topic Re: Question on JMP Means & R chart please in Discussions

Question on JMP Means & R chart please

Kate — Sun, 11 Jun 2023 11:25:22 GMT

See question in the above chart. How does the number of segments defined please? Thanks!

Re: Question on JMP Means & R chart please

Mark_Bailey — Tue, 28 Jun 2022 13:47:25 GMT

The X bar and R charts work with subgroups of size > 1. You have multiple measurements per subgroup level. The limits change depending on the subgroup size.

Can you show your data or describe it and how it is collected? Can you describe how you launched the chart platform (i.e., the variables used in the analysis roles)?

Re: Question on JMP Means & R chart please

statman — Tue, 28 Jun 2022 17:14:56 GMT

As Mark points out, this is due to varying subgroup sizes. I'm not sure how you intend to use control charts, and what questions you are trying to answer with the analysis, but varying subgroup sizes is not how Shewhart intended the control chart method be applied.

Re: Question on JMP Means & R chart please

Kate — Tue, 28 Jun 2022 20:00:12 GMT

Hi, Mark and Statman,

Many thanks! I also wonder how it is so, because I didn´t made any subgroup, and I don´t want. I just would like to understand how the 3 and 4 segments are generated automatically. since I need to explain the to people on the resulted plot. In my eyes, the number of the segments should be equal to the number of the x points. correct? or the number of the segments can be set with JMP? no idea.

Re: Question on JMP Means & R chart please

statman — Tue, 28 Jun 2022 23:28:16 GMT

Kate,

What version JMP are you running? What platform are you running? If you could share your data set, perhaps we could provide some guidance. You are running some sort of control chart platform, but there are multiple ways to create these charts. If you choose Analyze>Quality and Process>Legacy Control Charts>Xbar, you will get the following screen:

In the Sample Size section, you have 2 options either Sample Grouped by Sample Label (by one of your columns) or select Sample Size Constant and enter the number of observations in a subgroup. If you chose the first option, then subgroups may have unequal values within subgroup and therefore you may get varying control limits. Subgroups should be based on rational hypotheses.

Re: Question on JMP Means & R chart please

Mark_Bailey — Wed, 29 Jun 2022 13:55:37 GMT

Are you saying that the subgroup size is constant and equal to 1 for all the observations? Are you saying that you want an Individuals and Moving Range chart pair (XmR)?

Re: Question on JMP Means & R chart please

Kate — Wed, 29 Jun 2022 18:50:48 GMT

Hi, Mark and Statman,

Good to know there is such options. I used the control chart builder. Here I attach the data. Hopefully it would make it easier to have the problem explained. Many thanks!!

Re: Question on JMP Means & R chart please

tonya_mauldin — Thu, 30 Jun 2022 12:21:12 GMT

In the XBar & R chart of Error script attached to your data table, you will see that Target has been defined as the subgroup variable (Subgroup( :Target )). For Target =100 and Name=A, there are 9 observations in the subgroup. For Target=150 and Name=A, there are 18 observations in the subgroup. The control limits vary because of the varying subgroup sizes.