16. MSA – Attributive
For attributive MSA, it helps to analyze attributive results (PASS/FAIL) for evaluating whether the system is judging the attributive characteristics correctly. the judgement correctness, κ (Greek: Kappa) will be used to calculate the judgement correctness level.
For the example demonstrated below, 2 different CCD cameras are being evaluated.
There are 50 reference samples for the CCD camera to evaluate (10 pass sample and 40 fail samples), the judgement correctness and % matching value between the lanes are the main focus for the attributive MSA.
Where the Kappa value indicate the measurement system can perfectly judge the PASS/FAIL attribute category.
Equation above is only applicable when you have only pass/fail and two operators, refer to equation on the right for multiple responses and operators (minimum 3 for both).
Based on the analysis from MINITAB, both CCD camera can fully identify the PASS/FAIL either by themselves against the standard results or compared with each other. The Kappa value is 1 which is 100% agreement between the CCD cameras compared with the standard samples.
Minitab Analysis Summary Data
Appraiser vs Standard Accuracy Diagram