The view of risk perception adopted in this paper focuses on how individuals define, and hence feel about, the outcomes of a risk issue. Perception of risk is seen as encompassing a variety of attributes of a risk issue including wider beliefs that permit a risk–benefit trade-off. The convergence between this differentiated view of risk perception and expectancy-value attitude theory was used to provide a technique for the description and comparison of risk perception across various risk issues, and between different, socially relevant, groups. A survey of beliefs and attitudes of the general public toward the use of various energy systems (coal, oil, hydro, solar and nuclear energy) was carried out in Austria at a time of increasing concern with energy strategies and the controversy surrounding Austria’s first nuclear energy plant. The study showed that the public does conceive risk issues in differentiated terms, taking into account several substantive dimensions of both risk and probable benefits. While such dimensions might well be specific to the risk issue in question, it does seem likely that both risks and probable benefits will form part of belief systems in most instances where risk acceptance, or otherwise, is an issue. In general, energy systems were found to be associated with only environmental risk; however, nuclear energy was an exception in that the public, whether in favour of nuclear energy (pro), or against it (con), believed that it is also associated with psychological and physical risk. Those against nuclear energy also believed that it is associated with indirect (future-oriented and political) risk, while those in favour did not associate nuclear energy with environmental risk. Overall, those members of the public with pro nuclear attitudes believed more strongly than those with con nuclear attitudes in the economic and technological benefits of energy generation in general. The question was raised as to how an understanding of risk perception in the sense described here might be used in the service of policy making. It was suggested that, on issues that are the subject of public debate, policy makers might already possess sufficient information to understand the reasoning of public groups. This hypothesis was tested by asking a group of policy makers to take part in a role-play experiment. Their responses in the role of typical Austrian citizens either pro or con nuclear energy were compared with the actual responses of the public. In general, the policy makers were accurate, but there were some errors of attribution. In particular, the policy makers underestimated the public’s negative feelings about the psychological risk dimension and their belief in its association with nuclear energy. This mistaken attribution was especially large for those members of the public who were in favour of nuclear energy.