The end of Dunbar’s number: Have our social networks changed for good?

A cartoon of a social network

This article originally appeared in The Oxford Scientist’s Print edition, Networks, in Trinity Term 2022.

When the COVID-19 pandemic hit the world in 2020, our social circles shrank away to nothing. For many of us, the lack of social contact was hugely detrimental to our mental health and was a stark reminder of how social we are as a species. In fact, it’s long been thought that humans innately create social networks of approximately 150 people, a figure referred to as Dunbar’s number.

Dunbar’s number is the proposed limit on the size of our social networks, made up of people with whom we can sustain stable relationships. In a paper published in 1992, Professor Robin Dunbar imaged the brains of 38 non-human primate species. He paid particular attention to the size of the neocortex, the region of the brain that relates to cognitive abilities such as communication, planning and thought, finding that it was tightly correlated with the average size of the primates’ social groups. Dunbar then extrapolated to the human neocortex, predicting that humans are limited to social networks comprising of 148, usually rounded to 150, individual relationships. Any larger than this and we must fragment into smaller groups to maintain our innate social structures.

Of course, there are very few human societies left on Earth that have fewer than 150 people—millions and even billions are now the norm. But it is the definition of ‘relationship’ that is important when understanding this number. The apparent limit of 150 is on the number of relationships we have, where we know each person in a group and how they relate to everyone else. Writing in his 1998 book, Grooming, Gossip and the Evolution of Language, Dunbar describes these relationships as ‘people you would not feel embarrassed about joining uninvited for a drink if you happened to bump into them in a bar’, or the ‘people you know well enough to greet without feeling awkward if you ran into them in an airport lounge’.

Throughout history, it is surprising how often a figure of roughly 150 crops up. For example, the mean size of a neolithic English village was 160 people, Roman army units averaged 150 men, and even modern Christmas card networks max out at 153 people. In the past, humans would have needed large social groups for support and protection, but not so large that they would be unable to feed every member. A group of around 150 people appears to have been the sweet spot. 

Yet recent analysis has cast doubt on this number. In 2021, a Swedish research group reanalysed Dunbar’s original data, predicting that human social networks could sustain on average 290 relationships. More importantly, however, was the finding that the statistical error on this number was so large that possible estimates varied from social networks of 2 to 336 people. Analysis of an expanded dataset found even greater disparity, varying from relationships between 4 to 520 people. ‘It is not possible to make an estimate for humans with any precision,’ commented Andreas Wartel, a co-author of the study.

Real-world data on the size of human social groups also undermine Dunbar’s theory. A 2010 study calculated that people actually know between 472 and 611 people. Individual variation also means one number will certainly not apply to all people. Women are thought to be better at understanding other people’s perspectives than men, which has been linked to having larger social circles. Personality traits are also important, with extroverts having larger networks and introverts having smaller ones. Even the highly cited examples of 150-people networks have been criticised as overwhelmingly skewed towards rich, educated, and industrialised societies, with non-western cultures rarely mentioned. Confirmation bias may well be a factor in the popularity and acceptance of Dunbar’s number.

The theory also assumes that the human brain, although slightly larger, retains the same wiring and function as it did over 10 million years ago when we evolutionarily split from our primate ancestors. In the meantime, we have developed new cognitive abilities such as language and planning that affect our social structures. The size of our brains no longer equates to its abilities in the same way it does in monkeys. 

Besides, in 2022 there is another huge influence on our societies that researchers in 1992 could not account for—social media. A survey of Facebook users in 2016, also by Dunbar, found that over 50% of people had more than 200 friends. The average number of Twitter followers is larger at 707. Following a new account doesn’t push out the memory of another, so our brains must have the capacity to remember and keep track of more than 150 people. ‘This number would make sense if we still relied on a Rolodex and talking to people,’ says Angela Lee, a professor at Columbia Business School, ‘but that’s not the world we live in’. 

Once again Dunbar maintains it is the type of relationship that is important. Often the quality of the relationships on social media are low and one-sided, for example with celebrities or influencers, contradicting the definition of relationships as between people that know each other. Still, even on social media the number of 150 pops up. The mean number of Instagram followers is 150, and research suggests that outside of our top 150 friends and family the other people that we follow are just acquaintances or people we have never met. In fact, 60% of our time on social media is spent talking to just our 5 closest contacts.

Maybe this transformation in the structure of our social networks signifies a change in how we define ‘relationships’. References to airport lounges and pub trips certainly feel outdated since the pandemic, and even as we return to normal, our work and social routines may have been irreversibly altered. There is also evidence that 18–24-year-olds, who have spent more of their lives with the internet, have many more social media contacts than people aged 55 and over, yet they do not feel that the quality of their social lives is any less rich or meaningful. Perhaps over the last two decades we have adjusted to online relationships, needing less in-person contact and distributing our limited time and energy more broadly across more people. 

On the other hand, maybe Dunbar is right after all. Every other species on Earth develops stable social networks that have been largely unchanged for hundreds of thousands of years, why should humans be any different?  Maybe we do have a cognitive limit to the number of relationships we can sustain. But what sets humans apart is that we no longer rely solely on our brains for our relationships—we can develop new technologies to surpass our innate capacities, allowing us to manage more relationships than ever before. Despite the lockdowns and restrictions, our social networks keep growing, and who knows where this will take us.