Process Capability Screening by Group

[#1]

Hi, 

I have been using the process screening platform to calculate Cpk and Ppk values. Using the "by group" function I am able to use a separate table to define different LSL and USL values depending on the type of the process. 

One thing I run into is that by changing the type of the process data column from continuous to ordinal the Cpk calculated are different, particularly at low values of the "type of process data". I wonder why is that I don't think process type data goes into any equation that calculates Cpk and Ppk.

I am attaching an example with a screenshot showing two different process screening table and two jmp files, one being the spec limit and the other the data table with screening script saved.

Thanks a lot,

Gary

[#2 · Accepted solution]

When you used Condition 2 as the Grouping variable in Process Screening, you forgot to specify Condition as the Grouping variable in the Limits Specifications dialog.  You need to do this because Condition in your spec limits table does not have the same name as the Grouping column in your original data table (Condition 2).  You will get the results you desire if you specify Condition as the Grouping variable in the Limits Specifications dialog.

[#3]

Spot on Tonya. I overlooked this. But why was it able to still calculate with unmatching group name and some calculated values were correct and others not?

Thanks,

Gary

[#4]

JMP took the last row of the limits table as the limits for every condition (since you did not use a grouping column).  LSL=-1, Target=0, USL=1.  This is correct for every condition except for Condition=20 which has spec limits of -1, 0, and 2.  So for this example everything looks ok except for condition=20.