Our natural tendency to seek similarity and close social triangles when meeting and befriending people can lead to polarized societies. A recent study by CEU researcher Gerardo Iniguez and his co-authors highlights the importance of tolerance towards people who are different from us, especially in crisis situations such as the current pandemic.
Following many decades of study, researchers have identified two main ways in which we make new social connections: finding people similar to us (called “homophily”, from the Greek “homou” meaning together, and “philia” meaning friendship), and meeting friends of friends (called “triadic closure”, since befriending the friend of a friend means “closing a triangle” in the social network). Iniguez, assistant professor in the Department of Network and Data Science at CEU, designed a very simple mathematical model to describe how homophily and triadic closure can shape the structure of society, and then compared it with real social networks to measure how much homophily and triadic closure are actually needed for them to have the shape that they do.
An In-Built Tendency for Similarity Bias
The main finding of the study is that even a small preference for similarity when choosing new friends - together with the tendency to befriend friends of friends - can lead to an “amplification of homophily”. This means extremely polarized social networks composed of groups of similar people, but where different groups don’t have too many connections between them. “This is not ideal since it seems that our in-built social tendency to close triangles pushes us towards polarized societies”, explains Iniguez. “Our model - and its validation with data from real social networks - also shows that the dynamic combination of homophily and triadic closure can lead to what social scientists call a ‘core-periphery’ structure.”
The model shows a social network where a group of people (the “core”) dominates most connections in the network, and members in the other group (the “periphery”) are almost not connected between themselves. Iniguez’s modelling illustrates how a natural human tendency is to seek similarity and close social triangles, which in turn can lead us to societies where, for example, a minority might be disenfranchised in the sense that its members don’t interact as much between themselves as with the majority.
Perceiving Society Through our Social Brain
Studies such as this are crucial for understanding the mechanisms and processes that lead to the types of societies we live in, both off- and online. We may think that we’re quite tolerant about meeting people different to us, and that in general we’re open to making acquaintances with almost anyone, but as Iniguez’s model shows, even small biases in our decisions can accumulate, multiply in complex social feedbacks, and lead to societal effects such as extreme polarization, filter bubbles, echo chambers, and so on. The CEU professor points out that: “Scientifically, our results highlight the importance of treating social mechanisms together within dynamic mathematical models in order to understand how society actually evolves. To everyone else, it shows how tolerance to people that are different from us is extremely important, especially in crisis situations such as the current pandemic, since the way our social brain normally behaves might naturally lead to polarized and intolerant societies.”
The aim of the study was to show that, even in a simplified scenario where there is nothing apart from a tendency to seek similarity and close triangles, this similarity bias may drive the whole of society to polarization, showing how important it is to recognize feedback loops and dynamic processes when studying society. “In the future, we plan to refine our results, making the model more realistic, improving its validation with data from more social networks, and think about how these types of modelling efforts could help inform public policy and intervention strategies to decrease polarization in offline and online societies,” Iniguez concludes.
“Cumulative Effects of Triadic Closure and Homophily in Social Networks” by Gerardo Iniguez, assistant professor in the Department of Network and Data Science at CEU and his co-authors, is published in Science Advances.