Questions about Capability Statistics - Part 1

Minitab Blog Editor | 19 September, 2011

Topics: Capability Analysis, Quality Improvement

When you're conducting a Capability Analysis, do you know which statistic to look at?  We get a fair number of questions about this, so let's explore the question. 

In the graph below, you’ll see that there are two categories for Capability Statistics, Potential and Overall:


The Potential (Within) Capability statistics(Cp, CPL…) are based off of the estimate for within standard deviation. Well, what is within standard deviation? This simply represents the variation within your subgroups.  On the flip side, overall standard deviation takes into account variation from the entire data set and looks at the total process variation. The Overall Capability statistics (Pp, PPL…) are thus based off of this information. 

The Cpk and Ppk are the default “go-to” statistics when running this analysis. They both represent the minimum of the respective lower and upper values (CPL  and CPU, for example).  You’d want to compare your values to benchmarks to determine whether to improve your process; many industries use 1.33 as the benchmark value for both Cpk and Ppk.

There have been situations where customers have called in saying that their Cpk value didn’t match ours. This is due to a difference in the definition of Cpk in some industries. We have spoken to a lot of customers who have been calling our “Ppk” the “Cpk”.   Is this bad? Is there a reason that we should track Within Capability?

Minitab's product support page says this:

The within variation corresponds to the inherent process variation defined in the Statistical Process Control (SPC) Reference Manual (Chrysler Corporation, Ford Motor Company, and General Motors Corporation. Copyright by A.I.A.G) while overall variation corresponds to the total process variation. Inherent process variation is due to common causes only. Overall variation is due to both common and special causes. Cp and Cpk are called potential capability in Minitab, because they reflect the potential that could be attained if all special causes were eliminated.

There are situations where you may not have a legitimate subgroup size greater than 1 in your data set. Fortunately, Minitab still tries to provide an estimate for the within standard deviation by using Moving Range calculations.