实验室测量审核结果的评定方法研讨

Discussion on the methods for evaluating the results of measurement audits

对多种类型测量审核活动的结果评价方法进行较为系统的研讨。关于测量审核重复性评定:采用单一试样时,可依据方法重复性限r评定;采用能力验证分割试样时,可依据稳健统计实验室内Z比分(ZW)评价。关于测量审核再现性评定:采用能力验证试样时,可依据稳健统计实验室间Z比分(Z或ZB)评价;采用具有指定值的试样时,可依据95%置信水平临界差评定,不同标准方法对于精密度的表述方式不同,其临界差的计算方法也不同,本文介绍了根据不同标准提供的参数计算临界差的方法;当实验室提供不确定度时,其结果也可依据指定值的不确定度En值评定。

The various methods for evaluating the results of measurement audits were studied systematically. As for repeatability evaluation, the repeatability limit r can be employed when a single sample was used in measurement audits; while in case of split samples of proficiency test （PT）, the robust Z-score within laboratory （ZW） can be used for evaluation. As for reproducibility evaluation, the robust Z-score between laboratories （ZB） can be adopted for PT samples; while critical difference at 95 % confidence level can be employed for samples with assigned values. Since the precision of different standard methods was presented in different forms, their calculation of critical difference also varied. This paper also introduces the methods for calculating critical difference according to parameters of different methods. If uncertainty was provided, the reproducibility can also be evaluated by the uncertainty En.