The pattern of human life is profoundly influenced by the rotation of the Earth and the resulting day-night cycle. This circadian rhythm influences our biochemical and physiological states as well as our psychological and social behavior.
Our relationship with the day-night cycle can vary widely. Some people—night owls—prefer to get up and stay up late. Others—larks—are morning people who are most active early in the day. Others do not fit either “chronotype.”
Clearly, a person’s chronotype will have a huge impact on the people they interact with—it’s hard to interact with somebody who is asleep when you are awake. So it’s easy to imagine that night owls are more likely to interact with each other than with larks and vice versa. But the large-scale evidence for this kind of social behavior has never been gathered.
Today that changes thanks to the work of Talayeh Aledavood at Aalto University in Finland and a few pals, who have studied the social network and sleeping patterns of over 1,000 individuals for a period of a year. And their work produces some counterintuitive findings.
Their method is straightforward. The team gave 1,000 volunteer students smartphones equipped with an app that measures the phone’s activity—the times it is used, for example—and the number of people it calls or texts.
That gives the researchers the raw data to study each person’s pattern of daily behavior. “We use time-stamped data on ‘screen-on’ events from the smartphone data-collection apps to assign a behavioral chronotype to each participant,” say Aledavood and co.
The team then categorized people as “larks” if they had more early morning activity than expected—that is, activity between 5 a.m. and 7 a.m. They defined “owls” as people who had more-than-expected activity between midnight and 2 a.m. The team categorized the rest—more than half of all the participants—as intermediates.
Next, the team built a social network showing the links between all the participants. Each individual is a node in this network and is linked to another if they have communicated with each other via a phone call or text.
Finally, the team analyzed the social networks associated with owls and larks to see how they differ. In particular, they looked at how popular each node is, how likely a member of a group is to connect to others of the same group, whether they play central roles in the network, and so on.
The results make for interesting reading. “Evening-active owls have larger personal networks than morning-active larks, albeit with less frequent contacts to each network member,” say Aledavood and co. They also say owls are more central in the network.
The way members of these groups connect to others like them—their homophily—is the most unexpected finding. Aledavood and co say that owls connect to other owls more often than pure chance would suggest. So they are strongly homophilic.
But larks show no such tendency. “Surprisingly, this homophily is not visible in the case of larks,” says the team, clearly puzzled by this finding.
One potential explanation is that social gatherings tend to take place later in the day. So people who stay up late are more likely to take part and to organize them. “It is perhaps not surprising that there is a bias in favor of the evening-active chronotype,” say Aledavood and co. The researchers also suggest that larks spend more of their time alone and interact with fewer people because social events are much rarer in the early mornings.
That’s interesting work with implications beyond social networks. Researchers have long known that a person’s chronotype is intimately linked to all kinds of outcomes, such as academic performance, body mass index, and physical and mental health.
Other research has shown how other behaviors seem to be linked across social networks. For example, people who have a higher body mass index are more likely to be linked to other people who are similarly overweight.
It’s just possible that a better understanding of chronotypes could reveal important insights. What’s more, for network scientists and anthropologists, this is potentially low-hanging fruit because there are numerous large-scale data sets from mobile phone companies, for example, that could help tease these effects apart.
We’ll be watching to see what they find.
Ref: http://arxiv.org/abs/1709.06690: Social Network Differences of Chronotypes Identified from Mobile Phone Data