Quality in Bayesian chronological models in archaeology

Abstract

Bayesian chronological modelling is fast becoming the method of choice for the interpretation of radiocarbon dates in archaeological and palaeoenvironmental studies around the world. Although software enabling the routine application of the method has been available for more than twenty years, more than half of published models have appeared in the past five years. Unfortunately, the pace of development in statistical methodology has not been matched by the increased care in sample selection and reporting that is required for robust modelling. Barely half the applications considered in this article provide the information necessary to assess the models presented critically. This article discusses what information is required to allow the quality of Bayesian chronological models to be assessed, and provides check-lists for authors, editors and referees, in the hope of improving current practice.