His lecture was entitled ‘Understanding Uncertainty’, as is his website (understandinguncertainty.org). And he is passionate about this, and the need to communicate risk and uncertainty properly.
He began by explaining that we are faced by limited numeracy amongst the public, and need to respect people’s difficulties and their interest in risk. That is why he stressed the importance of risk communication, the main theme of his talk.
How can we communicate daily risks to people, he asked. In the sphere of health, one of his specialist areas, he said that what was needed was a ‘friendly unit of deadly risk’, namely, a micromort.
This, he explained, is a one-in-a-million chance of sudden death for some defined activity, whether it be walking, cycling or sky-diving.
This means you can use whole simple numbers to compare everyday activities. For emergency general anaesthetic, it is about 10 micromorts. Giving birth in the UK is about 80 micromorts, while going into hospital is about 75 micromorts. And, as Professor Spiegelhalter pointed out, this is for accidental deaths, as opposed to the reason why you entered the hospital in the first place.
Another unit he uses is the microlife, which shows how much life we lose on average when we’re exposed to chronic risks. A microlife is 30 minutes off your life expectancy. He explained that smoking two cigarettes, or drinking seven units of alcohol, would cost one microlife. Each day of being 5kg overweight also costs one microlife.
“Microlives encourage the metaphor that people go through their lives at different speeds, according to their lifestyle. For example, someone who smokes 20 a day is using up around 10 microlives, which could be loosely interpreted as their rushing towards their death at around 29 hours a day instead of 24,” he said.
“What we are doing here,” he explained, “is changing the imagery of risk. Using a different narrative, in this case, the narrative of accelerated ageing. And this has been found to be very powerful in persuading people to [adapt] their behaviour.”
One problem with risk communication is the issue of framing.
Professor Spiegelhalter said he looked for framing issues all the time, and gave one example: a poster campaign on the London Underground which stated that ‘99% of young Londoners do not commit serious youth violence’.
He said this was a positive frame, but he flipped it and was somewhat concerned to discover that, therefore, 1% do commit serious youth violence. With 1 million young Londoners, that meant, in his words, ‘10,000 maniacs running around London.’
Often, he said, the media chose to flip from a positive to a negative frame because it gets more attention. “What this emphasises, when talking about very large risks, is that you have got to do the ‘part to whole’ comparison to avoid framing bias. If you tell someone there is a 5% mortality rate, you have to tell them there is a 95% survival rate, almost in the same breath. They have got to be presented together.”
As part of risk communication, graphics can play an important role in helping to visualise the risks. The problem is how to illustrate these part-to-whole comparisons to give a balanced view. Professor Spiegelhalter pointed to a range of different ways of visualising probabilities, including pie-charts, bar charts, tree maps, and increasingly for risk communication in health, icon arrays (where you show comparisons of two groups of 100 human icons).
He also pointed to fan charts which show the possible way things might turn out—‘possible futures’—but without giving a central estimate. This latter point is important because people (and notably journalists) fixate on the central estimate and not on the range of the possibilities.
He said one of the most exciting developments he had seen in public risk communication was the way that NBC news had illustrated the possible paths of a hurricane from using different computer models.
It used a graphic described as a ‘spaghetti plot’ which showed all the different paths. “It uses a metaphor of ‘possible futures’, by showing the uncertainty of where the hurricane might go, the strength and variability of where it might go, and also some idea of which is the most likely path,” he said. “This idea of ‘possible futures’ is a very powerful image and narrative for communicating risk.”
One problem is that this talk of risk and probability may not reflect the individual’s risk. It is based on similar groups, or populations, and not the individual. “So, any risk communication must be in terms of metaphor and analogy,” he explained.
The world is increasingly facing deeper uncertainties, and when communicating deeper uncertainties to people, it is vital to realise there are many ‘publics’, many different groups of people, who are strongly influenced by their values, their culture, their feelings and their experience, said Professor Spiegelhalter. He pointed out that numbers are almost completely ignored, trust is incredibly important, and people will respond emotionally, using gut feelings.
His advice was: “Be clear about what you are confident about, but also say what you are not sure about. And then be confident about what you are going to do, and tell people what they can do. And good risk communication will do this all the time. And you should also emphasise that things might change. You have to be flexible and adapt, and be resilient with that adaptability. And we should acknowledge the deeper uncertainties. Some things you simply cannot put a number on, and we should admit this.”
He concluded that we have a duty to communicate the magnitude of a risk, even if it is a judgment or an approximation. And because one size doesn’t fit all—people vary—you have to adapt to the audience. And that is why, he said, you need multiple ways of presentation, a range of options including words, numbers, graphs. And he encouraged the audience to explore alternative metaphors, such as possible futures.
“Don’t just talk about worst-case scenarios in risk communication to the media,” he said. “You should communicate the likely, as well as the worst-case, scenario. The media will often take the negative worst-case scenario and we need to be prepared for that. By choosing the right frame and the right language, and being very careful about that, you might be able to get the right message across.”
He ended by quoting Donald Rumsfeld, whom he described as ‘the great sage of risk analysis’: ‘Sometimes there are unknown unknowns’. He concluded: “There are things we don’t know we don’t know and we need to have the humility to admit that this may be the case.”
Visit Professor Spiegelhalter’s website (understandinguncertainty.org) for articles, presentations and links to papers.
Please sign up here to our full-time mailing list to ensure that you receive our weekly newsletter.