How to perform capability analysis on data that is the same

[#1]

Hi everyone. I'm trying to run a capability analysis on some sets of data (n=30) where the values do not vary. As far as I am aware, this is not possible. Do I need to treat those sets of data as attribute? Please advise. Thank you!

[#2]

First, welcome to the community.  I must first ask if you have evaluated your measurement system for whether or not it has a effective resolution or discrimination?  It appears your measurement system is rounding to the point where there is no quantifiable difference in the measurements.  This will, of course, lead to an estimate of standard deviation of 0.  Unfortunately you do not provide enough context to suggest alternatives.  Do you consider the measurement system appropriate and adequate? 

[#3]

I was thinking along the same lines. What is the smallest increment of change in this measurement? Would such a small change remain within the tolerance allowed for this outcome?

[#4]

We have run into multiple instances of this. The resolution of the measurement is based on the resolution of the specification for the measurement. So, in this particular case, the resolution would be three decimals. Also, we are using a gage pin to measure this diameter. The diameter had no variance, as the same diameter gage pin fit the feature every time. Unfortunately, we have not performed a gage r&r to assess the appropriateness of the measurement system, but it seems appropriate for this feature. Does this make sense? Sorry! I am pretty new to all of this. 

[#5]

I'm not sure the motivation?  Is it required you provide a capability study?

Gage pins are not a continuous measurement system (essentially go/no-go).  They lack discrimination and likely will not provide measures useful for estimating standard deviations (unless the samples you are measuring are extremely variable). They purposely only report MMC (essentially the smallest diameter).  So if the hole is out-of-round, the gage pins will not assess this. Now, there may be variation in the measurement system as a function of how the pin is used (e.g., angle of entry, force applied, etc.).  It might be useful to assess the appraiser variation, but the only thing you need for the pin itself is calibration records.  While this may be all you need, it is not useful for traditional capability studies.  Can you use a continuous measurement system like an indicator (TIR) or non-contact measurements?

As a note, using the specification for determining the correct measurement resolution only impacts the gage capability with respect to that spec.  It does not consider whether the measurement system is adequate for process improvement.