In a Six Sigma project, Gage R&R comes after the process has been mapped, and before the calculation of the baseline capability. It is again an essential step in the Control Phase, to ensure the correct measurement of the critical parameters.
Gage R&R ensures that the data used for study is good data. It increases the confidence on the data and helps with good decision making. In a Six Sigma project Attribute agreement aids in taking a trustworthy decision on continuous data.
A case based on the study on off-quality products used 4 possible choices from the ERP. To ensure whether everything that’s classified as shelf life is really the issue or not, the appraisers were made to judge each sample twice. The analysis process is made easier using Minitab. .
In the existing study, the graphs that came up illustrated that while appraiser 2 is in complete agreement with the standard, appraisers 1 and 3 are unsure about the set standards.
Next came the Fleiss Kappa statistical analysis, which compares the results to a possibility. With a range from +1 to -1, and 0 for a random change, Kappa says that the closer the value is to 1, it is more likely that the appraisers will be able to differentiate between the categories.
When the Kappa value is subtracted from 1, it gives the percentage value for Gage R&R. If Kappa is .9, the Gage percentage is 1 minus .9, which is .1 or 10%. Now, the AIAG rule says gage percentage less than 10 is acceptable, and greater than 30 is not, and anything in between needs improvement.
This Six Sigma project was thus successful in reducing 50% of the problem by using fact based solutions with the help of attribute agreement analysis.