Hot new workshop:
Designed Experiments for Medical Devices
May 8-9, Minneapolis, MN
Dear Experimenter,
Here’s a fresh set of answers to frequently asked questions (FAQs) about design of experiments (DOE); plus timely alerts for events, publications and software updates. Check it out!
- Mark
My wry look at all things statistical and/or scientific with an engineering perspective.
Also, see the Stat-Ease blog here for tips on making DOE easy. For example, a recent posting provides a heads-up on “Design of Experiments (DOE): The Secret Weapon in Medical-Device Design”. Check it out!
Topics in the body text of this DOE FAQ Alert are headlined below (the “Expert” ones, if any, delve into statistical details):
  1. Workshop alert: New workshop - Designed Experiments for Medical Devices
  2. FAQ: Real or actual components—which should I use for mixture design?
  3. FAQ: Why is my mixture-component trace plot backwards?
  4. Expert FAQ: Can Design-Expert® software suggest a formulation that goes outside the initial design space?
  5. Info alert: “Know the SCOR for a Multifactor Strategy of Experimentation: Screening, Characterization, Optimization, and Ruggedness (SCOR) Testing” published in March ITEA Journal
  6. Events alert: Join us at the DOE/Design-Expert Users Meeting (a part of the 2019 Analytics Solutions Conference). Time is running out to reserve your place at special pricing!
  7. Survey Alert: Professors seek input on the use of supersaturated designs
P.S. Quote for the month: Fisher’s views on one-factor-at-a-time (OFAT) experimentation.
(Page down to the end of this e-zine to enjoy the actual quote.)
Workshop alert: New workshop - 
Designed Experiments for Medical Devices
You can do no better for quickly advancing your DOE skills than attending a Stat-Ease workshop. Our expert instructors provide you with a lively and extremely informative series of lectures interspersed by valuable hands-on exercises. Enroll at least 6 weeks prior to the date so your place can be assured—plus get a 10% “early-bird” discount.

Designed Experiments for Medical Devices (DEMD) New! 
May 8-9, Minneapolis, MN
Mixture and Combined Designs for Optimal Formulations (MIXC)
May 14–16, Cleveland, OH
Experiment Design Made Easy (EDME)
May 29-30, Orlando, FL
Modern DOE for Process Optimization (MDOE)
May 29-31, Orlando, FL

See this web page for complete schedule and site information on all Stat-Ease workshops open to the public. To enroll, scroll down to the workshop of your choice and click on it, or email our Lead Client Specialist Rachel Pollack at If spots remain available, bring along several colleagues and take advantage of quantity discounts in tuition. Or, consider bringing in an expert from Stat-Ease to teach a private class at your site.*

*Once you achieve a critical mass of about 6 students, it becomes very economical to sponsor an on-site workshop, which is most convenient and effective for your staff. For a quote, email
FAQ alert: Real or actual components—which should I use for mixture design?
Original question from a Senior R&D Manager: “I have always used Design-Expert to create my formulations, but admittedly, I never understood the difference between real and actual. When I go to the lab, it looks like I don’t want to create batches using the pseudo values. Real and actual look to be the compositions I want to use. But which of these two codings should I use?”

Answer: To keep things simple, go with the Actuals rather than the coded values (Real or Pseudo). For what it’s worth, the Pseudos get used 'under the covers' for all the calculations. The Real values are an intermediate code that can be interpreted as the proportion of a component to the total of the active ingredients. For example, let's say you experiment on the viscosifiers making up 12 wt % of the formulation, holding the other 90%, everything else, fixed in composition. A 6% actual concentration for a given viscosifier then translates to 0.5 on the Actual scale--it being 50% of the active ingredients.

For more details see Chapter 4 in Formulation Simplified, a practical paperback based on the Stat-Ease mixture class. It will serve well as a replacement for your missing notes.
(Learn more about mixture design by attending the three-day computer-intensive workshop on Mixture and Combined Designs for Optimal Formulations (MIXC). Click the title for a description of this class and registration details.)
FAQ: Why is my mixture-component trace plot backwards?
Original question from a Development Chemist: “I am somewhat confused by some Design-Expert output from a recent experiment. I ran a 6-component mixture experiment. I’m using a reduced quadratic model with good fit and no data transformation. The trace plot has me very confused. It seems to be backwards.”

Answer from Stat-Ease Consultant Shari Kraber: “In your case, Design-Expert defaulted to upper “U” Pseudo coding because the modeling works better by this inversion from the usual lower “L” Pseudo coding. This causes the graphs to appear reversed due to axes going low to high from right to left instead of the other way around. You can go back and reverse this, but we do not recommend this due to it creating inferior statistical results.”

For details, see Appendix 4A in Formulation Simplified: “Upper (“U”) Pseudo Coding to Invert Mixture Space”. – Mark
Expert FAQ: Can Design-Expert software suggest a formulation that goes outside the initial design space?
Original question from a Statistician: “Is there a way to use a model to suggest a formulation that is extrapolated outside the initial design space, and which pushes the overall desirability even higher? I expanded the limits on the components outside the initial design parameters, but the optimization algorithm did not seem to extrapolate. Beyond this, I’m interested in a strategy like steepest ascent, that is, a strategy for follow-up experiments that explore an optimal path.”

Answers from Stat-Ease Consultants Shari Kraber and Pat Whitcomb: “First expand the eligible design space by rebuilding the design with a larger space, and then copy the original data back into the new design. Finally, model and extrapolate.” – Shari

Pat adds: “Here’s a procedure for steepest ascent for mixtures after the initial design is run and the optimum found on a boundary:
  1. Set up a pure simplex for components in the original design.
  2. Copy the runs from the original design and paste them into the simplex design – use the Real values.
  3. Setup a response for Euclidian distance (Ed) for the next run from the optimum point.
    Ed = SqRt(Σ(x’i – xopi)2) - use the Real values.
  4. Decide on an Ed step size: something like 0.05 seems reasonable; this results in 20 steps to cover the simplex from base to vertex using real values.
  5. Run the optimization with an “in range” criteria on Ed having no lower limit and an upper limit of 0.05 to get the first run on the path of steepest ascent.
  6. Repeat step 5 increasing the upper limit by the step size each time to develop more runs the path of steepest ascent.
I haven’t tried this but think it will work. If there is interest I can develop this further and work out a couple of examples.”
Info alert: “Know the SCOR for a Multifactor Strategy of Experimentation: Screening, Characterization, Optimization, and Ruggedness (SCOR) Testing” published in March ITEA Journal
The March ITEA Journal, published by the International Test and Evaluation Association, presents a paper I wrote, titled “Know the SCOR”, which maps out a tried-and-true multifactor strategy of experimentation comprised of screening, characterization, optimization, and ruggedness testing (SCOR). Each step along the way is illustrated in a comprehensive case study on a welding process. The manuscript is posted here.
Aided by Design-Expert, Houghton International chemists used DOE for “Designing a Metalworking Coolant for Multiple Applications in Half the Time”. Find out how they did it by reading this Fabricating & Metalworking case study.
Events alert: Join us at the
DOE/Design-Expert Users Meeting
(a part of the 2019 Analytics Solutions Conference)
The DOE/Design-Expert Users Meeting, now in the US! On the heels of a highly successful series in Europe, we have brought this popular design of experiments (DOE) users meeting to the United States. Come meet your fellow users of DOE and Design-Expert software. Discover new ways to improve your DOE skill set. See what's new with Design-Expert.See the outstanding lineup of keynote speakers, session speakers and register here. Early bird registration ends April 15th, so hurry!

I will give a talk on “Strategy of Mixture Experimentation” to the “Using Data Science in Chemical Research” symposium at the Central Regional Meeting of the American Chemical Society in Midland, MI on June 5th. Register for this dynamic conference of chemists here.
Click here for these and other upcoming appearances by Stat-Ease professionals.
Survey Alert: Professors seek input on the use of supersaturated designs.
Our friends Byran Smucker and Maria Weese at Miami (OH) University need your input:
"We are seeking information from practitioners of experimental design to determine the current design and analysis methodologies for screening and sequential experimentation; awareness of supersaturated designs and their presence/absence in practice; and concerns about including supersaturated designs within a sequential experimentation framework.
This short survey includes multiple choice and free response questions concerning the practice and theory of screening designs, particularly supersaturated design. The results of this study will help inform us, and the larger research community, on the current practices and experiences of those using experimental design techniques.
We appreciate you taking the time to answer these questions and welcome all comments and feedback, which can be sent to Maria Weese (
Access the survey here:
Thank you for taking the time to fill this out."
I hope you learned something from this issue. Address your general questions and comments to me at:



Mark J. Anderson, PE, CQE
Principal, Stat-Ease, Inc.
1300 Godward St NE
Minneapolis, Minnesota 55413 USA
P.S. Quote for the month: Fisher’s views on one-factor-at-a-time (OFAT) experimentation-
“No aphorism is more frequently repeated in connections with field trials, than that we must ask Nature few questions, or ideally one question at a time. The writer is convinced that this view is wholly mistaken. Nature, he suggests, will best respond to a logical and carefully thought out questionnaire; indeed, if we ask her a single question, she will often refuse to answer until some other topic has been discussed.”
- R. A. Fisher (1926) The arrangement of field trials. Journal of the Ministry of Agriculture of Great Britain. 33, 503-513.
Trademarks: Stat-Ease, Design-Expert and Statistics Made Easy are registered trademarks of Stat-Ease, Inc.

Acknowledgements to contributors:
  • Students of Stat-Ease training and users of Stat-Ease software
  • Stat-Ease consultants: Pat Whitcomb, Martin Bezener, and Shari Kraber
  • Stat-Ease programmers: Hank Anderson, Joe Carriere, and Mike Brownson
  • Stat-Ease business staff—Cathy Hickman, Greg Campbell, and Rachel Pollack—who provide such supreme support!
Circulation: Over 7000 worldwide
Copyright © *|CURRENT_YEAR|* *|LIST:COMPANY|*, All rights reserved.

Our mailing address is:

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.