Shanin Quality Principles

Your potential employer is either (a) a narrow-minded fool or (b) a narrow-minded person of high intelligence, for applying this criterion, because he / she should want someone who can think, not produce a toolbox.  Do a web search for “Shainin”.  Dorian Shainin developed a number of quality engineering tools which allow the user to “listen” to a process and gain profound knowledge.  It is highly intuitive as opposed to statistically based.

I can send you the pages from the Six Sigma Handbook by Thomas Pyzdek who discusses Dorian Shainin’s Pre Control method which he developed in the 1950’s. So it is not a new method. You should be able to log onto a web site and find out more information but if you want these pages sent to you via fax please send me your fax number and I will try and get it back to you the same day as received.
Hope it helps, or at least a start. Let me know..
 

Thanks for the advice.  I bought a book called, “World Class Quality” by Keki Bhote that deals extensively with the Shainin DOE.  I am still suprised at how little I can find about Dorian Shainin.  His DOE technique does seem easier and less time consuming (so far, I am only half way through the book).  Thanks again for all of the help.

I went through the Shainin training duringthe early 90’s. It had some unique approaches to problem solving. The most useful was take a broken product and a good one and find out what is different.
 
Practical buty not very useful in todays complex product offerings such as Software. Another was the red light green light yellow light approach to inspection. Another rather silly approach to practical manufacturing.
Don’t waste your time unless you are investigating methodologies that just did not make the cut.
 
Remember not everyone can understand the complexities of the six sigma methodology so utilize the group knowledge on what works and what does not work.
Make your own conclusions.

Dear Kraig:
You are on the right track by purchasing “World Class Quality”.  Keki Bohte told me personally during a class I took from him that Dorian Shainin licensed him alone to write about his techniques.  This is another reason why you are having problems finding information.
Regarding other comments on this thread, that suggest these techniques are out of date, I agree that they go back a few years but I disagree with regard to improving supplier quality and occasional use.  I have been in several situations where we desperately needed to improve supplier quality and the supplier was desperately afraid of using statistics.  In those situations, I found Shainin techniques very helpful.  I have also used the techniques in house on occasion where I wanted to use very simple graphs to explain variation. 
Good luck with your studies and getting that job you want.
Sincerely,
KN –   https://www.isixsigma.com/library/bio/kniles.asp ; http://www.KimNiles.com

Dear Friend
There is an excellent book on Shinin( not Shanin) DOE by Prof Keki Bhote (Deciple of Prof Shinin)published in USA.
Basically Prof Shinin claims that his method of DOE is statistically much more scientific , simple , economical and accurate than any other method of DOE.
In my opinion this method is best suited for Manufacturing Processes . I don’t know how it can help in Design Process.
This book explains simple techniques of DOE for Problem solving.
I strongly recommend reading and practicing.
 

Stan,
You obviously stayed up late last night drinking. Your response shows your inept knowledge of anything happening in the 21st Centruy.
 
I hope you the best in whatever home your have been placed in.
 
 

Ron,
I forgot to tell you that I agree with your statement of Shainin not being right for things like software.
What you need there is even simpler, like process maps and standardized work, and getting the software folks to recognize they are not the customer.

Shainin techniques were developed decades ago, before the advent of software products like Minitab made the chore of statistical analysis easy. The fact that putting tools like AnOVa and Correlation and Regression in the hands of people with 4 weeks of ”training” doesn’t seem to be alarming anyone. That aside, Shainin tools were developed in the paper and pencil era. If you prefer to rely on a two trial evaluation of the repeatability of your measurement system using an X-Y graph, I say more power to you. I will stick with using the Average & Range method and AnOVa.
Caveat Emptor,
Marc Richardson
Sr. Q.A. Eng.
 

You are not comprehending the scope or DOEs.  Firstly, any DOE experiment, whether it is Taguchi, or traditional, takes time to understand, insofar as its effect.  Secondly, Red X, which Motorola latched onto 18 years ago saw the distinct advantage over the rigors or simple stat analysis.  GM, where I have a colleague and friend working, employed Shainin in one division and the results over Taguchi are astounding.  They have garnered every award and attainable kudo there is from up top over at GM.  The more mathematically rigorous methods do work, however, why not simplify if the results are equal or better?  There is no real question as to which method is superior, when spanning all business types and processes.

Who said anything about Taguchi methods? Don’t get me started on that topic. 
 
What you have told me is where Shainin methods have been applied but you have not answered my challenge so I will state it again,
 
Give me a reason that I would need to employ Shainin techniques when I have access to (and skills in) solid statistical analysis methods
 
You say “Motorola latched onto 18 years ago saw the distinct advantage over the rigors or simple stat analysis”.  Should I use it because Motorola and GM used it? What distinct advantages?  What do you mean by simple stat analysis? Be more specific.
 
Statman

To say that Motorola latched on to Red X is at best misleading and simple is not true.

Stan,
Shainin guards their methodology and techniques with caution.  They are consultants and this is how they make their money.  If you are open minded the Shainin techniques are useable and geared toward ease of use for the average shop operator.  It is like forensic science for the shop floor.  Do not discount some of Shainin ideas.  Keep an open mind. I am an avid Six Sigma guru and degreed statistician, though some of Shainin’s origins can be traced to statistical techniques already named (Tukey’s run test, but Shainin’s B vs C), they are presented in a simplistic nonengineering/statistician jargon which is beneficial to some audiences.
 As far as an employer requesting Shainin is a must on your employment AND you want to work there, try negociating.  Ask them for a trip to Michigan for Shainin’s Apprentice and Journeyman training.  If they won’t pay for this education, then you know this compnay or this specific manager is not worth YOUR time.  Keep looking. 

Can you define temporary. Thanks.

I’m not new to this site but I’m new to this forum and I am astonished at the level of arrogance that some of you have toward those who are trained in Red X strategies.  We are not all that different.  We strive to be the best problem solvers in the world, we strive to save our companies money.  We are the same, we just approach problem solving a little differently.  I use Red X you may use DMAIC, DFSS or DoE’s at the end of the day we solve the problem, we report out and move on the to the next issue.

I think that you just hit it on the head.  Why use six sigma if you don’t need to?  Red X is geared for the more simplistic problems.  Shainin can’t handle multiple interactions as well as six sigma, but it can out duel six sigma in speed to closure for simple problems.  The whole goal as you know is using the y to x strategy by converging on the main x that effects the y. Red X has been refined since 1983. I’m not saying that RedX is the best tool to use, but it is effective in our industry in quickly finding the families of variation in our process with just using a BOB/WOW sample. The simplistic approach makes problem solving available for the average production manager.  The key to this is the simple 9 step approach:
1. Does the RedX live in the parts or the assembly?
2. Which family offers the most leverage with the least effort?
3. Does everything needed to find the RedX live in this family?
4. What are the observed physical/geometric peculiarities that may assist along the way?
5. Which family has the larges source of variation?
6. What side contains most or all the BOBs?
7. What are the environment conditions the component is under?
8. Is the fracture partial or total?
9. Is the void progressive or does it happen all of a sudden?
How progressive are the voids on the 5 previous parts?
How progressive are the voids on the 5 subsequent parts?
Tools such as group or paired comparisons quickly get you down to a managable few possibilities.
 
As you know its all about splitting the dictionary by asking simple questions and observation.  You can easily isolate and leverage differences.  If you end up with 4 factors, you either didn’t do a good enough job in splitting the dictionary or you must use a different tool to find the answer such as six sigma. 
 

If I may sprinkle a little humor here.
We are a company that fully embraces Shainin (Lean and 6s too).
We are also a company that proudly advertises “Wow” products.

No!  I’m not a Shainin Salesman. I have zero benefit of preaching their teachings.  I find that their buisiness practices and isolationist attitudes create a stigma of their teachings and further alienates the professionals that learn and use their tools.  They are brilliant in that respect as they pigion hole us into our positions with little hope of branching out as we will not gain acceptance with the other teachings. With that said.  As you know what you said about doing it right, up front on the engineering side is right on.  But that is perfect utopia and it doesn’t exist right now. We have to have good reactive problem solvers until our DFSS is robust.

Christian,
Obviously our Marketing people aren’t Shainin-trained :-)  We’re an automotive supplier trying to punch our way out of the commodity bag, so to speak. 
We use Shainin mostly for the must-be-fixed-now manufacturing quality problems with well-defined Y’s, and Six Sigma mostly for transactional and non-quality operations improvements.  Sometimes the boundary lines are kinda fuzzy but the Shainin and 6s people still hang out in different camps.  That’s a shame, IMO.  They should all be part of a common toolset.
BC

Mario honey,
Could you be more specific? What do you think is nonsense?

BC,
We are begginning to role out transactional here with the Shainin RedX Masters being the first to go through.  I am one of them, so that will eliminate the dual camps. You are right on with the Shainin tools are great for quick hits with good discriptions of your Green Y.  I hope to become even more dangerous as I go through the transactional process. 
Cheers

Shainin’s isoplot has solved many measurement problems as you well know. Your claim it is stupid to use it gives you no credit, epscially when most semicondcutor measurement equipment is either automatic or semi-automatic. Not to mention your stupid sexist comment. (What is it about some Six Sigma consultants?)As you well know, Shainin proposed labelling each sample and measuring them twice, on two separate occasions, and then correlating the two measurements. What’s stupid about that?

Stan doesn’t necessarily need my input BUT I can tell you that as part of my development as MBB, he insisted on having me and the other MBB’s in training receive Shainin training. 
I am ever grateful to have received this very useful training even in light of the age of computers.  Stan is a supporter of Shainin’s principles….believe me.

Mario honey,
If you think you see me regurigitating SS dogma, you need to learn to read.
Shainin’s techniques are good, but to estimate one aspect of gauging issues visually is crazy if you have better analytical techniques. You can compare a measurement device to itself – repeatibility or one device to another – reproducibility but you can’t do both at the same time.
There are just better ways to do it, I can look at repeatibility and reproducibility and look at interactions of 2 or more systems. Why would I resort to this visual estiation to analyze? I do use the IsoPlot visual as a good way to explain what the better analytical method tells me.

Stan,
You bring up great points and I’m sure you have tons of experience to share with the group, however you destroy all of that with your condescending remarks and in your face rebuttals. Your remarks and short snippets; such as “Honey”, Zombie”, “Garbage Men” are not appreciated at all.  Hey I come to the forum to learn not fight for my relevancy. If its your opionion that I’m doing or using tools that you would not use, give me some alternatives, some constructive criticism. As far as isoplots, thats the only tool I know to use when validating my measurement system.  I’ve completed 23 projects and half of them I have used an isoplot to establish that I have an effective measurement system. Is there another way? If so is it just as quick?

If you have done 23 projects and only established that you have an
effective measurement system, I’ll bet you have left money on the
table. Odds are against finding them effective that often.Yes there are better methods that are quicker and provide better
guidance on how to resolve measurement issues. Look at
Measurement System Analysis – there is a good reference
published by AIAG. I’ll send you a copy if you give me a mailing
address. Don’t post it on here, email it to me @
StanMikel@aol.com.Another good reference is Evaluating the Measurement Process by
Wheeler and Lyday, ISBN 094532006X.As far as your critique of my style, as Popeye says, “I am what I
am”. Look at your posting touting the superiority of Shainin when
all you know is Shainin. I’ll engage in a rational conversation if you will.

Mario
He calls you Honey because he knows it bothers you. And second because your an idiot if you think charting reproducibility of an automatic system is a waste of time.
 

Let’s see .. we study the repeatability of twenty or so automatic measurement systems and find all of them meet the 10% rule. Gee, we’ve now got to go and use ANOVA on dependent samples to find out how much variation is contributed by operators.Whose the idiot?Go write another Six Sigma report flunky.

Mario Sweetie
How do you know when the reproducibility ends or becomes erratic?
20 machines cannot all reproduce the same and therefore have a range within themselves and as a whole. You should really read up on Lean automation principles and statistical charting of these processes.
 

Mike:I have no problem with Saguin – other than it smells and has far too many nodding donkeys.I have no problem with you either, or with Stan, or the other yo-yo.As for reproduceablity – I’ve always used that term in the context of operators. If you or someone else has taken a different approach and now call it equipment ‘matching’ reproduceability, I have no problem with that either. But why didn’t Stan make that point instead of called the Isoplot method stupid?What I do find suprising is that anyone, who is associated with your organisation, would seek to deny the use of the Isoplot within Motorola. Why would they do that … unless they have a selfish motives.Why would someone try to deny the existence of the Isoplot software widely used in Phoenix, Austin, East Kilbride, Toulouse, etc.What is an Isoplot – it’s just a plot of one dependent variable against another and calculate a correlation coefficient and I don’t know of anyone who has plotted one by hand for the past twenty years! Why is it stupid to study the correlation of one piece of equipment against another.More seriously, what happens when one of your client’s employees asks these two yo-yo’s a simple question? Are they going to be called honey, stupid, a zombie, etc. Or do they only have bravado in anonymity?I find it strange that a person of your calibre would associate yourself or try to defend such ignorant people.I’m out of here.

Van Belle has a nice, concise way to quantify the “agreement” of two measurement methods.  The same kind of data you’d get in an Isoplot. It breaks down into three components:  precision (correlation), bias, and scale.  We’ve used it to compare, say, a theoretical estimate with a physical measurement.
Here’s the reference:  http://www.amazon.com/Statistical-Rules-Thumb-Probability-Statistics/dp/0471402273/ref=sr_1_7/103-5804524-8659828?ie=UTF8&s=books&qid=1181240851&sr=8-7

Bivariate normal, to be precise.

Mario honey,Let’s try this one more time.Motorola semiconductor is not the whole corporation. In fact, if you
look at the capital drain and low profit contribution, they were an
insignificant piece.I didn’t say 52nd St was better, I did say they were more in line with
what the corporation was doing. Most of the corporation had ADT
as their primary tool set by 85 or 86.I laid out why IsoPlot is an inadequate tool. Why did you choose to
ignore that?

BC,
Could you explain what is meant by bivariate normal? Are you talking about two variables, each of which are normal? If not, does this imply something about joint probability distributions or something like that?
Thanks!

You need to learn to read.
ADT and Shainin are one in the same and IsoPlot, realistic tolerancing, MultiVari, Component and Variable Search, … were taught.
IsoPlot is a stupid analytical method.

I’ll make this my last post on the subject. You can have the last word.Let me get this straight:- Motorola won the Malcolm Baldrdige Award in 1988 using training methods (ADT) which you readily admit were the same as Shainin’s.- The training material included the use of the Isoplot which you consider stupid- But Motorola achieved considerable sucess using these materials.- After Mikel Harry changed the Shainin Techniques to ‘more advanced methods (tools)’ Motorola began to decline- You dislike Mikel Harry so much you use his name in your on-line email address- Your website claims yourself an inventor of SSI’m not going to call you stupid, an idiot, a zombie, or honey. I’ll let others decide how best to describe your approach to this forum. For myself I think you could be so much more than what you appear to be after reading some of your posts!

Mario W,
The plant doesn’t smell but it is unfortunately located in the middle of some farm land so when it gets fertilized there is a distinct oder in the air. The chicken processing plant on the other side of I-10 probably doesn”t help much either.
In terms of the R&R. If you are running automated test equipment that is not influenced by the operator then reproducibility doesn’t make sense in terms of the operators. If you use multiple machines to test the same product then you need reproducibility between machines. You simply use the different machines in the same way you would different operators and the reproducibility number will represent the agreement between machines rather than operators.
As far as Shanin stuff I have always liked the component search methodology and use it when appropriate.
Regards

Agreed ..Respectfully,
Mario

Mario honey,To your points — Motorola was requiring 40 hours of training for all employees
since 1983. ADT was just one of over 30 courses available in the
SPC series from MTEC. The part of Moto I was with gave no
emphasis to ADT after 1984.- The use of GR&R came into Motorola at least by 1986 and we
found MSA issues everywhere we looked and improved dramatically
by finally straightening out our SYSTEMS.- Mikel Harry did not change anything, he was one of many who
advocated other tools. John Ramberg, Tom Barker, Ed Heard, Skip
Weed, and many others contributed to the body of knowledge.
Mikel Harry had almost zero impact inside of Motorola. He certainly
did not influence Semiconductor. You may find it intriguing that Dr.
Harry advocated neither IsoPlot or MSA in his description of MAIC.
He saw looking at the measurement system as a tiny consideration
before doing a DOE only.- Yes, I have an email address with Mikel as part of it – it happens
to be a family name (easy to prove).- I have no website and I have never claimed to be an inventor of
SS. I claim quite the opposite. SS was the joint contribution of a lot
of smart people, not the least of which was Bob Galvin who allowed
this whole thing to happen. And thank you for the vote of confidence that I could be more.Now, what about the specific issues I explained to support my
assertion that IsoPlot is stupid?

Only Mikel, no Harry.

Stan,Some of us know who were on the committee. You left two people off. One was a woman and the other was Mario Perez-Wilson.From the tone of your posts I can understand why you left the woman out, but did you make a mistake with the other name by not including yourself in the list :-)Bernhardt

Christian,
If it works for you that is great keep doing it because the people that run your business in all probability only care about the results. If you deliver that you are fine. I hope you have more bandwidth incase you find that problem where it doesn’t work. How about that situation where there may be a more efficient way?
I have put this analogy up before but I still like it. There is a movie called “Tin Cup” where Kevin Costner is a golfer and gets into a situation where he breaks all his clubs except his 7 iron. Still shoots a good round and qualifies for a tournament. He is bragging about his accomplishment later and Don Johnson (another golfer) listens to his bragging and then asks him “why?” When I see people who are sworn to one guru, one tool, one methodology, etc it makes no sense. Essentially they have broken all their clubs except one. Management doesn’t hire a person because they only run Taguchi experiments or are a devout Shaninite. They buy results. when you accept a job you should be commiting to deliver results not prove that some guru, methodology, etc it is light and the way and you will never get to heaven with out some stamp on your butt.
Just my opinion.
Good luck

I can’t believe this thread is still going. If it works use it. Shanin or Six Sigma, Isoplots, or annova’s, if you can get the end results that satisfy the customer, who cares. Unfortuanetly we have been brained washed by some organizations that Six Sigma or Shanin is the only methodology and which ever one your not using sucks. I have absolutely no Shanin experience at all, do I think its a bad way to do things, NO, I have just not been involved in that culture. As for Mario and Stan, Its nice to see that some people still have passion about what they do, even if its a little immature at times. So how bout lets all agree to disagree and move on to what this Forum is all about.
CT

Christian,
I hope it works for you.
The real emotional high for this whole thing is fixing problems especially if they are chronic problems. There was an earlier comment about Americans not learning from anyone else. We have deployed in 18 countries other than the US. We learn stuff everwhere we go. That is the fun of doing this.
Best of luck in your new venture. If I can help you in any way please contact me at mike.carnell@CSintlinc.com.
Good luck

Mike:Of all the people who visit this forum I didn’t think you’d miss the point.No one doubts the contributions of Mario – or any other Motorolan at the time – including Mikel Harry and others who worked with Motorola, such as Juran, Shainin, and many others.That was not the point of my post.My point is if you seach for Mario’s name on this forum, you won’t find it. If he was here as you claim, why are his contributions missing?Therefore, we can only conclude he used a nom de plume. The question is why. We can only assume it was to call Mikel Harry, Shainin, and others who ask questions fools, idiots, and zombies.Just the opinion of someone who has been insulted by Stan a number of times over the years.

Ronaldo,
Yes it is a nom de plume, as are most names on this and any other forum. Mike Carnell is one of the few who post under their own complete name. Why? Answer the question yourself, you don’t use your complete name assuming Ronaldo is part of your given name.
I don’t post under any name to call anyone fools, idiots, or zombies. You may want to check, but I have only ever spoken with respect about Dorian and I’ve never called Mikel a fool, idiot, or zombie. I have said he is deceitful and takes credit for the work of others, because it’s true. I’ve called the whole 1.5 sigma proof smoke and mirrors because it is. I’ve made fun of his books because, with the exception of the one written prmarily by Schroeder, they are exercises to either have an ISBN # to sell as intellectual property (the white and green cartoon books) or intentionally meant to confuse.
I don’t recall insulting you directly, can you refresh my memory since you made the claim?

Blackberry! You should have held out for the iPhone – arriving June 29.iSixSigma by Blackberry is a pain because it takes soooooo long to
load all of the ads and mine at least will reverse the colors if I get
impatient and scroll down anyway.

ronaldo,
I hope the links from clb1 have satisfied you that Mario has published under his full name. If you read Mario’s book “Six Sigma” you will find he does a discussion concerning the 1.5 sigma shift and how he feels about it. Pretty straight forward and not under a nom de plume.
It is a pretty big jump to go from using a nom de plume to a the assumption that it is done for the purpose of calling Mikel, Shanin and others who ask questions fools, idiots and zombies. Actually I have used the term tool zombie on several occasions myself and I publish under my own name.
Just my opinion.

Stan,
Sounds exciting. We watch the fireworks over the Kemah Boardwalk from the back dock at the house on Thursday and Friday evenings.
Not sure about that Ratatouille. We were in Botswana last month and were offered a local delicasie called aerials. Turns out they are dried rats cooked over a fire. By the time they are done they resemble an aerial. I had to pass.
Regards