Hypothesis testing and confidence sets: why Bayesian not frequentist, and how to set a prior with a regulatory authority

Abstract

Acknowledgements: I am grateful to: Lalita Ramakrishan for eliciting the comments which provoked me to start writing; Pawe l Piwek for constructive criticism and inspiring sections 2.3 and 2.4; Filip Tokarski for inspiring section 9; Mark Owen for detailed comments and in particular for telling me to be less combative; Paul Edelstein for encouragement and for inspiring section 6; Mark Troll for encouragement and for reflecting the viewpoint of scientists less inclined to abstraction; and Brendan Kelly for suggesting possible further points to include. Nonetheless responsibility for the text and any errors etc. remains my own -reports of typos and errata welcome at the above email address.