Sceptical combination of experimental results: General considerations and application to epsilon-prime/epsilon

Abstract

This paper is meant as a contribution to the often debated subject of how to combine data which appear to be in mutual disagreement. As a practical example, the epsilon-prime/epsilon determinations have been considered.

20 pages, 12 figures. This paper and related work are also available at http://www-zeus.roma1.infn.it/~agostini/prob+stat.html