3. "Setting Specifications"
1:00- 4:30 PM - Program 4. "Process Validation and Acceptance Sampling"
November 5, 2008 (Day 3)
8:00-11:30 AM - Program 5. "Managing and Improving Processes"
1:00- 4:30 PM - Program 6. "Measurement Systems Analysis"
The following criteria were used to select the six statistical methods included in this seminar:
- The method is widely applicable in R&D and manufacturing.
- The method is underutilized in industry and wider use will lead to beneficial results.
- The method is being wrongly used, or wrong methods are being used, to solve practical problems.
- There are misconceptions regarding the method.
The combination of an excellent teacher, a convenient Bay Area location, a low price, and the focus on practical applications including the use of software makes this an excellent opportunity.
The series of tutorials given over 3 days are geared toward technical professionals who deal with data and make decisions based on data. This includes scientists, engineers, technicians, personnel from research and development, manufacturing, quality, analytical, engineering, supply management and regulatory affairs departments. Technical managers, "green belts" and "black belts" will also benefit.
Day 1: Monday, November 3, 2008
Program 1: Comparative Experiments – Demonstrating Change or Equivalence (8 to 11:30 AM)
A common situation in industry is to conduct experiments to demonstrate that a change, perhaps an improvement, has occurred, or that no change has occurred, thereby establishing equivalence. Thus, comparative experiments are conducted to show that the new process has less variability than the old, that the new formulation is better than the old, or to show that two product designs or two measurement methods are equivalent. Such problems are often analyzed using a t-test, an F-test or an ANOVA. Such hypothesis tests, when blindly applied, can lead to disastrous results. For one company, this SOP caused an unnecessary delay of three months and an unnecessary expenditure of over $250,000!
The purpose of this seminar is to provide practical guidance to conduct comparative experiments, and answer questions such as:
- What is wrong with the t-test and the F-test as usually practiced?
- What to replace these tests with?
- How to make multiple comparisons?
- How much data to collect?
- How to reduce sample sizes?
- How to use software to design and analyze comparative experiments?
Program 2: Screening Experiments – The Heart of DOE (1 to 4:30 PM)
Product and process development requires identification of key factors that control product and process performance. Screening experiments efficiently identify key factors and interactions. They provide a way to optimize product and process designs, and to troubleshoot manufacturing. Screening experiments are the most important class of designed experiments (DOE), and help shorten development cycle time and costs. Unfortunately, much product and process development proceeds using one-factor-at-a-time or other ad-hoc experimentation strategies – approaches that provide less information at higher cost. On the other hand, there are any number of examples where screening experiments, properly applied, have resulted in – making a better cake, brewing a better beer, improving drug formulations, reducing development cycle times, designing robust measurement systems, and improving the accuracy and consistency of ammunition.
The purpose of this seminar is to provide practical guidance on the effective use of screening experiments, by answering questions such as:
- Why design experiments?
- What are factorial designs?
- What to think about to make the screening experiment successful?
- How to evaluate a large number of factors with a fractional factorial design?
- How to design robust products and processes?
- How to easily implement screening experiments using software?
Day 2: Tuesday, November 4, 2008
Program 3: Setting Specifications (8 to 11:30 AM)
One key objective in R&D is to establish specifications for final product, in-process and raw material characteristics. There are two fundamentally different ways to set specifications: based upon functional relationships, and based purely upon data collected on the characteristic for which the specification is desired. Despite the fact that considerable attention is devoted to this subject in industry, specifications are often rather poorly set. This can result in setting arbitrary specifications, for characteristics that may or may not even be important. Once set, the specifications can be difficult to change. Badly defined, sometimes mismatched specifications lead to costly consequences.
The purpose of this seminar is to provide practical guidance regarding effective ways of setting specifications by answering questions such as:
- How to set specifications when the only available data is the characteristic for which the specification is desired?
- How to set specification by obtaining a functional relationship between an input and output characteristic?
- How to set functional specifications when there are multiple input and output characteristics?
- How to use software to implement the necessary statistical tools to develop specifications?
Program 4: Process Validation and Acceptance Sampling (1 to 4:30 PM)
Acceptance sampling plans are routinely used to accept or reject incoming raw material lots, in-process lots and finished product lots. There is often some misunderstanding in industry regarding how best to design such plans and the protection that they provide. One purpose of this seminar is to remove these misconceptions and to present a practical approach to design acceptance sampling plans. Process validation studies are routinely conducted to provide assurance that the process is ready to be used for manufacturing. Many different approaches are used in industry to design and analyze process validation studies.
The purpose of this seminar is to provide practical guidance for the design of process validation studies and acceptance sampling plans, by answering questions such as:
- How are attribute and variable sampling plans usually designed?
- What is the most important information you need to know to design acceptance sampling plans and how to get it?
- What is the practical approach to design acceptance sampling plans to reduce risks and costs?
- What is the connection between acceptance sampling plans and process validation?
- What is the practical approach to design a process validation study to reduce the risk of failing validation while reducing the cost of sampling?
- How to use software to design sampling plans and validation studies?
Day 3: Wednesday, November 5, 2008
Program 5: Managing and Improving Processes (8 to 11:30 AM)
Control charts are implemented in manufacturing with the hope of improving manufacturing processes, but control charts can deliver on this promise only if the manufacturing process meets certain criteria. Capability and performance indices can be used as an at-a-glance-display to rapidly determine whether the process meets these criteria, namely, where control charts may be the effective process management tool and where other statistical tools need to be used. Variance components analysis is one such statistical tool that provides the detailed information necessary to make cost effective process improvements. Using the ideas presented here the manufacturing department of one company was able to quickly develop and implement a focused quality improvement plan. The quality department of another company reduced their voluminous quarterly quality reports while getting greater management attention to the true quality issues.
The purpose of this seminar is to provide practical guidance on the use of statistical tools to cost-effectively manage and improve processes by answering questions such as:
- What is the logic behind control charts and how to select the proper chart?
- What are the four process capability and performance indices?
- How to use these indices as an at-a-glance-display to manage and improve processes?
- How to use variance components analysis to estimate the % of variability caused by each cause?
- How to target variance reduction efforts most cost-effectively?
- How to use software to implement this methodology?
Program 6: Measurement Systems Analysis (1 to 4:30 PM)
Measurement systems serve two decision-making purposes: they are used to make specification related (accept/reject) decisions, and to make product & process improvement decisions. Large measurement variability and inaccuracy can lead to wrong decisions and significantly increase product and process development cycle times and costs. With a single application of measurement systems analysis, one company reduced measurement standard deviation by over 50%, resulting is smaller sample sizes and faster product development. A second company avoided costly wrong decisions and delays by the proper analysis of method transfer data.
The purpose of this seminar is to provide practical guidance to conduct measurement systems analysis and answer questions such as:
- What are the acceptance criteria for measurement system variability and bias?
- How do these criteria differ for destructive and non-destructive measurements?
- How to design measurement system validation studies?
- How to statistically assess if the measurement system is adequate for the task?
- How to improve the measurement system?
- How to use software to conduct measurement systems analysis?
Dr. Anand Joglekar is a leading statistics educator and consultant to industry. Dr. Joglekar received an M.S. in engineering from IIT, Bombay, India and a Ph.D. in engineering with extensive statistics education from University of Wisconsin, Madison. He has held positions as manager of statistics department and manager of new product development at Pillsbury, and manager of strategic planning and director of quality improvement at Honeywell. In 1990, Dr. Joglekar founded Joglekar Associates which is focused on helping industry achieve business successes through statistical methods. He has implemented statistical methods in areas such as research, product design, process design, manufacturing, analytical science, and management. He has taught statistical methods to thousands of industry participants. His customers include major companies in automotive, biotechnology, computer, chemical, defense, drug delivery, food, medical device, mining, packaging, pharmaceutical, semiconductor and many other industries. He has many publications and conference presentations and is a sought after presenter in industry. He is the author of Statistical Methods for Six Sigma in R&D and Manufacturing, published by John Wiley & Sons in 2003.
- The location of the course will be Michaels at Shoreline, 2960 N Shoreline Blvd, Mountain View, CA 94043
- Additional instructions will be provided to those who register for the event by Oct 24, 2008.
- A continental breakfast and lunch will be provided on site each day at no additional charge.
- A Wine & Cheese reception will be provided at the end of lectures on Day 1 .
- Hardcopy lecture notes from each day will be provided to all attendees.
- For attendees who require hotel reservations, numerous hotels are nearby in Mountain View and Palo Alto
List of Nearby Hotels
- The closest airport is San Jose, followed by San Francisco, and then Oakland.
- Early Registration (on or before October 6, 2008) : $245 for each day; $595 for the 3-day program
- Late Registration (October 7-24, 2008) : $295 for each day; $695 for the 3-day program
- Advance Registration and payment required. No drop-ins will be allowed.
- Payment must be made with a Credit Card via PayPal. If this is against your company policy, contact Lothar Kleiner at the email address in the "Contact Information" section.
- No reservations will be accepted after October 24 or after the class size limit is reached (90 attendees for Nov 3 and 4; 72 attendees for Nov 5.)
- No reservation will be considered complete until payment is received.
(1) Begin the registration process on at the bottom of this web page and follow the instructions there.
Remember, you can register for any day, any two days, or all three days. If you register for anything other than the complete three-day course, please select the menu option in the pull-down menu that corresponds to the day or days that you will be attending. Don't forget to press "Add". We will need your name, affiliation, and contact information as requested.
After selecting "Verify", you will be prompted to “Register by Credit Card”. This will enable you to pay the registration fee via PayPal. You will receive an email notice when payment has been confirmed. We are not accepting payment by check at this time. If this is against your company's policy, please contact Lothar Kleiner at the email address in the "Contact Information" section.
If your company requires it, the GGPF Tax ID# will be provided to you by e-mailing a request to Nayan Ashar at email@example.com . Please note that the conference registration fee covers the short course expenses and is not a tax deductible donation.
(2) Your registration will not be complete until payment is received.
(3) You will be given a receipt when you arrive for the class. If you need a receipt sooner, contact Lothar Kleiner at the above email address and ask for one to be mailed to you.
(4) Reservations will be accepted in the order received until the class size limit is reached. We can accommodate 90 attendees on Day 1 and Day 2 and 50 on Day 3 and anticipate selling out. Early reservations are encouraged. No reservations will be possible after attendance limits have been met.
(5) If someone other than the intended attendee is performing the registration process, please ensure that it is the intended attendee whose name appears in the appropriate field in the registration.
(1) Cancellations by you: allowed until Oct 24 - you will receive a refund minus a fixed $50 administrative cancellation fee regardless of the number of days you registered. You must cancel in writing or e-mail and have a verifiable acknowledgment from us that you have cancelled in time. No cancellations allowed after Oct 24. Registrants who fail to attend and who did not cancel in time will not receive a refund. If you personally cannot attend, another attendee from your organization may substitute (by arrangement only; contact Lothar Kleiner).
(2) Cancellations by us: in the unlikely event that not enough registrations are obtained, the class will be cancelled. If this happens, you will be notified by October 24. Either your credit card or your check will not be cashed or if cashed, you will receive a full refund from the GGPF.
For course content, contact Lothar Kleiner, firstname.lastname@example.org
For GGPF web page issues or registration difficulties, contact email@example.com
If you can’t reach either of the above, contact Clayton Henderson, Clayton.Henderson@HitachiGST.com