Remarks on “Reply to the responses to the comments on “uncertainty profiles for the validation of analytical methods” by Saffaj and Ihssane”

Abstract

This letter provides an opportunity to expand further on the problems associated with using tolerance interval for share validated analytical methods and to estimate the measurement uncertainty. Indeed, we propose at clarifying points that have been raised by Rozet et al. (Talanta 100 (2012) 290–292) as some of them merit to be discussed.In that respect, we demonstrate here that β-content, γ-confidence tolerance intervals will provide perfect estimates of the routine uncertainty. In particular, we confirm that there is “no statistically significant difference” between the uncertainties estimated by our methodology with those obtained from the routine phase. Obviously, this is an opportunity to show which of the two types of tolerance interval i.e., β-expectation tolerance interval or β-content, γ-confidence tolerance interval allows a perfect estimate of the routine uncertainty.Furthermore, we prove that the β-expectation tolerance interval does not provide an adequate balance between consumer risk and producer risk.