The majority of the public in industrial countries believes that pollution and low doses of radiation are threats to good health. As a matter of fact, when these putative risks are compared to those originating from lifestyle, they appear very small. In particular, the risks associated with low doses of irradiation, even when they are assessed with the most pessimistic models, appear extremely small. Public anxiety is fuelled by the uncertainty regarding the magnitude of this risk and the use of the linear no threshold (LNT) hypothesis, which gives credence to the concept that even the smallest doses are harmful. There are a number of scientific and epidemiological data currently under debate that are not consistent with the LNT hypothesis. For example, no difference in the incidence of cancers or of birth defects has been observed between regions with low or high natural irradiation. This inconsistency between perceptions and data underlines the role of psychological factors studied since 1957 which should be placed in the perspective of the public's present attitude toward risk and technology. Social amplification or attenuation of risk may occur in several ways. Fearful concern about radiation began in 1955, with the beginning of the Cold War, when the possibility of a nuclear holocaust appeared very real. Analysis of the data shows that these fears of technology could have a detrimental effect; they should therefore be investigated and understood.