Truly not my strong suit either - but from what I understand:
Why use it? ..it's quick - few calculations -just "Count/Graph then compare to what can be considered a "Normal population curve"... result is a "Probable estimation of % outside or inside spec.. in the "whole" population of all similar products produced --- Call it a simple - brief- snapshot of a population/at that given time period... If you choose another set of samples the result may be quite different from day to day... or from sample set to sample set.
Their use is fairly easily exlained to the end user..
Regards
Jim
[This message has been edited by Jim Biz (edited 10 May 2000).]