Comparison of artefact standards by more than two laboratories

Laboratory comparisons of artefact standards may follow a "star-shaped" sequence ABACA... or a "round-robin" sequence ABC... A, where A is the pilot laboratory. For each of these sequences, two analysis procedures are described: "overall constant drift" (OCD) and "separately fitted lines" (SFL). SFL is well suited to cases where there are changes in artefact drift, or transportation shifts. With the round-robin sequence, both OCD and SFL in particular assign different uncertainties to laboratories' realizations of a physical unit, depending on the laboratories' positions in the sequence, and also induce significant interlaboratory correlations. These disadvantages are less pronounced with the star sequence, where SFL in particular assigns equal uncertainties regardless of position in the sequence.