Signed-up members of the at-risk register

What makes some people feel more at risk than others? asks Lennart Sjoberg

Striking differences between the attitudes of different European peoples to radiation risks are emerging in a five nation study. In a European Commission project, which I coordinate, we are investigating public perception of radiation risk 10 years after the Chernobyl accident.

We have found, for example, that the Spanish are more worried about radiation risks than are the Swedes. The greatest perceived risk in Sweden stems from east European nuclear power; yet, in our survey, we found that the Spaniards considered the risk to be just as high as did the Swedes - despite the fact that Spain is much farther away from eastern Europe than Sweden.

Another interesting finding is the large number of people who believe they may get cancer as a result of the Chernobyl accident - much larger than the expert estimates. The Swedes estimated that 300 people were actually killed in the accident - a number far higher than the official figure.

In open questions about worry, unemployment emerged as a key risk, even in Norway, where unemployment is quite low. People of course vary in how they perceive risk and which risks they worry about most. What are the reasons for such variations?

Women are more worried than men about most risks. This is most clearly the case when it comes to general risks; risks to others, or to "people in general". When personal risks are studied, risks as seen by someone as they pertain to him/herself, the gender gap in risk perception is reduced. Other factors are less important than gender, but people with lower incomes and less education tend to magnify most risks.

Experts differ from the public when it comes to risk perception. They tend to regard the risks posed by hazards such as nuclear waste or power plant operations as smaller than the public does. But in the case of "lifestyle" risks, like drinking alcohol or smoking, they perceive the risks as being higher. Lifestyle risks are complicated because they pose large differences between personal and general risks. Most people see the alcohol risk as small to themselves but as considerable to others. This is obviously important when considering strategies for communicating risks to the public.

The difference between experts and the public can be dramatic. A large sample of the Swedish public and experts working on nuclear waste issues were asked about the current approach to solving the nuclear waste problem. Very few among the public were satisfied with the solution, while most of the experts were. It is hard to explain why experts in some fields perceive risks as so small. We found, in the nuclear waste study, that they rated non-nuclear risks at the same level as the public, even a radiation hazard such as domestic radon. I believe that the phenomena of perceived competence and familiarity may provide some answers.

Two major attempts at explaining risk perception originated at the end of the 1970s. Research started earlier in the decade, when it was shown that risk acceptance seemed to depend not only on objective risk, but also on a factor named "voluntariness". Other dimensions, such as whether technology was new or not, were suggested, and later collected by the American psychologist Paul Slovic.

They form the basis of the psychometric model of risk perception largely used today. According to this model, risk perception can be described by two factors: new versus old risk, and "dreaded" risk. Nuclear power scores highly on both newness and dread, which was believed to explain the strong public opposition to this technology. However, more current work suggests that newness is hardly the most important factor, neither is "dread". Opposition to nuclear waste siting is mostly accounted for by the hazard being "unnatural".

The empirical basis of the psychometric model can be further criticised. If individual risk perception, rather than mean risk perception across individuals in a group, is studied, it is found that only about 20 per cent of the variance of perceived level of risk is explained. If tolerance of or demand for risk reduction are studied, the level of explained variance is even lower, perhaps about 15 per cent.

Anthropologist Mary Douglas has suggested a different approach, which became known as the "cultural theory" model. In a book written with Aaron Wildavsky, an American political scientist, she suggested that people can be categorised as various types - egalitarian, hierarchical, individualist or fatalist according to their values, world views and the group that they belong to. Scales were published whereby these typologies could be determined. The results seemed to suggest a strong relationship between such world views and people's perceptions of risk.

Researchers have tried to use these scales with only been partial success. There is a pattern in most data which gives some support to the model, but it is very weak; explaining about 5 per cent of the variance of perceived risk, perhaps up to 10 per cent in a few cases.

How, then, is perceived risk to be understood? Is there a better alternative? In my work, I have studied perceived nuclear waste risk. The model we have formulated is very different from those described above with three factors emerging as important determiners of perceived risk: attitude to nuclear power; specific fear of radiation and general risk sensitivity.

The model is similar for general and personal risk. Attitude is assumed to be antecedent to perceived risk, structural modelling supports this assumption. Specific fear is measured as fear of natural background radiation. General risk sensitivity is the mean- rated risk of several non-radiation risks. Some people are worried about virtually all risks, others about almost none.

Work is now underway to test the generality of our findings. The psychometric model and cultural theory have both failed - probably because they are attempts to explain risk perception on the basis of general factors. But risk is specific, and each new major hazard to which people are exposed, or believe they are exposed, brings a host of new factors and old ones in new combinations. That is why risk perception research is so fascinating and of lasting importance.

Lennart Sjoberg is at the Centre for Risk Research, Stockholm School of Economics.