A generalized statistical approach for interlaboratory comparisons with linear trends is proposed. This new approach can be applied to the general case when the artifacts are measured and reported multiple times in each participating laboratory. The advantages of this approach are that it is consistent with the previous approaches when only the pilot labboratory makes multiple measurements and applies to the cases that there either exists a trend or not. The uncertainties for the comparison reference value and the degree of equivalence are also provided.
Pub Type: Journals
Comparison reference value, degrees of equivalence, generalized least square estimator, linear regression, uncertainty.