Compatibility verification of certified reference materials and user measurements

Abstract

A problem that frequently occurs in metrology is one of assessing compatibility of data obtained by a user laboratory with the specified values and uncertainty estimates from the certificate of analysis. The user's data are summarized by an estimated measurand value and a confidence interval, which is typically based on a repeatability standard deviation, but may include other variance or bias components. If the lab's interval and the certificate interval do not overlap, or more generally when the ‘no-bias’ hypothesis is rejected, the user may seek guidance on how to confirm this lack of compatibility or how to rectify it. The suggested two-stage statistical approach demonstrates a confidence interval whose width is similar to that of the certificate, and a compatibility test of guaranteed power for the given bias magnitude. Practical computationally simple formulae for each stage sample size are provided.

NIST Reference Materials: Utility and Future

“…NIST provides conceptual and computational tools to facilitate transfer of RMPs and material certification technologies to secondary standards producers and users who wish to develop their own higher-order materials. These include ways of evaluating chemical purity (21,22,(46)(47)(48), estimating measurement uncertainty (40,(49)(50)(51), evaluating bias (52,53), designing and analyzing calibration experiments (54), transferring metrological traceability with minimal increase in measurement uncertainty (55,56), establishing traceability of binary SRM mixtures (57), and demonstrating CRM comparability (58,59).…”

Bias detection and certified reference materials for random measurands

“…In another approach the lab may want to have its interval's width to be proportional to that of the certificate as discussed in [15]. Indeed the ratio of (expected) widths of these intervals, C m = width(CRM interval)/width(lab's interval), can be considered as one of the measurement capability indices used in quality control.…”

“…If compliance is rejected, i.e. when lab's coverage interval does not agree with the CRM specifications, a lab may want to pursue the bias correction by taking more measurements involving the same or different CRM [15].…”

“…The NIST Special Publication 260-181 [11] gives some recommendations for CRM selection, in particular in the analysis of foods and dietary supplements, see also [16]. The statistical requirements that a responding lab must meet before attempting a method validation, as well as some rules concerning the necessary lab's sample size, are discussed in [15]. The general role of uncertainty in making decisions about the significance and limitations of measurements (the bias checking being one of them) is reviewed in [13].…”

Random effects model for bias estimation: higher‐order asymptotic inference

“…The choice of an appropriate reference material is discussed in NIST (). Statistical requirements on a ' measuring system are reviewed by Rukhin ().…”