Distance Estimation in Cosmology

Abstract

In this paper we outline the framework of mathematical statistics with which one may study the properties of galaxy distance estimators. We describe, within this framework, how one may formulate the problem of distance estimation as a Bayesian inference problem, and highlight the crucial question of how one incorporates prior information in this approach. We contrast the Bayesian approach with the classical `frequentist' treatment of parameter estimation, and illustrate -- with the simple example of estimating the distance to a single galaxy in a redshift survey -- how one can obtain a significantly different result in the two cases. We also examine some examples of a Bayesian treatment of distance estimation -- involving the definition of Malmquist corrections -- which have been applied in recent literature, and discuss the validity of the assumptions on which such treatments have been based.