RE: Where is the Evidence?

["Carlos H. Mendez S." <qteam@guate.net>]

Frank Voehl wrote:" I have often wondered about the 90% rule of Dr. 
Deming's:  that 90% or more of variation is caused by the system....Where 
is the evidence?"

I have made myself the same question, and since I have not been able to 
find the evidence, I decided I would search for it, if there was any,  in 
the systems I have been involved in.

I decided to do two little experiments. One in a lab setting, where the 
process is highly technology dependent, and for that matter, I assumed that 
the results would be largely dependent on the process itself. The second 
setting was in education, where I assumed that the results would be less 
dependent on the educational process, and more on the students.

CASE 1: Laboratory setting.
In the first case, we chose two lab operators, and had their boss train 
them in a new analytical method that used technology as the bases for the 
analytical results, instead of titration, for example, which demands more 
skill from a person. We had them both analyzed 25 samples of the same 
analytical standard  and under presumably same conditions, and plotted the 
control charts. We did this every day, until the operators, through their 
training and elimination of special causes,  had achieved statistical 
control.

Since they were both analyzing the same chemical standard, with the same 
equipment, with the same chemicals, the difference in the control chart 
from the operators should be more or less related to the effect of the 
operator itself.

Interestingly, both charts had the same total random variation, but their 
averages where shifted just a little bit one from the other. One operator 
reported an average of 54.2% in the standard composition, while the other 
reported 53.6 %, or something like that. I don't remember the exact data, 
but I do remember the shift in the average was considerable less than 1%. 
In other words, 99% came from the system they were in.

This one percent can also be disputed as being considerably less, as it was 
assumed that many factors did not affect, for example, the way the 
operators learned and were trained. Even so, in this case, it was clear to 
us that more than 99% of the results came from the process itself.

CASE TWO: Educational setting.
I have talked about this before in other posts. What I have done here is 
try to stratify the data from my students evaluations to find the different 
ways the education process might be affecting them. As I have said in other 
posts, things I have done, for example,  is see if there is a correlation 
between an specific test grade and the number of classes my students where 
having while taking my course. In this case I have found that a students 
grade could easily be lower by 10 points with every additional class he/she 
was having. For example, a student who gets a grade of lets say 45, and is 
taking 3 simultaneous classes, could easily have gotten 55 if he/she was 
having only 2 classes. In this case, around 15% of his/her grade is depe  
ndent on how many simultaneous classes the student is having.

I have done similar studies with exam's format, and other factors. In 
general, what I have found is that around 60% of my class grade is system 
related.

The above two cases might not be very scientific, nor conclusive, but they 
have been useful for me to find value in taking into account the high 
impact systems have in the performance of people and processes, even in 
settings where it might seem systems don't have a large impact.

 As a manager, I have come to conclude that I will learn more of the system 
I'm managing, if I assume that most of the results come from the system 
itself, than from the people. What I have observed, is that managers who do 
the other way around, quickly abandon the search to understand, and fall 
into blaming individuals,  limiting themselves from a learning that might 
result profitable to their companies.

Hope this is useful, and thanks for reading,

Carlos Mendez
Invenia, S.A.
Guatemala, Central America
qteam@guate.net <mailto:qteam@guate.net>
"I used to have three theories on how not to have babies.
Now I have three babies, and no theories!"
"Yo tenia tres teorias sobre como no tener bebes.
?Ahora tengo tres bebes y ninguna teoria!"
                                                        Anonimo
============================================================================