Process Capability: Subgrouping vs. no subgrouping?

[#1]

Hi,

I am currently working on a Cpk Ppk analysis (JMP Pro 17.1) and I am very unclear on what subgrouping is and whether it is needed in my analysis. I have the same lab results dataset however did the Process Screening/Capability slightly differently:

1. Process Screening --> Process Variables = Grade A (leaving Grouping, Subgroup, Time and By column empty).

Result:

2. Process Screening --> Process Variables = Grade A, Time --> Year Quarter (leaving Grouping, Subgroup and By column empty).

Result: 

Method 1 control chart defaults to IMR control chart whereas method 2 default to Xbar control chart which I'm guessing is the reason for the difference in Cpk. However, I do not understand why that 2 different controls are used in this and which would be more accurate for my analysis since there's quite a large differences in the Cpk numbers. Would love to get any insight on this.

Would also appreciate if someone would help to explain what N subgroup actually mean in my analysis here (is it averaging my data set each quarter?). 

Thank you.

[#2]

Hi @ConditionalAvg2 ,

You are right about Method 1 & 2 (further reading: https://www.jmp.com/support/help/en/17.2/index.shtml#page/jmp/overview-of-the-process-capability-pla... )

I would prefer using Ppk (if you noticed, regardless of with/without subgroup the value is the same because it considers the whole dataset standard deviation).

[#3]

Whole Statistical Thinking for Industrial Problem Solving (STIPS) course  that is offered (for free) is great, but if I remember correctly (I did STIPS two or three years ago) Quality Methods parts did have good discussion regarding subgroups and selecting them. I think you can just take single module if you want to (though I think it is recommended to do them in order).

Statistical Thinking for Industrial Problem Solving - Quality Methods