Exclusive: shock headlines induce horror

How much do the media affect public perception of hazard? Quite a lot, according to a recent American survey. John Davies reports

If we know more about health hazards than we used to it is thanks to television, newspapers and magazines. But the mass media seem to have contributed to some common misperceptions, too.

This is one conclusion to be drawn from research carried out for John D. Graham of the Harvard School of Public Health's Center for Risk Analysis. Or as Professor Graham puts it: "Clearly the mass media are a compelling and persuasive force I People appear to have a high degree of confidence and trust in television and the print media." People in the United States, he continues, "tend to rely on [the media] for information" about hazards to health rather than, for example, on the advice of their own doctors.

Graham's analysis is based on a random sample of adult Americans surveyed last November. Questions focused on eight (possible) hazards: radiation from medical X-rays; pesticide residues on fruit and vegetables; depletion of the earth's ozone layer by chemicals; global climatic warming from carbon dioxide pollution; electric and magnetic fields (EMFs) emitted from power lines; natural radon gas in homes and buildings; environmental tobacco smoke (ETS) exhaled into the air by smokers; and dust and particles in city air.

The 1,000-plus respondents from all over the US, interviewed over the telephone, were asked to say how confident they were that any given item was hazardous, using a scale from zero to ten; ten would mean complete confidence that a hazard existed, and zero, complete confidence that there was no risk. The interviewees could not reply "don't know" or "not sure"; instead a response of five or near five would be recorded if they had no firm opinion on the subject.

First, however, all respondents had to react to two "test" hazards, to ensure they understood the meaning of the zero to ten scale. These were "heavy smoking" and "listening to relaxing music". Obviously the first was expected to receive high scores of nine or ten; while, as Graham drily notes in his paper, "we are aware of no serious hypothesis that listening to relaxing music is hazardous to people's health, and there certainly has been no publicity about this possibility."

The bar charts of "hazard confidence scores" that accompany Graham's paper do indeed show that heavy smoking got a huge percentage of tens (its mean score was 9.1), and that the noughts predominated when it came to relaxing music. But none of the other items registered a majority in favour of non-hazardousness. Exposure to electric and magnetic fields, the item that got the most diffuse range of responses, still found more people believing in the danger posed by EMFs than not - although, according to Graham, its mean score of 5.5 suggested "lack of confidence one way or the other".

What is interesting, though, is the huge gender gap that he sees in the research (see table). In a finding that is "quite consistent" with similar surveys, women much more than men feel that such things as the quality of city air or pesticide-treated fruit are a threat.

The gender difference is at its largest for global warming and smallest for medical X-rays - which anyway rank near the bottom of the table in terms of perceived danger. "With X-rays, the benefits may be influencing people's judgements," observes Graham. "But it's hard to say. We haven't done an analysis of what goes on in people's heads."

There were other, less marked, differences. As well as being asked the standard demographic questions (age, sex, level of education, children in household, and so on), interviewees were questioned about their own tendency to risky behaviour and quizzed on their mathematical literacy, knowledge of actuarial risk and their social values. So Graham's analysis also found that parents and people with egalitarian values were more likely to show confidence that the eight phenomena were hazardous. On the other hand, the better educated were less likely to be sure about the hazards, as were those with actuarial knowledge. (The respondents had good actuarial risk knowledge if they correctly chose traffic crashes over handguns, and heart disease over cancer, as the more common causes of death in the US.) But Graham's paper, which will be presented to the conference next week, ends with another correlation. This correlation is between the frequency of news stories about a hazard and its high score in the survey respondents' assessments. In other words, the more something gets mentioned in the media, the more likely it is to be seen as a hazard (see table).

The figures come from a two-year search through Nexus, the computerised database that logs all references in major newspapers and magazines in the US. Although Graham concedes that further study is needed, he feels he is on to something.

Is it just coincidence that of the eight selected hazards the two most written about - ETS and ozone depletion - had the highest mean hazard confidence scores? And if it isn't, what should we conclude?

"People's perceptions influence their actions, and those of politicians and other public officials," says Graham. "This is one of our problems. Sometimes public perceptions of risk are too low, sometimes too high I There is a tendency to respond to what people believe or are thought to believe, but we need leadership."

Having been in Paris when the BSE scare began, he was "fascinated" to observe "a clear example of the way a hazard was covered by the media and how it had a powerful effect on politicians. BSE is not the most serious public health issue, but it was clearly very newsworthy."

Graham, who is a professor of policy and decision sciences, is clearly exercised by the need to "get public attention to more familiar hazards" - such as road crashes, death from accidental falls in the home or drowning in boating accidents. "Traumatic injury in daily life of the more frequent kind doesn't make for interesting reading," he comments.

"The mass media are businesses, and respond to what they think the customers will be more interested in I We need other institutions to counteract their dysfunction."