Deming Discipline*

[David Kerridge <dfkerridge@mac.com>]

Robert Ferrell asked

.............................. Is there one educational discipline that is best
suited for the Deming Philosophy? If so, what is it? What would you
  recommend to someone, wanting a Deming education, just starting college?

...................................... what specific educational 
disciplines are
most in harmony with Deming Philosophy?

As Jean-Marie Gogue said, in his reply, the Deming philosophy is
a new discipline. Unfortunately it is one that is not (as yet) taught.
So we must consider what is presently available.

WED never called himself anything other than a "Consultant
in Statistical Studies". In "Out of the Crisis" chapter 16 he
suggests that an expert adviser to management should have
a research level education in statistics, and then considerable
practical and theoretical experience, gained by studying under
a master. He explained, when asked in a seminar, that by master
he meant "master statistician".

The difficulty about that is that most statistical education is
so very bad. Bad enough to be one of the "Diseases and
Obstacles" in Chapter 3 page 131.

Shewhart mentions two stumbling blocks to understanding
his new theory. The need to understand statistics, and the need
to understand "certain concepts from modern physics". It is
worth remembering that both Shewhart and Deming were
research Physicists first, and then self-taught statisticians.

That must have helped them to think about statistics as
a branch of scientific method, rather than the way it is
often presented, as a branch of mathematics. Of course,
probability theory *is* a branch of mathematics, and it is
useful to understand it. But inference must be understood
as scientific method. Abstract mathematics encourages
abstract thinking, which is fine. But the real skill is to
integrate this with concrete and practical insight.

I remember a student complaining "When you talk
mathematics, I understand it. When you talk about
practical things, I understand it. But when you keep
switching, I get lost." I replied "That's what statistics
is about."

Experience in statistical consulting is very valuable because
it requires you to work with experts of many kinds. Many of
them, though intelligent and thoughtful, are inarticulate
outside their chosen field. They may feel that assumptions
that they make without thinking are obvious to everyone.

The statistician, to be more than a hack,  must translate these
unstated assumptions into statistical models, and explain
them well enough so that the subject expert can check the
correctness of this translation. And then translate the
result of using the models back to the expert.

So the statistician is compelled to work to work closely with
others, and to understand their ways of thinking. This ability
to understand that there are other ways of thinking than one's
own is rarely taught, but is essential.

So this appears to be what Deming said, and my view of
why he said it. But all knowledge is helpful in understanding
the System of Profound Knowledge. We must not put non-
statisticians off. In fact I suspect that the best amount of
statistics to learn in academic courses is either everything or
nothing. It is easier than most people suppose to teach
yourself, though a good grounding in mathematics is
helpful. Statistical principles can be learned without
mathematics, but there are few if any texts to help you
do it. But I have worked with medical researchers who
had a better grasp of statistical principles than most
academic statisticians.

Many "introductory courses in statistics" given as part
of degrees in other subjects do more harm than good.
So perhaps this is why WED suggested that statistics
should be studied to the level at which you can do
research in it. By then you should have seen the weaknesses
of current teaching as well as its strengths.

But above all, good teaching, regardless of the subject matter,
is valuable. Good teaching is based on deep thinking about
concepts and principles, reasoning and questioning.

Hack teaching, based on cramming facts and rote learning
is harmful. Even worse if it is based on brilliant and
persuasive presentation without real depth. It is so much
harder to unlearn. Hack teaching of statistics is common.

For the present, we should also remember that there is
little demand for specialists in the Deming Philosophy.
So to have an expertise that is in demand may be a life-
saver. Dr. Deming himself earned his living as an adviser
on surveys for many years, while his other skills were
not appreciated.


-- 
Best wishes

David

dfkerridge@mac.com