Comparisons between reproducibility standard deviations (SDR) derived from proficiency tests and from collaborative trials: mycotoxins in food

摘要

Both proficiency tests and collaborative trials provide information about the reproducibility standard deviation (SDR) of reported results from interlaboratory studies. Prima facie we would expect the former to provide greater SDR values on average. Collaborative trials require the participants to use a closely defined analytical procedure, while no such restriction applies in proficiency tests. Therefore we would expect the individual biases of different measurement principles, procedures and laboratories to contribute an extra uncertainty to the results of proficiency tests. In this paper we compare the SDR statistics derived from the determination of mycotoxins in foodstuffs. Overall there was little difference between the sources, with the implication that proficiency test results could, in this application sector, be used as a benchmark for uncertainty.