Stat-Ease
If you are having trouble viewing this email view it online.
 
Vol: 14 | No: 1 | Jan/Feb '14
Stat-Ease
The DOE FAQ Alert
     
 

Dear Experimenter,
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 the previous DOE FAQ Alert click here.

To open yet another avenue of communications with fellow DOE and Stat-Ease fans, sign up for The Stat-Ease Professional Network on Linked in. A recent thread features “Choosing the constant to add to zero when log transforming data.”


 
Stats Made Easy Blog
 
 

-

 
Topics in the body text of this DOE FAQ Alert are headlined below (the expert ones, if any, delve into statistical details):

1:  Software link: Version 9 of Design-Expert® software released!
2:  FAQ: Split-plot design for taste testing
3:  Reader response: Stat-Teaser article on “Regressing the Rupee’s Plunge”
4: Events alert: 5th European DOE User Meeting July 10-11 in Cambridge, UK
5: Workshop alert: Series of classes in San Diego, California
 
 


PS. Quote for the month: Data may speak, but does anyone actually listen?


- Back to top -


1: FAQ: Version 9 of Design-Expert® software released!

We are pleased to announce the release of Design-Expert version 9 (DX9), which features split plots for hard-to-change (HTC) factors, definitive screening designs (DSDs) and many other very useful new statistical tools and wonderful interface enhancements.  Click here* and scroll down to the “Features” tab for more details on what’s new in DX9.  Then, assuming you like what you see, scroll back up to the link for the free 45-day trial or buy it directly**.  You will not be disappointed!

*P.S. Let us know how you like the completely-revamped Stat-Ease web site, keeping in mind it being only a month old and still under development.

“I REALLY like the look and navigation of the new web site!”
—Applied Statistician, Motor Vehicle Industry

**P.P.S. Annual license holders will automatically receive a free upgrade to version 9 in early February.


- Back to top -


2: FAQ: Split-plot design for taste testing

Original question from a Professor of Chemical and Biomolecular Engineering:
“Please thank Mark for his excellent webinar on Real Life DOE.  I plan to ask my DOE class next semester to watch the recording after they’ve become comfortable using DX.

I have been meaning to ask your opinion of the following.  I’ve been using Montgomery’s text for several years and require a term project as he recommends.  They start with a half-fraction of a five-factor two-level design with 16 runs and then follow up with response surface methods (RSM) on the two numeric factors that create the biggest effects.  Many students select some sort of taste test.  I discovered the following problem when multiple tasters are used.  Sometimes tastes are so different for different persons that this can obscure actual preferences.  For example, Americans tend to favor sweet while Indians like spicy.  In such a situation, DX gives misleading results if you use either the average response or the individual responses as replicates.  So I recommend they analyze each person’s response as a separate result.  If these are similar, then you can use the average and, if sufficient persons, the standard deviation as well.  Have you experienced this and, if so, how do you handle it?"

Answer from the Stat-Ease Consulting Staff:
“As advised here by the Society of Sensory Professionals, split-plot designs serve well in situations like this where taste-testers may not be consistent, that is, they interact with the samples being assessed.  As illustrated in the DX9 screen shot below (green circles added to highlight), specify the sweeteners and spices as hard-to-change (HTC) factors and panelists being easy to change (ETC).*  Then start serving up the samples—enough to divvy up to the panel according to the randomized run order provided by the software.

*(Note that choosing whether any given variable should be HTC, ETC or blocked depends on many considerations.  When in doubt, it’s always best to consult with a statistician and/or subject-matter expert.)”

P.S. from Consultant Wayne Adams: “In other situations than the one outlined here, making panelists the HTC factor might be more convenient.  Panel-testing generally focuses on finding a food that works for everybody rather than figuring out the individual effect of people.  Typically, as the professor points out, the effect of individuals is blocked out or averaged.  Changing the blocks to whole-plot groups doesn’t change the experiment at all but it opens up the estimation of person-by-food interactions.”

Split Plot Design for Taste Testing
Split plot design for taste test with inconsistent panelists

(Learn more about factorial split plots by attending the new half-day computer-intensive workshop, Factorial Split-Plot Designs for Hard-to-Change-Factors.   E-mail [email protected] for details.)


- Back to top -


3: Reader response: Stat-Teaser article on “Regressing the Rupee’s Plunge”

Comment from Nico Laubscher, Statistical Consultant, InduStat Pro, South Africa:
“Please accept my congratulations Brooks on this excellently written article on “Regressing the Rupee’s Plunge (The Dangers of Happenstance Regression)” in the September issue of the Stat-Teaser, which explains (or perhaps how not to explain) the decline in India’s currency.  Economists always find one reason or another to explain responses such as the value of the Rupee, etc. I wish they would see your note and take it seriously.

This type of explanation is of course well known on the stock exchange where all sorts of reasons are found by economists to explain the variability in the market. What I liked about your note is how you used DX to do some of the computations. It was nice that you point out that one can see correlations in the graph columns node (although other software may be more efficient to look at many variables simultaneously).  This may have alerted some users at the availability of correlation coefficients when using DX.  Also, the idea of the hold-out data set to verify the model (one of the basics in data mining, given that you have sufficient data) was nice.  Of course, the difficulty with economic data like the value of the Rupee or the index of the stock market is autocorrelation.  The holdout sample may have the problem that it is not representing the current situation if it is selected at random from the entire data set.

Before I rant on, let me say it was a very stimulating read, strengthening some of the views I have about explanations for economic indicators from happenstance data.”

Response from Stat-Ease Consultant Brooks Henderson:
“I am preparing a follow-up article to go in the next Stat-Teaser that shows off the new DX9 color-correlation matrix, which, as Nico suggested would be good, looks at many variables simultaneously.”
Smiley Face

- Back to top -


4: Events alert: 5th European DOE User Meeting July 10-11 in Cambridge, UK

We are pleased to announce the 5th European DOE User Meeting July 10-11 at Downing College in Cambridge, UK, co-sponsored by PRISMTC.  See further details on the event at their conference site, including information about pre-meeting DOE workshops.  The meeting itself and Downing College lodging in particular must be booked early, so you are advised to make your plans as soon as possible. A call for speakers will go out soon.

Click here for a list of upcoming appearances by Stat-Ease professionals.  We hope to see you sometime in the near future!

P.S.  Do you need a speaker on DOE for a learning session within your company or technical society at regional, national, or even international levels?  If so, contact me.  It may not cost you anything if Stat-Ease has a consultant close by, or if a web conference will be suitable.  However, for presentations involving travel, we appreciate reimbursement for travel expenses.  In any case, it never hurts to ask Stat-Ease for a speaker on this topic.


- Back to top -


5: Workshop alert: Series of classes in San Diego, California

All classes listed below will be held at the Stat-Ease training center in Minneapolis, Minnesota unless otherwise noted.  If possible, enroll at least 4 weeks prior to the date so your place can be assured.  Also, take advantage of a $400 discount when you take two complementary workshops that are offered on consecutive days.

*Take both EDME and RSM in the same week to earn $400 off the combined tuition!

** E-mail [email protected] for pricing on the complete series of San Diego classes.

See this web page for complete schedule and site information on all Stat-Ease workshops open to the public.  To enroll, click the "register online" link on our web site or call us at 612-378-9449.  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 a private workshop, which is most convenient and effective for your staff.  For a quote, e-mail [email protected].


- Back to top -


I hope you learned something from this issue. Address your general questions and comments to me at: [email protected].

Please do not send me requests to subscribe or unsubscribe—follow the instructions at the end of this message.
Sincerely,

Mark

Mark J. Anderson, PE, CQE
Principal, Stat-Ease, Inc.
2021 East Hennepin Avenue, Suite 480
Minneapolis, Minnesota 55413 USA

PS. Quote for the month—Data may speak, but does anyone actually listen?

 
"Never underestimate the difficulty of changing false beliefs by facts.”

—Henry Rosovsky, Harvard economic historian

Trademarks: Stat-Ease, Design-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, Shari Kraber, Wayne Adams and Brooks Henderson
—Statistical advisor to Stat-Ease: Dr. Gary Oehlert
Stat-Ease programmers led by Neal Vaughn
—Heidi Hansel Wolfe, Stat-Ease sales and marketing director, Karen Dulski, and all the remaining staff that provide such supreme support!

Twitter-SmileyFor breaking news from Stat-Ease go to this Twitter site.

DOE FAQ Alert ©2014 Stat-Ease, Inc.
Circulation: 6300 worldwide
All rights reserved.


 
  Subscribe