Werner Hässelbarth

Controlling uncertainty in calibration

Abstract A rigorous error-in-both variables procedure for linear and nonlinear curve fitting is described, including determination of the variance/covariance matrix of the fitted parameters. The curve parameters are determined by regression according to the maximum-likelihood functional relationship approach. The variance/covariance matrix is calculated by uncertainty propagation from the variances and covariances of the calibration data.

Rückführbarkeit analytischer Messungen

Die Ruckfuhrung von Mesergebnissen — das Knupfen von Kalibrierketten — bis hin zu anerkannten Normalen ist kein Selbstzweck. Die Forderung nach Ruckfuhrbarkeit dient vielmehr einem hoheren Zweck: der Einfuhrung einer Beweispflicht fur die Genauigkeit von Ergebnissen, hier neben der Prazision vor allem der Richtigkeit. Das gilt gleichermasen fur die physikalische Mestechnik wie fur die (quantitative) chemische Analytik.

Approaches to Data Assessment and Uncertainty Estimation in Testing

The following chapter presents a selection of feasible approaches to data assessment and uncertainty evaluation for single measurands (univariate case), multiple measurands with and without a functional relationship (multivariate case), and some aspects of semi-qualitative and qualitative testing.

Uncertainty — The key topic of metrology in chemistry

Abstract At the Second EURACHEM Workshop on Measurement Uncertainty in Chemical Analysis the author had the pleasure of chairing a working group on chemical metrology. This note presents some propositions arising from the preparation of, as well as from the discussion at and after, the working group session.