Social distancing has been a key part of the pandemic response: we all know that our chance of infection is reduced if we minimise the contact we have with others. Yet there are countless stories of people covertly meeting up with friends and family even at the height of lockdown.
Clearly many of those who disregarded the rules did so because of a desire for social interaction and support. But a new study in Humanities & Social Sciences Communications suggests there may be another reason too: we simply underestimate the risk of contracting Covid-19 from friends.
Across a series of five studies, Tobias Schlager from the University of Lausanne and Ashley V Whillans from Harvard Business School found that people consistently feel that friends are less of a Covid risk than strangers.
In the first study, participants assessed the likelihood of catching Covid from a friend, family member, stranger, work colleague, or acquaintance, rating the risk of each “compared to any other person”. The team found that participants actually believed most of these people posed lower risk than the average person, but crucially they rated friends the lowest risk and strangers the highest.
If participants had been around strangers more than friends during the pandemic, it would make sense that they’d rate strangers as higher risk. So in the next study the team asked participants to imagine they had gone to the supermarket and met a friend, family member, cashier, stranger, and colleague. They rated the likelihood of contracting Covid from each of these people using the same scale as before, and also indicated how long they’d expect to talk to them.
Participants rated friends and family as posing significantly lower risk than the average person. They also rated the cashier as significantly higher risk than the average person, followed by the stranger. These results held even when the team controlled for the expected length of time interacting with each person, so it wasn’t simply that participants saw strangers as a higher risk due to spending more time around them.
In another experiment, participants imagined they were out to dinner and were approached either by a friend they hadn’t seen in a while or a stranger, who invited the participant to join them. People were more likely to say they’d join a friend’s table than a stranger’s. Perhaps unsurprisingly, the team’s analysis showed that this was partly because participants were more interested in having dinner with the friend, and also because they thought the friend more likely to follow social distancing rules. But there was another reason too: participants rated the stranger as a greater Covid threat (believing they were more likely to be infected, for example), and this belief also contributed to their decision.
Finally, the team also found some evidence that people may physically distance themselves more from strangers than friends. Participants positioned a figure on their screen to indicate how close they would stand next to someone else in a park; those who read they were meeting a friend placed the figures closer than those who read they were meeting a distant colleague. However, participants who were also reminded about the dangers associated with spending time in close proximity to others put a greater distance between themselves and the friend.
Overall, then, the work provides quite convincing evidence that people underestimate the risk posed by friends. And yet, the authors note, close contacts actually play a key role in spreading the disease. Counteracting the perception that our friends are less of a risk to us could therefore be an important part of pandemic communication strategies.
The results are consistent with other work finding that familiarity and trust is related to reduced perceptions of risk, the team notes. They are also reminiscent of the illusory superiority bias, in which we think of ourselves as “better than average” in various ways. Some work has found that we also believe that our friends or partners are superior to other peers — so perhaps this is a similar effect, where we see our friends as “better” than average at not spreading disease.
There are some limitations. In particular, all of the studies are based on self-report data and imaginary scenarios. The authors note that in the middle of the pandemic it would not have been possible to conduct a real-world experiment; still, it would be interesting to use other methods such as virtual reality to look at how people behave towards friends and strangers.