The Safeguards Advantages of Naive Bias Correction

In a measurement control situation where periodic measurements of known quantities (standards) are interspersed among production measurements of unknowns, handling biases is to base the correction to an unknown soley on the observed error of the closest (in time) standard measurment. The statistical properties of this "naive" bias correction method are derived, and its utility for safeguards objectives is discussed. It is pointed out that no statistical evaluation of the data is needed to make such a correction, that the statistical properties of the corrected estimates are independent of the nature of systematic errors or biases, and that in most cases the correction results in considerably improved estimates of the total amounts of material in strata containing many items. By contrast, this correction method should generally not be applied if one is interested only in the contents of individual items. Because facility operational and administrative needs do not necessarily coincide with material accountancy goals, it is argued that it is reasonable that measurement and measurement control data should be treated differently for the different purposes. Thus, the bias correction should not be applied to book values, but accumulated by strata and used for the correction of inventory differences. a. simple method of