The criteria for acceptability depends on the type of study.

According to AIAG^{1}
guidelines, if your measurement system's variation is less than 10% of process's
variation, then it is acceptable. To evaluate your process variation, compare the
Total Gage R&R contribution in the %StudyVar column (%Tolerance, %Process) in
your output with the values in the table.

Percentage of process variation | Acceptability |
---|---|

Less than 10% | The measurement system is acceptable. |

Between 10% and 30% | The measurement system is acceptable depending on the application, the cost of the measurement device, cost of repair, or other factors. |

Greater than 30% | The measurement system is not acceptable and should be improved. |

This table contains corresponding guidelines using variance components. To
evaluate your variance components, compare the %Contribution column in your
output with the values in the table.

###### Important

Percentage of variance components | Acceptability |
---|---|

Less than 1% | The measurement system is acceptable. |

Between 1% and 9% | The measurement system is acceptable depending on the application, the cost of the measurement device, cost of repair, or other factors. |

Greater than 9% | The measurement system is not acceptable and should be improved. |

The AIAG also states that the number of distinct categories into which the measurement system divides process output should be greater than or equal to 5.

Guidelines for Wheeler's EMP study classify the measurement system into a class with
the intraclass correlation coefficient. Wheeler (2006) ^{2} describes the calculations,
output, and classifications for the EMP crossed study. In practical terms, the
coefficient explains how well the measurement system detects a shift in the process
mean of at least 3 standard deviations. First and second class measurement systems
usually have a high probability to detect such shifts with a limited number of tests
and subgroups on a control chart. For third class measurement systems, the typical
analysis adds tests to the control chart to increase the probability to detect a
shift in the process mean. A fourth class measurement system usually requires
improvement to monitor a process or for process improvement activities.

Classification | Intraclass Correlation | Attenuation of Process Signals | Probability of Warning, Test 1* | Probability of Warning, Tests* |
---|---|---|---|---|

First Class | 0.80 - 1.00 | Less than 11% | 0.99 - 1.00 | 1.00 |

Second Class | 0.50 - 0.80 | 11 - 29% | 0.88 - 0.99 | 1.00 |

Third Class | 0.20 - 0.50 | 29 - 55% | 0.40 - 0.88 | 0.92 - 1.00 |

Fourth Class | 0.00 - 0.20 | More than 55% | 0.03 - 0.40 | 0.08 - 0.92 |

The two criteria lead to different conclusions. The classifications for Wheeler's EMP studies are less strict than the classifications for gage R&R studies that follow the AIAG methodology.

Intraclass correlation | Percent of process variation | AIAG | EMP |
---|---|---|---|

99% | 10% | Acceptable | First class |

91% | 30% | Marginal | First class |

80% | 45% | Needs improvement | First class |

50% | 71% | Needs improvement | Second class |

20% | 89% | Needs improvement | Third class |

0% | 100% | Needs improvement | Fourth class |

The development of the AIAG criteria in the automotive industry is from a tradition of processes that require high precision from the measurements to meet tight tolerances. The development of the EMP criteria comes from a tradition that uses the measurement system to detect shifts in the process average for process improvement activities.