Sander van der Linden is Director of the Cambridge Social Decision-Making Lab, Department of Psychology, University of Cambridge. His edited answers are in italic to distinguish them from my commentary.
Why should we care about COVID-19 statistics?
As we come out of lockdown, it is important that members of the public understand the risks of COVID-19.
Why is the perception of risk so important?
The pandemic is as much a behavioural as a biological problem. We need a vaccine, but we also need people around the world to coordinate their behaviours to help slow the spread of the virus.
Behavioural science is also relevant in terms of how to communicate the science to the wider public, how to communicate uncertainty and risk, and how to protect people from the onslaught of fake news and misinformation about COVID-19.
Human behaviour is notoriously variable, and difficult to change and predict. I think it’s one of the biggest challenges of this pandemic. If we don’t get it right, there’s a chance the spread of the virus will pick up again as restrictions are relaxed.
Can we get a sense of the risks from official statistics?
Overall, the UK has been hit particularly badly and people are right to be concerned. I am interested in what that means psychologically.
Generations of politicians of all persuasions have sought ways to frame statistics to tell the stories that suit their narratives and this does not always help the public to understand what is going on.
Spiegelhalter, who is the chairman of the Winton Centre for Risk and Evidence Communication at the University of Cambridge, pointed out that the official death rate falls short of the actual death rate. The same goes for the amount of testing when compared with the capacity for testing.
And while he agrees that death rates are difficult to compare between nations because of differences in the way data are gathered, as the Government has emphasised, he still believes that massive differences between the death rates seen in countries such as the UK and US – compared with South Korea and Germany – are still worth investigating.
What do we mean by ‘safe’?
I am interested in whether or not people’s perception of risk is accurate, and what it means for that perception to be accurate because everyone has different risk preferences, tolerances and so on.
There are many dimensions, from the risk to yourself to the risk you pose to others. In other words, the accuracy of risk perception is a very difficult question.
Journalists love to ask if an activity is safe. If they mean, is the activity ‘free from the risk of harm’, then the answer has to be ‘no’: the risks may be vanishingly low but no human activity is safe, that is, zero risk.
It is more helpful to think of how we can reduce risk, just as we do with daily dangers, like cycling and crossing the road, so as to make an activity ‘safe enough’.
How certain can we be of the risks we face?
When it comes to COVID-19, we can’t.
We have been exploring uncertainty in its many forms and you can break it down into two basic kinds – firstly, ‘epistemic uncertainty’, which refers to things we don’t know but could come to know through the scientific method of theory and experiment.
Secondly, aleatory uncertainty (In Latin, alea means dice or gambling) — unknowns about the future due to randomness, chance or luck.
Uncertainty about the development of the pandemic and, most importantly, when it will end, will be down to a mix of both kinds.
Is there such a thing as ‘behavioural fatigue’?
I was sceptical from the outset about the idea of ‘behavioural fatigue’, that people would grow tired of self-isolating, which is one reason given for delaying lockdown.
Much of the literature is about the negative psychological impact of isolation on people’s mental health, which is a very different question from how long people can stay inside and change their behaviour, in the sense of not going outside.
We will have to get used to wearing masks – what does psychology tell us?
Face masks are mandatory on public transport in England and this measure is now being adopted as lockdown is being relaxed.
Social norms are important for their introduction: the more people who wear masks, the more people will join them. You can boost their adoption with role models of people wearing masks, cues from government officials, or celebrities advocating this new norm.
There is also what we call a ‘spillover effect’: if you’re outside and you see a mask, it reminds you maybe not to shake other people’s hands, and that you need to keep your distance.
What can we say about the risks of COVID-19 at the moment?
David Spiegelhalter relies on statistical science to turn the variability that we see in the world into meaningful numbers.
He likes to remind us that we’re all going to die sometime, and the rate at which we do so is faithfully recorded in the life tables provided by the Office for National Statistics.
Every year around 600,000 people die in the UK while a study by Imperial College estimated that if the pandemic went unchallenged, there would be around 510,000 deaths. Roughly speaking, he says that getting COVID-19 is like packing a year’s worth of risk into a week or two.
There are 7 million school-children between 5 and 14 in England and Wales, and, up to 29 May, there had been two deaths with COVID-19 registered in this age group. More may be registered later, but it’s clear they have faced a tiny risk even at through the peak of the epidemic, said Spiegelhalter.
In comparison, nearly 2 per cent of all those aged over 90 in the country have so far died from COVID-19, in a nine-week period.
These are average risks, he emphasised: they are higher for men, for ethnic minorities, for occupations such as bus drivers and care-home workers, for those living in deprived areas, and those with pre-existing conditions, although disentangling the individual contributions of each of these factors is not yet possible.
So, what are ‘the risks of dying from COVID-19’?
‘Please permit me a rant’, said Spiegelhalter. He points out that phrases like ‘the risk of dying from COVID-19’ are ‘deeply ambiguous’: that could mean the risk of dying among people who are infected (this is called the Infection Fatality Rate, or IFR), or the risk of dying among people who are not currently infected (this is the Population Fatality Rate, or PFR.)
For example, the risk of dying if infected, if over 45, is very roughly the same size, but of course additional to, a normal year’s risk, he said.
Our understanding of COVID-19 is evolving. How do we cope with uncertainty?
While there are still many unknowns about coronavirus, we have done research to show that the public can deal with uncertainty when it is quantified and made numerical.
Importantly, being transparent about uncertainty does not harm the public’s trust in the facts or in the source of those facts.
How did you study the link between trust and uncertainty?
Using online experiments, we weighed up reactions to uncertainty expressed in statements about various subjects, from the number of tigers left in India to the increase in global average surface temperature between 1880 and 2010.
We repeated this in different countries and with a field study on the BBC News website, expressing uncertainties with a numerical range or percentage, or qualitatively, using a word such as “estimated” or “approximately.”
We found that precise numerical statements were more effective both in conveying uncertainty and in maintaining trust.
When we recently repeated this experiment with COVID-19 death rates, we got the same result.
How do people perceive the pandemic risk around the world?
We did a study of public attitudes of around 7000 people across Europe, America and Asia between mid-March and mid-April and found that the baseline concern about the coronavirus was high in all countries, though people in the UK have the highest overall levels – more than Italy or Spain or the US – while those in South Korea are relatively least concerned.
Insights into risk perception are important because, given we have no vaccine or effective treatment, we are relying on people changing their behaviour to put the brakes on this pandemic
We think this is the first comparative evidence of how to perceive the risk of COVID-19 around the world and this perception will influence, in part, the willingness to adopt protective behaviours such as physical distancing.
And, indeed we found that greater concern about the virus did indeed correlate with taking a number of preventative public health measures such as increased hand washing or wearing face masks.
We also have new evidence to suggest that, as the UK infection and death rates decline, so does the concern, as you would expect.
We also found that “prosociality”, or a belief in the importance of doing things for the benefit of others, was linked to heightened concern about the virus in nine of the 10 countries. As my co-author, Claudia Schneider, has pointed out, the UK’s ‘clap for our carers’ campaigns help us to publicly signal prosocial intentions.
Men typically had lower levels of concern about the virus than women, despite the fact that, on average, COVID-19 appears to be considerably more dangerous to men if contracted.
Does your survey support claims that COVID-19 causes more problems in countries with ‘ILLIBERAL POPULIST LEADERS’?
Political ideology was less significant for risk perception overall, although a more conservative outlook was associated with lower levels of concern in the UK and the US.
We also found that an “individualistic worldview” – inferred from a belief that governments meddle too much in our lives – related to lower levels of concern about the risks of coronavirus.
While this worldview is famously associated with certain US states, it was also significantly related to risk perception in several other countries, such as the UK
The media shape public attitudes – how did they report your findings?
The media can produce quite different impressions, depending on how they frame a story.
When it comes to the two-metre distancing rule, the risk of infection could double if reduced to one metre, according to one report, while keeping one metre apart can ‘slash the risk of catching coronavirus by 80 per cent’ according to another.
Both represent the same study – the difference reflects the way the risk is framed, that is, if the risk under discussion is relative or absolute (a ten-fold rise in cancer risk is not so alarming if it turns out to be a tenfold rise in a tiny absolute cancer risk).
It’s complicated, as the media can attenuate or amplify risks with their headlines, and that in turn could influence people’s behaviour, which could in turn increase (or decrease) the objective risk.
For example, if people think you only need to keep 1m distance based on the latter headline, the infection rate might go up, deaths might rise, and the risk factor could change. That’s why accurate communication is so important.
Moreover, the absolute risk to individuals might be small, but that doesn’t remind individuals of the risk to vulnerable populations. Therefore without social context, considering risk in all of its psychological dimensions is easily misunderstood or interpreted (e.g. as risk from an individual’s perspective and not at the societal level).
When it comes to the two-metre rule, a Lancet comment on the same study reported: ‘a reduction in risk of 82% with a physical distance of 1m in both health-care and community settings …every additional 1m of separation more than doubled the relative protection.’
The two-metre rule is now under review by Government.
How influential is fake news?
Sadly, fake news is very influential.
We studied a climate change disinformation campaign, the Oregon Petition, which in 2007 falsely claimed that over 31,000 American scientists had rejected the position that humans caused climate change.
In a survey of 2000 people, we asked about the consensus around climate change, telling some in advance of seeing the petition that in reality, 97% of climate scientists agree that humans are responsible for climate change and pointing out flaws in the petition.
For instance, among its tens of thousands of names are people like the deceased Charles Darwin and the Spice Girls, and that fewer than 1% of the signatories are climate scientists.
What surprised us is that, compared to what people perceive is the size of the scientific consensus on climate change, the misinformation neutralised the correct data.
There is a lot of fake news about COVID-19 – what can we do about it?
Fake news is a problem and the constant stream of misinformation about COVID-19 is a major challenge. For example, concerted disinformation campaigns have the potential to undermine public willingness to vaccinate if people do not believe the vaccine is safe.
To show how to deal with fake news, we designed an online game with partners at DROG, a Dutch anti-misinformation platform (you can play the Bad News game here) in which players enter a fictional social media environment to “walk a mile” in the shoes of a fake news creator.
After playing the game, we found in a study of 15,000 people that they became less susceptible to future exposure to common misinformation techniques, an approach we call ‘prebunking.’
We also find, in the preliminary results of another study, that numeracy was important – the higher the numeracy score, the less vulnerable to misinformation.
Just as administering a vaccine – say a weakened dose of a virus – triggers the production of antibodies to confer immunity against future infection, the same can be achieved with information.
We are building a COVID-19 version of the game. By using the new science of ‘prebunking’ to actively expose people to severely weakened doses of the tactics used to produce fake news, people gain psychological immunity (or mental ‘antibodies’) against misinformation.
Anti-vaccination propaganda, false cures, homemade remedies, lunatic conspiracy theories about 5G, even rumours that the pandemic is a hoax, are among the social media posts spreading misinformation about COVID-19.
One analysis of 100 million individuals expressing views about vaccination, emerging from the approximately three billion users of Facebook, concluded: ‘Our theoretical framework reproduces the recent explosive growth in anti-vaccination views, and predicts that these views will dominate in a decade.’
Vaccine hesitancy is already taking a huge toll. Even before the COVID-19 pandemic, The World Health Organisation estimated that a further 1.5 million deaths annually could be avoided if global coverage of vaccinations improved.
How else can you spot fake information?
One way is to use a surprising finding known as Benford’s Law, the observation that the first digits in all sorts of data sets – electricity bills, street addresses, stock prices, house prices, population numbers, death rates and more – are not evenly distributed such that the digit “1” is the most frequent, followed by “2”, “3”, and so down to “9”.
In other words, in many naturally-occurring collections of numbers, the leading significant digit is likely to be small.
Malcolm Sambridge of the Australian National University and Andrew Jackson of the Swiss Federal Institute of Technology, ETH Zurich, found that records of cumulative infections and deaths from the United States, Japan, Indonesia, and most European nations adhered well to the law.
However, they did reveal some anomalies, for instance in data from the Czech Republic.
HOW CAN I FIND OUT MORE?
David Spiegelhalter of the Winton Centre for Risk and Evidence Communication provides his own updates.
There is more information in my earlier blog posts (including in German by focusTerra, ETH Zürich, with additional information on Switzerland), from the UKRI, the EU, US Centers for Disease Control, WHO, on this COVID-19 portal and Our World in Data.
The Science Museum Group is collecting objects and ephemera to document this health emergency for future generations.