Report from SGR Conference 'Sharing the Responsibility' (1997)

A Richter Scale for Risk?

The Scientific Management of Uncertainty versus the Management of Scientific Uncertainty

Report written by Alan Mayne, SGR
 

John Adams began his presentation by expressing the hope that he could make some connections with the previous session, on the SGR Climate Train and the forthcoming Kyoto Conference on climate change. His presentation was supplemented by a paper, included in the Conference Pack, which gives more extensive details, and contains literature references including his own book "Risk" (1995, UCL Press, London). Risk analysis and assessment are about how to behave scientifically and responsibly in the face of uncertainty. The Royal Society studied risk assessment in 1982, and held a conference on "Science, Policy and Risk" on 17 March 1997 [reported by Tim Foxon in SGR Newsletter (May 1997), No. 15, pp. 4-5. Ed.]. Whereas the 1982 report was confident, the 1997 conference was presented as a contribution to a wide-ranging debate about risk in the scientific community. John Adams himself became involved in the risk debate when Professor Lewis Wolpert of University College London chaired a symposium on BSE.
 

Some Basic Concepts and Models of Risk

Risk management involves balancing risks and rewards, and Professor Adams outlined a simple model of this process. Everyone has a propensity to take risks, which varies from one individual to another, and is influenced by the potential rewards of risk taking. An individual's perception of risk is influenced by his own and other people's experience of accident losses, which in turn result from risk. The more risks an individual takes, the greater his rewards and losses are on average. Individual risk-taking decisions represent a balancing act, where perceptions of risk are weighed against the propensity to take risks; this phenomenon is called 'the risk thermostat'. When considering how the balancing act is performed, it is helpful to distinguish between three classes of risk: (1) directly perceptible risks, occurring in everyday life; (2) risks perceptible with the aid of science, such as infectious diseases; and (3) virtual risks, such as BSE, which scientists do not know or cannot agree about.
 

Directly Perceptible Risks

A toddler learning to walk is a conscious risk manager, who does a physical balancing act. The risk thermostat is a conceptual, qualitative model of how our behaviour in risky situations is influenced by our perceptions and attitudes. The following examples show that it explains why certain 'safety precautions' against road accidents sometimes have results contrary to those expected by many 'experts'. Safer brakes were found not to reduce car accidents. The introduction of seat belts in cars, even by legislation making their use compulsory, often has little effect on the rate of fatalities and serious injuries in car crashes. Indices of death by accident and violence in 31 countries show that everyone is his own risk manager, who 'gets round' safety measures by taking more risks. Each person tends to have his own 'acceptable' level of risk. Professor Adams accepted that the reduction in road accidents was correlated with the introduction of anti drink-driving legislation and related action by police forces around the world. He also gave a fascinating insight into the meaning of disaggregating risk assessment. For example, it is on average much more dangerous to drive than to fly by plane for most journeys, but the difference between these risks is dramatically less for a middle-aged, sober male air passenger, taking a short-haul flight, who has a relatively low rate of driving accidents. The message here is that one must beware of simple averages!
 

Risks Perceptible with the Aid of Science

The literature on risk and safety is overwhelmingly dominated by this class of risk. The 'rational actor paradigm' is central to this literature; here, the advice of risk experts about how to manage risk is based on their judgement about how a rational optimiser would and should act, if possessing all relevant scientific information. They often express their dismay at the inability of ordinary people to use such information sensibly, and seek ways of making their risk-taking decisions more rational and better informed. However, they tend to ignore the action of the risk thermostat. For example, a box diagram of a pharmaceutical company's risk assessment process tends to be an elaborate version of the bottom half of the diagram outlining the operation of the risk thermostat. Attempts to quantify various types of risks led to scientists' and decision makers' demands for a 'Richter Scale for Risk', which would attach quantitative levels to the orders of magnitude the probability of occurrence of well-known types of risk.
 

Virtual Risks

'Virtual risks' are meanings which we impose on uncertainty, when we are faced with it. The public is confronted by uncertainty whenever scientists disagree about certain specific risks, or confess that they do not know. Whether or not virtual risks are imaginary, they have real consequences, because people act on the meanings that they impose on uncertainty. All of us, including scientists, perceive virtual risks through 'cultural filters'; the cultural filters of scientists are usually called 'paradigms'. Several examples of virtual risks were discussed. The assessment of risks from the vast majority of potential carcinogens, that have not yet been tested, is virtual. Very different dose-response relationships can be derived from the same experimental data, for low levels of exposure to toxins, although the relationship becomes predictable for high levels. Global warming and climate change are typical examples of virtual risks.

A typology of cultural filters has been developed to help account for the different meanings that they impose on uncertainty. This typology reveals four basic types of attitude to risk that people may have, sometimes as dominant approaches, and sometimes in combination. Egalitarians view environmental threats as punishment for technocratic arrogance and failure to obey nature; they urge a retreat to sustainable practices. Individualists consider nature to be robust, and argue that our best protection is the power of science and technology over nature. Hierarchists, typically in government, assure everyone that things are under their control, and commission more research that they hope will prove it. Fatalists are resigned about their lack of power to guide events, and enjoy life 'while the going is good'. Typical quotations of expressions of these attitudes were given. To conclude, science has very effectively reduced some uncertainties, but is much less effective in managing uncertainty. The scientific risk literature says little about virtual risks. There are still some huge uncertainties in today's human situation, about which science cannot yet provide any definite answers.