Abstract

Measurement error modeling is crucial to any assay method. Realistic error models prioritize efforts to reduce key error components and provide a way to estimate total ("random" and "systematic") measurement error variances. This paper uses multi-laboratory data to estimate random error and systematic error variances for seven analytical chemistry destructive assay methods for five analytes (Gallium, Iron, Silicon, Plutonium, and Uranium). Because these variance estimates are based on multiple-component error models, strategies are described for choosing and then fitting error models that allow for lab-to-lab variation.

This work is licensed under a Creative Commons Attribution 4.0 License.