This article reviews the tools available to assign realistic uncertainties in the field of radioactivity measurements following the prescriptions of the Guide to the Expression of Uncertainty in Measurement. A presentation of different distributions that can be encountered in the analysis of radioactivity measurements is listed, and explained using simple but pertinent examples. Two different methods, a simplified coincidence method and γ counting, are considered, for each of which a detailed uncertainty budget is given. This leads into the characteristics function and the law of propagation of errors that can be used in the evaluation of the total uncertainty. The treatment of correlations, which occur when using the same set of sources and the same decay period for both methods, is also detailed.

Subsequently, the analysis of the results of a comparison of activity measurements made by various national laboratories is presented with special care regarding the treatment of correlations that may arise.

Finally, the paper also deals with the KCRV determination and the evaluation of degrees of equivalence between laboratories in the general case where the results of BIPM, CCRI and RMO comparisons are linked together. In this case too, as far as is practical, a method to treat correlations is described.

While the illustrations in this paper are focused on radioactivity measurements, most of the material presented can, when appropriate, be used in other domains of metrology.