Type 1 Gage R&R - Analyzing Run Chart and I-mR

[Mr.Agi]

Hi All, I'm new in this forum (first post actually), even if I read the forum I don't have an answer for this question.

I'm analyzing an appraiser to measure the airtight capability
I measured 20 times the same pumps.
Before starting the test I analyzed the values, not normal distributed.

The Run Chart tells me that there is an oscillation, (p-value 0,048) so I can suppose the process is instable.

The I-mR chart tells me that the process is under control, no problems.

How is it possible? I mean, is it possible a process to be unstable but under statistical control?

Thanks all for your reply

[bobdoering]

In Reply to Parent Post by Mr.Agi

I'm analyzing an appraiser to measure the airtight capability
I measured 20 times the same pumps.
Before starting the test I analyzed the values, not normal distributed.

The Run Chart tells me that there is an oscillation, (p-value 0,048) so I can suppose the process is instable.

The I-mR chart tells me that the process is under control, no problems.

How is it possible? I mean, is it possible a process to be unstable but under statistical control?

For a process to be stable in traditional SPC using Western Electric rules, the data will randomly jump about the mean. So, the calculations may analyze the data and may not detect an actual oscillation, and calculate control limits no problem. In essence, it will appear to be in statistical control.

The issue is that the data does not tell the story, it illustrates the story. Or, in other words, it supports the theory not generates the theory. So, the question is: what is the theory? Should there be oscillation? Can it be explained? Is it truly oscillation, or some kind of measurement error? Bottom line, what are the variations of the process?

[Bev D]

can you post your data?

[Mr.Agi]

Thank you for your reply.
Here attached an excel file with data

[bobdoering]

In Reply to Parent Post by Mr.Agi

Thank you for your reply.
Here attached an excel file with data

Here are some charts to look at.

[Mr.Agi]

I use software MINITAB to perform my analysis. The software gives me also a P-Value, so that I can accept or refuse an hypotesis. In the Oscillation case the P-Value is 0,048, so very close to 0,05. Maybe a lack of experience is the problem, the data actually seem random distributed. Thanks

[bobdoering]

In Reply to Parent Post by Mr.Agi

I use software MINITAB to perform my analysis. The software gives me also a P-Value, so that I can accept or refuse an hypotesis. In the Oscillation case the P-Value is 0,048, so very close to 0,05. Maybe a lack of experience is the problem, the data actually seem random distributed. Thanks

Be careful with the 0.05 conclusion.

[Bev D]

what is mbar?