We deal with the analysis of on-off measurements designed for the confirmation of a weak source of events whose presence is hypothesized, based on former observations. The problem of a small number of source events that are masked by an imprecisely known background is addressed from a Bayesian point of view. We examine three closely related variables, the posterior distributions of which carry relevant information about various aspects of the investigated phenomena. This information is utilized for predictions of further observations, given actual data. Backed by details of detection, we propose how to quantify disparities between different measurements. The usefulness of the Bayesian inference is demonstrated on examples taken from cosmic ray physics.

We propose an analytical solution to the on-off problem within the framework of Bayesian statistics. Both the statistical significance for the discovery of new phenomena and credible intervals on model parameters are presented in a consistent way. We use a large enough family of prior distributions of relevant parameters. The proposed analysis is designed to provide Bayesian solutions that can be used for any number of observed on-off events, including zero. The procedure is checked using Monte Carlo simulations. The usefulness of the method is demonstrated on examples from gamma-ray astronomy.

We focus on the primary composition of cosmic rays with the highest energies that cause extensive air showers in the Earth's atmosphere. A way of examining the two lowest order moments of the sample distribution of the depth of shower maximum is presented. The aim is to show that useful information about the composition of the primary beam can be inferred with limited knowledge we have about processes underlying these observations. In order to describe how the moments of the depth of shower maximum depend on the type of primary particles and their energies, we utilize a superposition model. Using the principle of maximum entropy, we are able to determine what trends in the primary composition are consistent with the input data, while relying on a limited amount of information from shower physics. Some capabilities and limitations of the proposed method are discussed. In order to achieve a realistic description of the primary mass composition, we pay special attention to the choice of the parameters of the superposition model. We present two examples that demonstrate what consequences can be drawn for energy dependent changes in the primary composition.

We describe a straightforward modification of frequently invoked methods for the determination of the statistical significance of a gamma-ray signal observed in a counting process. A simple criterion is proposed to decide whether a set of measurements of the numbers of photons registered in the source and background regions is consistent with the assumption of a constant source activity. This method is particularly suitable for immediate evaluation of the stability of the observed gamma-ray signal. It is independent of the exposure estimates, reducing thus the impact of systematic inaccuracies, and properly accounts for the fluctuations in the number of detected photons. The usefulness of the method is demonstrated on several examples. We discuss intensity changes for gamma-ray emitters detected at very high energies by the current gamma-ray telescopes (e.g. 1ES 0229+200, 1ES 1959+650 and PG 1553+113). Some of the measurements are quantified to be exceptional with large statistical significances.