[en] Saffaj et al., recently proposed an uncertainty profile for evaluating the validity of analytical methods using the statistical methodology of γ-confidence β-content tolerance intervals. This profile assesses the validity of the method by comparing the method measurement uncertainty to a pre defined acceptance limit stating the maximum uncertainty suitable for the method under study.
In this letter we comment on the response (T. Saffaj, B. Ihssane, Talanta 94 (2012) 361-362) these authors have made to our previous letter (E. Rozet, E. Ziemons, R.D. Marini, B. Boulanger, Ph. Hubert, Talanta 88 (2012) 769–771). In particular, we demonstrate that β-expectation tolerance intervals are prediction intervals, we show that β-expectation tolerance intervals are highly usefull for assessing analytical methods validation and for estimating measurement uncertainty and finally we show what are the differences and implications for these two topics (validation and uncertainty) when using either the methodology of β-expectation tolerance intervals or the γ-confidence β-content tolerance tolerance interval one.