'Presuppositions, faith and statistics: an ecologist's view' The science of statistics has diverse origins, but a common theme is the desire to know the true value of a quantity that can only be measured with error, and to use this knowledge for technical purposes involving decision- making. The discipline of statistics therefore concerns the accuracy of beliefs and the justification of actions, and is now widely used in diverse areas of science, engineering and policy. The approach to statistical inference that currently predominates in many natural and social sciences is based on the school of Frequentism, which arose early in the 20th century in the context of evolutionary biology. In this paper I will argue, following Andrew Hartley, that Frequentism embodies a reductionistic motive whereby decision-making is reduced to questions of numerical inequality. The problems of this objectivist approach tend to elicit a humanistic reaction in the "indirect frequentist" approach, where expert judgement intervenes in arbitrary ways in processes of inference and decision-making. A more promising alternative has long been offered by the Bayesian school of statistical inference. Again following Hartley, I will argue that the Bayesian paradigm, in which inferences are explicitly based on prior beliefs, is a mathematically and experientially coherent approach to evaluating the accuracy of beliefs and the justification of decisions. Because it is not inherently reductionistic, I will also argue that it does better justice to the multi-faceted nature of reality as presented to us by its Creator. I will finish by sharing my first-hand experience with some challenges in the use of Bayesian-inspired statistical methods and surveying their current uptake.