The Bayesian approach to reflectivity data

We outline the basic principles of Bayesian probability theory and illustrate its use with reflectivity data. This approach provides a unified rationale for data analysis, which both justifies many of the commonly used analysis procedures and indicates some natural extensions that enhance their potency. Thus, for example, we find that the ubiquitous least-squares apparatus of parameter estimation is easily adapted to tackle the more intriguing question of model selection. A free-form solution application is also presented, as is a discussion of the difficult but important question of optimal experimental design.