ATI,
Look at the reading Pete recommended. Also, think of it this way:
Cpk used to be called “short-term” capability, now it is more appropriately called “potential” by Minitab. The reason is how std. dev. is calculated. The std. dev. for Cpk uses the R-bar/d2 method.
Ppk used to be called “long-term” capability, now it is more appropriately called “overall” by Minitab. Ppk uses the “standard” long for calculating the standard deviation, hence overall. It ignores between or within point to point variation (for the most part), and instead considers the overall standard deviation.
Try this demonstration. Run a capability analysis on the following sets of data. The first is a simple random set of data, and the second is the exact same data but sorted.
84.23680.61983.50777.15991.22368.94169.78094.82966.31165.18566.27182.47259.55266.65355.09384.10088.99274.08877.048101.31476.66673.67475.80689.21764.46886.64165.58882.49274.31978.62394.11075.44861.48257.15959.33080.07161.05059.04053.21878.73171.33878.72877.07184.17871.14777.88174.12776.685
and,
53.21855.09357.15959.04059.33059.55261.05061.48264.46865.18565.58866.27166.31166.65368.94169.78071.14771.33873.67474.08874.12774.31975.44875.80676.66676.68577.04877.07177.15977.88178.62378.72878.73180.07180.61982.47282.49283.50784.10084.17884.23686.64188.99289.21791.22394.11094.829101.314
You will notice a huge difference in the two estimates even though the data are the same.
Let me know if you have any more questions on this.
Good Luck,
Dr. Scott
.
