Uncertainty analysis for measurement of measurand

Abstract

For measuring the true value of a measurand, vague statistical concept of measurand results in inefficient uncertainty analysis of measurement. The vagueness is caused by the fact that true value of the measurand is an unknown parameter such as population mean or median and the measurement of this true value is a random variable. Generally a parameter may be estimated remarkably more efficiently than the prediction of the random variable. The classical uncertainty analysis in the literature has been developed based on the structure that a measurand is a random variable. This misspecification of statistical concept costs serious price of sacrificing efficiency in terms of length of the uncertainty interval. The purpose of present study is to formally formulate a statistical model for the true value of measurand and provide an uncertainty analysis for the measurement of this true value. (C) 2010 Elsevier Ltd. All rights reserved.