Making sense of behaviour changes during COVID-19 using behavioural science

“Science is inherently unpredictable” Malcolm Gladwell states in his recent Revisionist History podcast (series 5, episode 3). In the case of behavioural science, it is an especially unpredictable science, if we are to even consider it a science at all. 

No science is obvious or predictable because the environment in which an experiment is carried out can lead to results being corrupted. This unpredictability is the reason we rely upon experts to conduct research and establish knowledge. 

Over the last decade behavioural scientists – experts – have been raised to positions of power in government and in 2020 put under an intense spotlight. As COVID-19 has necessitated behaviour change, government has used behavioural science to inform its response. Steven Taylor, author of ‘The Psychology of Pandemics’ puts it, “pandemics were caused and contained by the way that people behaved. Pandemics are controlled only when people agree to do particular things”. 

In addition, under the current circumstances the media, the community we live in, the behaviour of friends, family, and society as we see it has become more prevalent in shaping our behaviour because the usual conventions no longer apply. No longer are we sure what behaviour is legally required or socially acceptable. In short, rightly or wrongly, it is very easy to view COVID-19 as a macrocosm behavioural science experiment, full of complexity, with all of us as test subjects and life or death consequences. 

So now, after some time has passed and with the benefit of hindsight we are able to review which behavioural science biases have been influencing us in recent months and what effect they have had. It is this I will attempt to do over a series of articles in the coming weeks, beginning with how the threat of COVID-19 was initially assessed...

Confirmation Bias and Probability Weighting: Assessing the Threat of COVID-19

Initially when faced with the novel coronavirus, Asian nations over-weighted the unknown, small probability and prepared for the worst. By contrast, in March the UK government under-weighted a known, large probability. Why? 

Ashok Sethi, Co-founder and Director at Behave Consulting commented during NudgeStock (1:20:38) that “in Shanghai public transport and shops never went through a full lockdown but consumer behaviour changed. There was homogeneity of behaviour in China. They thought it was not just safe practice but a sacred duty… but the most important factor was probability weighting… the behaviour change seemed disproportionate, it was obvious people were overweighting a small probability, being on the cautious side”. 

Compare China to the UK and the difference is stark. In the UK we were dealing with a known risk. The impact of COVID-19 was documented across Asia and Italy, yet David Halpern, Head of the UK government’s nudge unit and member of the Scientific Advisory Group for Emergencies (SAGE), warned of behavioural fatigue when choosing to delay entering lockdown. He suggested that compliance with the rules deteriorates over time so the recommendation was made to delay entering lockdown. However this is backed by scant evidence. A review of the supposedly supporting evidence found that non-compliance does increase over time but crucially, the study was conducted in the context of extending the time armed forces were deployed. Unsurprisingly the results of that study, concluded “a clear and explicit policy on the duration of each deployment” was required. Clearly, the environment in which that study was conducted effected the results. The use of these findings to support the idea of behavioural fatigue was not applicable to the current situation.

Halpern’s desire to express his view led to confirmation bias. He cited the study that supported his view best, not a study that was sufficiently comparable. As a result, opinion was dressed up as science. In this instance it appears confirmation bias was more powerful than probability weighting. 

Probability weighting would normally mean we overstate potential future risks versus potential rewards. However, the UK government’s desire to be right and confirm their beliefs superseded a cautionary approach which would have likely led to rewards/a more favourable outcome. Unfortunately, this conclusion only perpetuates the popular view that ego and the want to be right, is making a mess of politics and having a negative effect on decision making.

Article 2 follows later this week... Short Term Rewards & Conformity: Happy Birthday Hand Washing and Mask On-Mask Off