Bayesian data analysis techniques, together with suitable statistical models, can be used to obtain much more information from noisy data than the traditional frequentist methods. For instance, when searching for periodic signals in noisy data, the Bayesian techniques can be used to define exact detection criteria for low-amplitude signals - the most interesting signals that might correspond to habitable planets. We present an overview of Bayesian techniques and present detailed analyses of the HARPS-TERRA velocities of HD 40307, a nearby star observed to host a candidate habitable planet, to demonstrate in practice the applicability of Bayes' rule to astronomical data.

Submitted to the conference proceedings of the RoPACS meeting "Hot Planets and Cool Stars" (Nov. 2012, Garching), 14 pages