A novel nested design approach for estimation of uncertainty arising from sampling in feed

Uncertainty arising from sampling is nowadays a controversial issue, if it is necessary to be taken into account in total uncertainty. As in the quantification of analysis uncertainty, there are two principal approaches in order to estimate uncertainty of sampling: the bottom-up and the top-down approach. A novel top-down approach for sampling uncertainty is proposed in this study using nested analysis of variance (ANOVA). Determination of the oxytetracycline (OTC) content in feed was utilized for estimating uncertainty arising from sampling using the new approach. The advantage of nested ANOVA is that the effect of every factor on the total uncertainty can be calculated. In the present study, factors such as physical preparation of sample, lots of production and particle diameter were examined in the experimental design. It was found that the effect of different lots had the highest uncertainty among the factors. Finally, the results of nested ANOVA were compared with the uncertainty calculated using other top-down approaches. Classical ANOVA and robust ANOVA (RANOVA) did not show significant differences in their results. On the other hand, nested ANOVA estimated higher uncertainty of sampling in comparison with the other approaches.