RE: Question regarding stable systems and correlations

[Rip Stauffer <rstauffer@bluefirepartners.com>]

Hello, All,

I'm not sure I completely agree with the majority on this question, or maybe
we have our signals mixed a bit. 

One difference between common cause and special cause variation, and one of
the reasons we look at it this way, is that common cause and special cause
require different types of action. Special or assignable cause variation
usually allows us to work on one easily identified variable, signalled by
the special cause indicator in the chart. Common cause requires work on the
system. While I agree that it is usually more costly, I'm not sure I agree
that it costs more than running the process badly. If I have a stable system
that produces a lot of bad product, I'd better work on the system (the
common cause system), given of course that I want to stay in business.
"Common cause" does not denote "causeless." It does indicate that there is a
causal system at work. 

Wheeler once described a common cause system as a lot of things happening
inside a system that appear random, and described special cause this way:
"Every once in a while, things sort of line up and resonate, drive
performance in one direction or the other, and you get a signal that lets
you know something is going on." (This is a paraphrase from my notes). His
point was that, when this happens, you get a chance to work on a piece of
the causal system that will reduce the overall variation. This is how the
Japanese at Tokai Rika doubled the performance of their system over the 18
months of "The Japanese Control Chart." This is what I call the passive side
of SPC and improvement; incremental improvement, if you will. 

On the active side, I don't know about anyone else, but I've had quite a bit
of luck with modeling through regression and multiple regression followed by
model testing and optimization using DOE. Of course, maybe I've just been
lucky enough to work on systems that were merely stable, not in a state of
maximum control. 

Certainly, the Red Bead represents a stable system. I suppose you could run
some correllation on the fact that you have 20 percent defective in your
incoming material and you get about 20 percent defective in production. We
don't do the math because it's too obvious, but we could certainly work on
that common cause system and reduce the variation; I get suggestions from
every class as to how we could improve it.


Best regards to all,

Rip

Rip Stauffer, Senior Consultant
BlueFire Partners
1300 Fifth St. Towers, 150 So. Fifth St.
Minneapolis, MN 55402
612-344-1027
mailto:rstauffer@bluefirepartners.com
http://www.bluefirepartners.com/