Special issues in the analysis of nuclear physics data: SUPA Distinguished Visitor lecture

From the first laboratory course to the design and construction of an experiment, from the first simulations to the final analysis of the data from your experiment, a sophisticated treatment of uncertainties and the proper application of statistical methods are essential.

The aim of these lectures is to give a few insights in specific statistical methods with some concrete examples of how the methods are applied to data analysis of nuclear physics experiments. Starting from basic requirements and standard statistical distributions, these issues will be briefly discussed to demonstrate how they are related.

Techniques for fitting data will be reviewed. The treatment of systematic uncertainties, as well as methods to combine results from different experiments which may have common uncertainty sources will also be covered.

The test for background effects, even when no signal is observed, allows limits to be placed on the size of possible unidentified effects. These can provide severe constraints on theoretical models. Methods for calculating upper limits taking into account several uncertainty sources will also be considered.