Requested Case Study (was "Rule of thumb for SPC")

[Kromkowski@aol.com]

>>If I construct a run chart and find a point or two outside the Upper
Control Limit, I exclude these points and recalculate the new UCL for the
range.<<

It has been my view that one should only exclude a point and recalculate, 
only _if_ you have determined the reason for the "special" cause.

>> Is there some sort of a rule that I can apply that determines the maximum 
number of points that I can exclude?  If say 20% of all the points are 
outside the UCL of the R-chart, do I exclude all? Or is the process so 
chaotic that the rules of SPC do no longer apply? >>

It seems to me that these questions as well as the others asked, want me to 
ask how it is you are using the control chart and what control chart you are 
using.  I think it would be very instructive to the DEN (and for you and your 
organization) for you to present a short case study.  

Tell us _briefly_ and _consicely_, what you are measuring, what chart(s) you 
chose, give us the actual data and tell us how it was gathered, and tell us 
how you have been using or plan to use the chart(s).

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