It is stated in the article that “The company was also using an incorrect method of calculating the process capability. In the calculation of Cp and Cpk, the most important consideration is the use of standard deviation that is estimated as the population standard deviation – free of special causes estimated from the subgroups formed. The sample standard deviation calculated by the company included variation due to special causes, which is more appropriate for the calculation of Pp or Ppk.”.
According to my QS handbooks it is the sample sigmas that defines the Cp and Cpk indicies and the over all (population) sigma that defines the Pp and Ppk indicies and not the other way around.

]]>