A Perspective on the Construction of Nonsymmetric Uncertainty Intervals

Abstract

This paper presents a new method for constructing nonsymmetric uncertainty intervals, one that is based on estimates of “expected values” and “variances” associated with deterministic errors that one constructs from estimates of “upper bias limits” and “lower bias limits” for measured variables. On the assumption that upper bias limits and lower bias limits specified by the user correspond to 95 percent confidence intervals for normally distributed deterministic errors, the uncertainty intervals determined by the new method reduce to approximate 95 percent confidence intervals for the true value of the measured variables.