DEN and Six Sigma

["Dr A D Burns" <aburns@q-skills.com>]

John Hunter's comments are correct in that most quality philosophies have
aspects that are useful and some may follow Demings 14 points to greater or
lesser degrees than others.  Each quality philosophy seems to have a "life"
of about 4 years, before a new wrapper is applied to the old content.  Six
sigma is no different.  It still uses the basic seven tools of quality (and
throws in a few dozen more) and still talks about reducing variation.
However, the difference with six sigma, is that its fundamentals are flawed.
That is, there are flaws in the very nature of the "SIX sigma" stuff :

1. It is a specification based philosophy.  Defects counts relate to the
specification, not just the process.
2. Defects counts are an incorrect way to estimate sigma for variables.
3. Most processes do not follow normal distributions (call centres please
note).
4. The +/- 1.5 sigma floating mean used in 6 six sigma calculations is
incorrect (see my previous note on its origins)
5. Implications that "6" sigmas are better than "3" is as much nonsense as
saying that feathers are better than fruit.

As Wheeler states " the best that can be said is that the (six sigma)
numbers represent a triumph of computation over common sense."

Regards,

Dr Tony Burns
aburns@q-skills.com