Re: Pareto and others

[Mike Townes <mdtownes@iamerica.net>]

At 6/1/99 10:08 AM, rderoeck@alphaind.com wrote:
>In Wheeler and Poling's book: "Building Continual
>Improvement" they write about a Flat Pareto chart where
>the 80/20 rule does not apply. They argue that flat Paretos
>(a changing major defect category each month/week) is a
>result of unpredictable processes upstream. Doesn't it
>therefore seem logical that there is some relationship
>between the Pareto principle and commom/special cause
>of variation?

I can see where unpredictable processes upstream could lead to a flat 
HISTOGRAM; we teach that as being mixed processes.

I see a lot of PARETO diagrams that show no young mountain, but rather a 
gradual, gentle and rather consistant slope from the highest to the lowest 
(think 100, 99, 98, 97, etc.).  My experience, and Kume's Book, suggest 
that this may simply mean that I have not viewed the problem from enough 
perspectives to gain appreciation of true causal factors.  For example, I 
might look at some sick folks as a group and begin to investigate.  If I 
stratified them by height, I might find that they represent the general 
population and show no significant hump (another SCIENTIFIC word, like 
YOUNG MOUNTAIN).  If I looked by age, the same thing might happen.  If I 
looked by hair color, the same thing might happen.  It could be, though, 
that I would find all the sick folks in a large hump called MALES.  That 
being the case, I might investigate further and find that the issue is 
prostrate cancer, something not possible for females to get.

This example may be crude, but is intended to cement the notion that good 
investigation probes from many angles or perspectives.  Most of those 
"chapters" in the investigation produce no significant YOUNG MOUNTAIN, but 
if one keeps looking diligently, the cause will eventually surface..

All this is to say that good quality work is more like Columbo, or Murder 
She Wrote or Mrs. Polifax than it is like John Wayne who comes in with guns 
blazing, slays the black-hatted scoundrel and rescues the damsel.  A PARETO 
diagram is simply one way to separate perspectives that do not make a 
difference from those that do.  I believe Wheeler calls this finding the 
signal among the noise.

Michael D. 'Mike' Townes, Quality Specialist, USPS
mailto:mdtownes@iamerica.net
214-819-8797, PO Box 225459   Dallas, TX  75222-5459
====================================================================