Different % Gaugae R&R results in Variability Gauge report

[#1]

In a Variability Gauge report there are two different %Gauge R&R results cfr. below

Why is result 41,1% in Summary Gauge R&R statistics different from Variance Components result = 16,89% ? Which result must be used to report %Gauge R&R?

[#2]

Those are different, because they are different things. First one is calculated using Gauge R&R and Total Variation and other contains variance components. What are you supposed to report? There are many things which could be reported from Gage R&R report, it could for example be % of Tolerance (for Gauge R&R)

[#3 · Accepted solution]

Yes, I see your confusion. The top output is the ratio of standard deviations, which is pretty classic if you do any internet searches of the Gauge R&R literature (e.g., Gauge R&R < 10% is good). But the issue with standard deviations is they are not additive, whereas the variances (standard deviation ^ 2) is. So, it's the classic a^2 + b^2 = c^2, but a + b /= c.

The bottom report is looking at variance components, where the total variance would be 62.07887 + 305.42440 = 367.5033. You have 62.07887 / 367.5033 = 16.89%. The top is just going to be the square root of that, so 41.1%.

The JMP help menu has good documentation as well Gauge R&R Report.

[#4]

OK jszarka, thanks, the calculation by variance components is not what is generally used in quality so it is confusing. Would be better to give two different names to both calculations: Gauge R&R is computed using standard deviations but on the other hand indeed the calculation by variance components is statistically correct.