Issue: Volume 2, Number 4
Date: April 2002
From: Mark J. Anderson, Stat-Ease, Inc.
Here's another set of frequently asked questions (FAQs) about doing design of
experiments (DOE), plus alerts to timely information and free software updates.
If you missed previous DOE FAQ Alerts, go to the links below.
Feel free to forward this newsletter to your colleagues. They can subscribe
by going to http://www.statease.com/doealertreg.html.
I offer the following links as cold
appetizers:
http://nsidc.org/news/press/larsen_B/2002.html
shows with images and movies how an ice shelf the size of Rhode Island (larger
than Luxembourg!) broke off from Antarctica this winter. As a resident of the
frozen far north of the mainland portion of the USA in the middle of a continent,
I can't get too upset about ice melting and oceans rising, but I imagine it's
unsettling for many people on this planet. In any case, this will stimulate
discussion on the reality of global warming and its causes.
http://www.peepresearch.org/nitrogen.html
explains why Peeps cannot be found in Polar regions. I think all the Peeps came
to my home for the recent Easter holiday. They won't last long!
Here's what I cover in the body text
of this DOE FAQ Alert (topics that delve into statistical detail are rated "X"):
1. FAQ: Why equations in actual terms provide misleading
information about the relative effects
2. Address alert: Stat-Ease is moving to a new suite (#480)
3. Simulation alert: Try your hand at baking better pizza
4. Reader feedback: A look at DOE for non-manufacturing
areas such as marketing (see Volume 2, Number 3, March 2002, FAQ #1)
5. Info alert: Link to the latest review of Design-Expert®
software
6. Workshop alert: Tuition has been reduced for private
DOE workshops presented at your site
PS. Quote for the month - Knowledge knows no boundaries.
1 - FAQ: Why equations in actual
terms provide misleading information about the relative effects
-----Original Question-----
From: Taiwan
"Both coded and actual factor equations [shown below]
give the same predictions; but why do the A, B and C factors show the effect
to the response in different ways? For Final Equation in Terms of Coded Factors,
the response was inversely proportional to these three factors, while for Final
Equation in Terms of Actual Factor, the response was proportional to these three
factors."
Final Equation in Terms of Coded Factors:
Y = 12.99 - 0.93A - 2.06B - 0.24C + 0.009375A^2 - 0.38B^2 - 0.14C^2 - 1.19AB-0.22 AC-0.18BC
Final Equation in Terms of Actual Factors:
Y = 7.01 + 84.43A
+ 10.71B + 0.20C + 13.09A^2 - 6.29B^2 - 0.001596C^2 - 180.89AB - 0.87AC - 0.076BC
Answer (from Pat Whitcomb, consultant
and principal of Stat-Ease):
"To get the actual equation, replace each term in the coded equation with
its coding formula*:
Coded = (Actual value - mean)/(range/2)
For example, the coding formula for factor A [in this case] is:
Acoded = (Aactual - (0.028 + 0.082))/((0.082 - 0.028)/2)
Substituting the coding formulas
for all factors in the coded equation creates many changes. The impact depends
on the class of terms:
- linear term replacement results in new linear coefficients and corrections
to the intercept
- quadratic term replacement results in new quadratic coefficients and a further
corrections to the intercept
- interaction term replacement results in new interaction coefficients, further
corrections to the each main effect in the interaction and more corrections
to the intercept.
The corrections from the interactions
can be large and opposite in sign from the linear terms. They can change the
sign on the linear terms. This is what happened to you. Always use the coded
equation to explain effects. Although both equations give the same predictions;
only the coded coefficients are proportional to the observed effect. This is
what provides the motivation to code factors in the first place."
*[Refer to the section on "Coding
factor levels" in the "Engineering Statistics Handbook" at http://www.itl.nist.gov/div898/handbook/pri/section7/pri7.htm.)
(Learn more about developing good
predictive models by attending the "Response Surface Methods for Process
Optimization" workshop
(see http://www.statease.com/clas_rsm.html).]
2 - Address alert: Stat-Ease is moving to a new suite (#480)
Over the next week or so, Stat-Ease
will move up from its current suite on the first floor to a larger one on the
4th floor of the same building. Our new address is:
Stat-Ease, Inc.
2021 E. Hennepin Ave., Suite 480
Minneapolis, MN 55413-2726
Note that the last four digits of
our zip code changed along with the suite number. Our telephone, fax and e-mail
remain the same but due to movement of lines there may be a short disruption
in these communications. Thank you for your patience.
Next month I will provide photographs
of our new office area. The most important addition will be a training room
for our computer-intensive workshops.
3 - Simulation alert: Try your
hand at baking pizza
Bill Hathaway of MoreSteam.com
offers a new simulation* for you to try. He says "Here is a sneak peek
at the latest DOE animation: http://www.gramtel.net/customers/moresteam/index.html (Update 3/07: Link no longer available ).
I have not yet developed an introduction, but the basic story will be about
a high-end pizza delivery chain using wood-fired ovens that is getting complaints
about taste and crust consistency. They want to be able to use cheaper flour
(not imported from Italy) and have trouble consistently getting Hickory wood
to fuel the ovens. So they want to know how these factors (and others) effect
taste and consistency."
Feel free to try your hand at making
better pizza. Then let me know what you think. Feel free to make suggestions
on modifications. I will pass your comments on to Bill.
*(The pizza-making simulation is the second one developed by MoreSteam.com.
In my January e-mail I provided a temporary link to a beta version of a simulator
that shoots virtual boulders from a trebuchet. Eventually, Bill hopes to put
several simulators together for use in their Six Sigma training on DOE. Stat-Ease
has no financial interest in MoreSteam.com, but they use "DOE Simplified:
Practical Tools for Effective Experimentation" (co-authored by me and my
partner Pat Whitcomb) as their DOE textbook, which comes with a free trial copy
of Design-Ease® software. See http://www.statease.com/doe_simp.html
for details on this book.)
4 - Reader feedback: A look at
DOE for non-manufacturing areas such as marketing (see Volume 2, Number 3, March
2002, FAQ #1)
-----Original Message-----
From: Fred
Re: DOE FAQ Alert, Volume 2, Number 3 - March, 2002 (see http://www.statease.com/news/faqalert2-3.html)
"Thank you for sharing! I recently completed Six Sigma Black Belt training. The training included DOE at a level that gave us just enough information to be dangerous. I am very interested in the application of DOE to service industries. As Drucker has said, "All companies are service companies, some of them also produce products." More than 70% of our economy is currently a service economy within which I have focused most of my quality management experience. I am struggling with the complexity of DOE and whether it is provides greater value than other methodologies in delivering quantitative understanding of key variables in service industries.
FYI, I am on your mailing list because I attended one of Stat-Ease's seminars at an ASQ [American Society of Quality] conference probably 6-7 years ago. I am truly thankful for your generous sharing of your knowledge. Thanks for your help."
Answer:
Fred, I'd be surprised if you can find any methodology other than DOE that will
do better at delivering quantitative understanding of key variables in service
industries, particularly the interactions that almost certainly affect the outcome.
The interactions are the gold nuggets you want to screen out with the aid of
two-level factorial designs. Stat-Ease will be leading a session on "The
Use of DOE in Non-Manufacturing Environments" at the Spring Research Conference
of the Quality and Productivity (Q&P) section of the American Statistical
Association (ASA) on June 5-7 in Phoenix, Arizona. It will be a great opportunity
to hear how DOE and other statistical tools can be used to enhance productivity
and improve the quality of products and services. For more details, see http://ceaspub.eas.asu.edu/ie/2002qprc.htm (no longer available).
5 - Info alert: Link to the latest
review of Design-Expert software
The April 1 edition of "Analytical
Chemistry" features a review of Design-Expert software, version 6. It's
posted on our site: http://www.statease.com.
Look for the icon of the magazine. Click on this or the text "Design, Analyze,
and Optimize with Design-Expert" to see the complete article written by
Katherine T. Alben from the New York State Department of Health. She provides
a fair assessment of how our software performs for the specific applications
of interest to the readership of this particular publication. However, it gives
anyone interested in doing DOE an idea of what Design-Expert will do for them.
6 - Workshop alert: Tuition has been reduced for private DOE workshops presented at your site
See http://www.statease.com/clas_pub.html for schedule and site information on all Stat-Ease workshops open to the public. To enroll, click the "register online" link at our web site or call Stat-Ease at 1.612.378.9449. If spots remain available, bring along several colleagues and take advantage of quantity discounts in tuition.
If you've got a group of 6 or more professionals that need DOE training, consider bringing in an expert from Stat-Ease to teach a private class at your site. We recently reconfigured our pricing to make smaller classes more cost-effective. Call Sherry Klick, Stat-Ease Workshop Coordinator, at 612.378.9449, extension 18 to get a quote.
I hope you learned something from this issue. Address your questions and comments to me at:
Mark J. Anderson, PE, CQE
Principal, Stat-Ease, Inc. (http://www.statease.com)
Minneapolis, Minnesota USA
PS. Quote for the month - Knowledge knows no boundaries.
"Science knows no country,
because knowledge belongs to humanity, and is the torch which illuminates the
world. Science is the highest personification of the nation because that nation
will remain the first which carries the furthest the works of thought and intelligence."
- Louis Pasteur
Trademarks: Design-Ease, Design-Expert and Stat-Ease are registered trademarks of Stat-Ease, Inc.
Acknowledgements to contributors:
- Students of Stat-Ease training and users of Stat-Ease software
- Fellow Stat-Ease consultants Pat Whitcomb and Shari Kraber (see http://www.statease.com/consult.html
for resumes)
- Statistical advisor to Stat-Ease: Dr. Gary Oehlert (http://www.statease.com/garyoehl.html)
- Stat-Ease programmers, especially Tryg Helseth (http://www.statease.com/pgmstaff.html)
- Heidi Hansel, Stat-Ease marketing director, and all the remaining staff.
Interested in previous FAQ DOE Alert e-mail newsletters? To view a past issue, choose it below.
#1 - Mar 01, #2 - Apr 01, #3 - May 01, #4 - Jun 01, #5 - Jul 01 , #6 - Aug 01, #7 - Sep 01, #8 - Oct 01, #9 - Nov 01, #10 - Dec 01, #2-1 Jan 02, #2-2 Feb 02, #2-3 Mar 02, #2-4 Apr 02 (see above)