Featured
Research by Northeastern network scientist Iacopo Iacopini is helping explain the dynamics of group interactions in complex social systems.
LONDON — The holidays bring people together. But have you ever wondered why a big dinner table conversation often splinters off into lots of smaller ones?
A Northeastern University network scientist is helping to explain what is happening during those situations by characterizing the dynamics of group interactions in complex social systems.
Associate professor Iacopo Iacopini found, during a recent collaborative study, that a person’s time spent in a group depended on how long they had been part of it and how large that group had grown.
The larger the group, the less likely it was that an individual remained part of it, the paper found.
Iacopini, a London-based researcher with Northeastern’s Network Science Institute, and his colleagues reached their conclusions after analyzing three different datasets featuring people of different ages — preschool children, freshman students at a university and academics attending a conference.
By using the data and analyzing the children and adults’ movement, they were able to see how groups formed, how long they stayed together and how the groups then broke up.
The findings were published in the journal Nature Communications in the paper “The temporal dynamics of group interactions in higher-order social networks.” Iacopini co-wrote it with Marton Karsai from the Central European University in Austria and Alain Barrat from the University of Toulon in France.
The study looked at what factors influenced how long a group stayed intact. The team found that the “long-gets-longer” — the more time spent in a group influenced how much longer a person was likely to stay in it.
“We actually found that the probability of leaving the current group for an individual,” Iacopini says, “is inversely proportional to the time spent there.
“So basically, the longer you’re in a group, the longer you try to stay in the group. In the same way, it is a higher chance that you leave a group that just started, as opposed to a group that was there for a long time.”
Helping to explain why big group conversations are prone to breaking up into smaller units, Iacopini says his study found that the durability of a group can also be impacted by its size. It is what is known as the social phenomenon of “schisming,” when a conversation reaches its cognitive and social limit for a group and breaks down into smaller parts.
“The group size has an impact, so the bigger the group, the higher the chance that you leave the group,” Iacopini says.
“So there are these two things — the chances of leaving a group depends on the time you spend there and also on the size of the group. And again, that kind of makes sense, right?
“You can witness this at any dinner, basically. If the size of the table, in terms of the number of people, increases above a given threshold, it naturally splits into multiple conversations because cognitive and social limits are reached.”
What surprised the researchers about the results of their study was that there were so many similarities in the way groups of all ages formed and dispersed. Despite the settings being “completely different” from one another, there were “not big differences” between how groups evolved, Iacopini adds.
“Essentially, we didn’t see major differences apart from obvious ones which are basically size — the biggest size a group can reach — and time,” he says.
“The interactions of kids are way faster than interactions of adults. Conversations are longer, there is more meaning, kids might just be playing. But basically, if you scale time and if you adjust for how fluid the system is, you can more or less map it [the group formations] onto one another.
“From a more scientific point of view, we were surprised to find all this common phenomenology in different contexts. The group formation, evolution, dynamics — even across age groups — seemed to be consistent.”
The takeaway the researchers took from their investigation was that there appears to be “common robust mechanisms determining group formation and evolution” even across age groups.
Iacopini says the novel data analysis, newly developed characterization tools and modeling “represent an important step forward in this interdisciplinary research.”
The results “open new avenues within the area of complex systems and the field of computational social science,” he says, and could prove useful when researching both biological contagions, like the spread of infectious diseases, and behavior contagions, such the adherence to imposed restrictions, vaccination and gathering habits.