Estimation of asymmetry in physics

Abstract

The notion of asymmetry arises in many physical experiments, being a natural expression of the relative difference between two quantities. In this paper, we discuss the estimation of asymmetries between the rates of two event processes subject to a common background. Such asymmetries have many applications in solid-state physics - a measurement of the difference in arrival rates of elctrons with one of two possible spin orientations to a detector, for example. We describe three techniques for estimating asymmetries from observations subject to uncertainty: a naive method based on an unbiassed estimate and a Gaussian approximation, a bootstrap approach, and a Bayesian approach. Differences in the interpretation of the estimates are discussed and their performances are compared by means of a simulation study. The aim is both to review the methodology on estimation of asymmetries and to contribute to the more general discussion of the relative merits of frequentist and Bayesian inference.