Uncertainty in the media: a mixed picture on swine flu

As of the 23rd May 2022 this website is archived and will receive no further updates.

was produced by the Winton programme for the public understanding of risk based in the Statistical Laboratory in the University of Cambridge. The aim was to help improve the way that uncertainty and risk are discussed in society, and show how probability and statistics can be both useful and entertaining.

Many of the animations were produced using Flash and will no longer work.

BMC Public Health is one of those online, open access journals that are becoming more and more prominent in academic publishing. A couple of recent articles throw rather different lights on the communication of health messages to the public. One is pretty positive about how Australian television reported the 2009 swine flu epidemic. The other reports an analysis of school books in Spain, and I'll return to that in another blog entry.

A team from the Universities of Sydney and Canberra looked at the reporting of swine flu (H1N1 influenza) on the news and current affairs broadcasts of five Australian television channels, between April and October 2009. They come to fairly positive conclusions about the reporting, while making suggestions on how it could be improved in similar future situations.

What particularly caught my eye was that the researchers explicitly considered how well uncertainty had been communicated. You probably recall that there was huge uncertainty about how big the epidemic would eventually turn out to be, as indeed is usually the case with such global epidemics. So the message to be got across was quite complicated - the epidemic might be very big, so there had to be preparation for that risk, but on the other hand it might turn out not to be huge. The researchers felt that this uncertainty was got across adequately and appropriately.

The messages on what the public should do about the epidemic threat weren't always so appropriate. The researchers draw attention to the fact that conflicting advice was given - sometimes in the same news bulletin. For instance, in a single bulletin, two different government health authorities said "I think it’s prudent that
all households think about their readiness" and "there is no need at this point for people to start thinking about those preparations…".

What struck me about the examples of advice, quoted in the paper, is that they show little sign of referring to the uncertainty. Presumably the fact that the uncertainty was so great is one reason why the advice was sometimes contradictory. But, judging by the quotations in the paper, most of the advice seems to have been pretty dogmatic - of the "you must do this" or "you really should not do that" kind. It reads rather as if the authorities (and the media reporting them) are perfectly happy in pointing out to the public that there's a lot of uncertainty, but much less happy about trusting the public to take account of that uncertainty when advising them what to do. Doesn't feel quite right to me.



In this type of situation the authorities may be criticised if they do warn, and criticised if they don't. If I advised them that the probability of a serious epidemic was low they might reassure the public, in which case an epidemic would be more likely. If I advised them that an epidemic was likely they would warn the public, making an epidemic unlikely. This is similar to Keynes' reflexive probability in investments. If the risks from not warning the public are high then one needs to convince them that the probability is high enough to justify easy precautions (e.g. 'bin it') but not high enough to have a serious economic impact (e.g. by staying at home). Thus the probability one is trying to convey has everything to do with public sensitivity and little to do with the objective conditions of the virus. I am starting to blog on this at djmarsay.wordpress.com, which has references.