Changing Windows

[Kromkowski@aol.com]

>>Inspector A does have more defects found than the others (using a
u-chart), but if you remove the two 1/3/97 inspections, his/her
remaining inspections are statistically in line with the other
inspectors (average number of defects per inspector).<<

Defects per inspector gets you to a U-chart?  Maybe, I could agree with 
defects per 10s of inspections, plotted by Inspector.

I have to say I'm still confused about what is being measured and how and 
when ("before the change" or, as I had originally perceived "after the change 
to see if the installation was correctly done").  

I think that these kinds of case studies can be helpful, but what happens 
after the data is plotted is just as important.  As I indicated in the "Brick 
Study",  what we should see next in all of the case studies is 

the results of investigation for  "special causes" when appropriate, 

the results of investigation for "common causes" of variation and centering, 

the theories for improvement based on these investigations, 

the tests of the theories and 

the resulting data so that the process of improvement can continue.

JDKromkowski
======================================================================