Exclusive statistics: simple treatment of the unavoidable correlations from key comparison reference values

Abstract

Measurement capabilities of national metrology institutes (NMIs) are being analysed with respect to a reference value (the key comparison reference value, KCRV) based on measurement data obtained during the key comparison. Thus the KCRV is correlated with the data of the contributing laboratories. As rigorous treatments of correlations are not routinely presented, a simplified approach to quantifying these inevitable correlations is appropriate. We discuss one method of analysis that neatly quantifies these effects and is particularly easy to describe to all end users, even those not adept with correlation analysis. Each laboratory's result is compared with the KCRV, in whichever way it is calculated (mean, weighted mean, median, etc.), recomputed including only the other laboratories but excluding itself. Exclusive statistics provide a simple procedure for examining candidate KCRV methods to ensure that the correlations introduced by the KCRV have been properly considered.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 24 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.