Key comparisons: the chance for discrepant results and some consequences

Franco Pavese


The paper reports an analysis of the results of complete sets of key comparisons available on the BIPM repository (KCDB), which shows that the majority of them show some discrepant results. The main reason of this fact is due to unknown or ill-estimated systematic effects, and indicates that the usual method of 'correcting' for known systematic effects is inefficient. The paper then discusses some of the possible reasons and an alternative approach, starting from a proposed revision of the meaning of systematic effect/error.

Full Text: