The standard recipes followed by HEP and astro-HEP scientists to decide whether the data bear enough evidence to support a discovery claim for a new phenomenon, or how a confidence interval can be set on a measured parameter, are an important ingredient in today's research in fundamental physics. Understanding where those standards come from, and what is the rationale behind them, is a necessary step in order to look beneath the surface of published scientific results. This talk will discuss in particular a few notable topics: the "five-sigma" criterion, its history, and its merits and pitfalls; the classical limit-setting procedures and their arbitrariness; and the issues of the ubiquitous simple-versus-composite hypothesis test, where the related paradox by Jeffrey and Lindley remains an active topic for debate in the statisticians community.
[Organizers: Andrea Giammanco and Marco Drewes, CP3]
