skip to content

SPEAKER SPOTLIGHT: Professor David Spiegelhalter, Chair of the Winton Centre for Risk and Evidence Communication and Dr Sander van der Linden, Assistant Professor in Social Psychology, Department of Psychology

Professor David Spiegelhalter is Chair of the Winton Centre for Risk and Evidence Communication at the University of Cambridge and Dr Sander van der Linden is  Assistant Professor  in Social Psychology in the Department of Psychology at the University of Cambridge. They have both collaborated on a new book, Risk and Uncertainty in a Post-Truth World, about which they will be speaking at the Cambridge Festival of Ideas on 19th October, 1-2pm.


Question: Do we need more education at an earlier age about statistics so we can understand risk better?

Professor Spiegelhalter: In an era of rampant misinformation, we all need to be able to take a suitably sceptical approach to claims being made by people who want to influence us. Many of these claims involve numbers, and this means that data literacy is an essential skill in the modern world. We need young people to be data literate, which should help them to understand risk better.


Question: From what age do they need to learn?

Professor Spiegelhalter: Primary school children can be taught to critique claims, which has been shown in large-scale experiments in Uganda. We made a BBC World Service documentary on this work: You can handle the truth.


Question How should we teach probability, given it links to so many areas? Does it need to be included in different lessons?

Professor Spiegelhalter: It just so happens that Jenny Gage and I have written a book – Teaching Probability – which shows how to teach the subject in a way that has been shown to be effective for both children and adults. This primarily involves the idea of expected frequencies – for example, instead of thinking of the risk of an average post-menopausal woman getting breast cancer as being 10% or 0.10, we think of what we would expect to happen to 100 such women: 10 would be expected to get breast cancer. This use of whole numbers clarifies the people we are talking about, helps communication since we can visualise 100 women and allows straightforward analysis of more complex situations.


Question: What about adults - including journalists?

Professor Spiegelhalter: Communicating risks in terms of expected frequencies is popular with both journalists and doctors, but they need training and help. To use a simplified example, if a treatment such as five years of combined hormone replacement therapy increases the risk of breast cancer by say 20%, this is a relative increase over an absolute risk of 10 percentage points – both journalists and doctors struggle to communicate this. But if we think of 100 women, 10 will get breast cancer anyway, and this is increased by 20% - to 12 – if all 100 women take HRT. So there are two extra cases of breast cancer due to these 100 women taking five years of HRT, and someone may consider this risk worthwhile if their quality of life is sufficiently improved. This way of telling the ‘story’ helps comprehension and informed decision-making.


Question: We are so inundated with reports and research studies, which often appear to conflict with each other. Is it not just a question of understanding statistics, but also teaching people the skills to know which reports to take seriously?

Dr van der Linden: Absolutely. It can be very confusing. For example, some studies may claim new health benefits of foods and drinks we previously thought were bad for us, and vice versa. Or reports might come out that another famous finding “did not replicate”. I think the key answer here is that no single study is going to provide any conclusive answers. I like to think of science as a cumulative effort: it takes many studies, many replications, over many years to establish what we call “weight of evidence”. There will always be uncertainty about research results, we need to accept that, but rather than putting our faith in a single study, our judgments should be guided by the weight of the evidence. To give an example, the fact that a study does not replicate doesn’t necessarily invalidate the original finding. You might, however, adjust your internal credibility meter so that you now evaluate the finding with less certainty and confidence than before until more evidence becomes available. This implies cultivating public attitudes toward evidence that are open to uncertainty though!


Question: If people are generally risk averse, do we need to approach risk in a different way?

Dr van der Linden: It’s a bit more nuanced than that. Yes, when confronted with uncertainty people will often attempt to reduce it. For example, one purpose of human communication is to reduce uncertainty about what someone else is thinking. But Kahneman and Tversky’s ‘prospect theory’ (for which Kahneman won the Nobel prize) predicts something very interesting. It suggests that people are risk-averse mostly in domains in which there is something to be gained. In other words, if I make you choose between a certain reward (say £50) or a gamble (a 50% chance you’ll get £100, but with a 50% chance you’ll end up with £0), most people will just take the £50, thanks very much! On the other hand, when we’re in a loss domain, i.e. when we have to lose something (£50 for sure, or there’s a 50% chance you’ll lose £100 pounds but with a 50% chance you’ll lose 0), all of the sudden people become more risk-seeking! I mean, if you have to lose something anyway, why not take a gamble, right? What I take away from this, and what we find in our own research, is that context matters a great deal for how people will react to uncertainty. For example, when it comes to basic facts and numbers in the media, people are not uniformly “allergic” to uncertainty!


Question: What can be done to mitigate concerns about risk?

Dr van der Linden: I think there is a common concern among producers of official statistics, governments and media organisations that people will react negatively to uncertain information. For example, think about the number of people who might be affected by a disease outbreak. If I provide a single estimate (e.g. 10 million), it looks like I’m precise and know what I’m talking about. If, on the other hand, I am honest and admit that this is difficult to measure and there’s lots of uncertainty (we estimate between five and 12 million people), the audience might think you’re incompetent, become worried, trust you less, and so on. But in our research we actually don’t find this to be true. When exposed to uncertainty people do think the numbers are more uncertain and they feel more uncertain, but they don’t necessarily trust them (or the communicator) less. I think the more we openly talk about uncertainty, the more common its reporting becomes, the more used to it people will get, and the more we can mitigate concerns about negative public reactions to uncertainty. But to not run the risk of sounding too certain myself, as I said before, context is key, the nature and dosage of uncertainty matters, the context in which it arises and how it is communicated to people. And that all, of course, is part of the exciting psychological study of judgment in conditions of uncertainty.


Question: What role should academics play in helping the public to understand risk generally?

Dr van der Linden: An active one! This is something that we also cover in our new book. Navigating risk and uncertainty in - particularly -  an era of fake news and post-truth, where facts are increasingly contested, is going to be a massive challenge. Having said this, I think scientists have an opportunity and perhaps even an obligation to share with the public what they learn so that everyone can benefit from these insights. A wonderful example is the Cambridge Professorship for the Public Understanding of Risk, which David Spiegelhalter held for many years. He has done so much to inspire and engage the public with the issue of how to grapple with risk in every judgment and decision. Unfortunately, David can’t clone himself, but I think we could all learn from his example of how to do better.