All scientific data and their interpretation are subject to some degree of uncertainty. Examples are method-ological variability and inter-individual, inter-species, and inter-strain differences. Such uncertainties in knowledge are compensated for by reduction factors. There is, however, insufficient information on all sources of uncertainty to provide a rigorous basis for establishing reduction factors over the whole frequency range and for all modulation patterns. Therefore, the degree to which caution is applied in the interpretation of the available database and in defining reduction factors is to a large extent a matter of expert judgment.