Some Thoughts on Random Errors, Systematic Errors, and Biases

An error of measurement may be defined as the "magnitude and the sign of the difference between the measured value and the 'true' value" [1]. The subject of measurement errors is of great importance in the area of special nuclear materials (SNM) accountability. Key quantities that measure the level of SNM accountability performance, material unaccounted for (MUF) and shipper-receiver difference (S/RD), are influenced by measurement errors, and much effort is expended in evaluating the sizes of reported MUF's and S/RD's relative to the combined effect of errors of measurement. In these applications, individual measurement errors may be quite large, and their effects cannot be ignored. Further, there are many sources of error that contribute to an overall index of performance such as MUF, and the problem of how to combine their effects is a very important one.