An approach to the analysis of key comparisons data is proposed using a model taken from meta-analysis metrology. Under the assumption of Gaussian distributions and equivalent qualification of all participating laboratories the analysis of data from interlaboratory studies is explored. The model leads to a class of weighted means estimators for the consensus value and to a method of assessing the uncertainty of the resulting estimates. Derivation of Bayes estimators shows that these also belong to the meta-analysis class.
Pub Type: Journals
DerSimonian-Laird estimator, interlaboratory study, key comparison reference value, Mandel-Paule algorithm, median, meta-analysis, uncertainty of Type A, uncertainty of Type B, weighted means