Sorry for the delay. It has been a hectic week.
In my last post I stated the conclusions so far:
- There is no statistical difference between sides of the test chamber
- There is a statistical difference between operators, but it is very small compared to the difference between parts
After verifying that there was no effect due to side of the test chamber, I ran the data through the standard crossed R&R study.
- Total R&R = 8.2%.
- Repeatability = 7.47%
- Reproducibility = 3.4% Bob and John were virtually identical, but Kelly had significantly more variation in results, particularly on parts 4, 9 & 11.
I believe that your original concern back at the start of this thread was on whether you could perform an MSA on this measurement system in response to a customer request. The answer is yes, and fortunately for you the results are quite good.