The analysis of cosmic ray data

There has been a number of new statistical results in recent years with application to the particular problems met in analysing cosmic ray particle and -ray data. Some specific statistics have been developed which deal with long-standing problems, such as the scan statistic for detecting bursts of counts and point sources in two dimensions. This review describes the available methods for the problems of detecting signals in time series and two dimensions and recommends the best statistic for each. New approaches to old problems have been developed based on Bayesian methods, which attempt to relieve the concerns often expressed about `degrees of freedom'. The background to these methods is given and some recent results on some of the same problems are given.