Cultural traits — the information, beliefs, behaviors, customs, and practices that shape the character of a population — are influenced by conformity, the tendency to align with others, or anti-conformity, the choice to deliberately diverge. A new way to model this dynamic interplay could ultimately help explain societal phenomena like political polarization, cultural trends, and the spread of misinformation.
A study published in PNAS outlines this novel approach. Presenting a mathematical model, SFI Complexity Postdoctoral Fellow Kaleda Denton with colleagues at Stanford University — former SFI Post-baccalaureate Fellow Elisa Heinrich Mora, SFI External Professor Marcus Feldman, and Michael Palmer — expand on previous research to offer a more realistic representation of how conformist and anti-conformist biases shape the transmission of cultural traits through a population.
“The idea behind this research was to come up with a better way to mathematically represent how individuals make decisions in the real world,” says Denton. “If we can do that, we can then scale things up to see what would happen in a population of 10,000 people over the long run.”
Traditional models of conformity often assume individuals gravitate toward the average or “mean” trait of a population. This concept works well if the most popular traits in the population are near this mean, which may be the case for, say, working hours or food portion sizes. However, the mean is a poor indicator of popularity in other cases; for example, if most people fall on either the far left or far right of a political spectrum, but the mean lies in the center.
To address this gap, the authors designed a model that incorporates trait clustering. In this model, individuals conform by adopting traits that are more clustered together (e.g., variations of a far-left political belief) rather than the mean trait in the population (e.g., the centrist view). Anti-conformists, on the other hand, deliberately distance themselves from the traits of their peers, creating polarization.
Using computer simulations, the team analyzed how traits spread across populations over multiple generations. Conformity often led to groups clustering around specific traits, but not necessarily the average. Anti-conformity created a starkly different pattern: a U-shaped distribution, with individuals clustering at the extremes and leaving the middle sparsely populated.
One significant finding was that populations rarely converge to a single trait unless the unrealistic assumption of perfect behavioral copying is imposed. Instead, even small variations in how individuals interpret or adopt traits result in persistent diversity.
“These outcomes align with what we observe in the real world, where cultural practices and ideologies don’t simply average out but instead maintain significant variation,” Denton says.
The research also challenges the notion that conformity always leads to homogeneity. The model shows that under certain conditions, conformity can sustain diversity, while anti-conformity amplifies polarization.
Denton sees broad implications for the study. “This framework could help explain voting behavior, social-media trends, or even how people estimate values in group settings,” she says. “It offers a way to understand how individual decisions aggregate into societal patterns, whether that’s consensus-building or polarization.” This model can be tested on real-world data in future studies.
“We’re excited to see if this framework works in different scenarios,” Denton said. “The ultimate goal is to understand how individual choices influence entire populations over time.”
Read the paper "Conformity to continuous and discrete ordered traits" in PNAS (January 17, 20204) at https://doi.org/10.1073/pnas.2417078122
Read More
"Mathematical Framework Explains Evolutionary Dynamics of Behavior" (Siam News, June 2, 2025)