Studying the Emergence of Social Norms

Social norms emerge from the dynamics of individual social interactions. Peter Froncek, a Ph.D. student in Computational Social Science at GMU, is investigating how this happens, how they are adopted, and the process used in various network topologies. He shared his early findings in a presentation in the Friday Seminar Series in the CSS department on April 3, 2015 in a talk titled, “A Community Approach to Norms on Various Spatial Topologies”.

“Social science should strive for explanations in terms of realistic, empirically observed social mechanisms
“ — Hedstrom 2005, (quoted by Froncek)

Froncek built a model to test various network topologies to try to replicate a more realistic version of the network diffusion of norms. This is because not all networks are equal. Below are the three network types he worked with in his model.

Image from Froncek’s presentation.

The model started with a base that was built off of the work of Robert Axelrod. In 1986, Axelrod developed a norms games, which Froncek described as, “an agent-based model which sought to explain the individual-level processes which lead to the establishment or collapse of a norm.”

Using the Axelrod model as the foundation and testing norm emergence on three networks, Froncek applied the following model specifications…

• ‘Agents interact ONLY with their neighbors.’
• ‘The payoff matrix is the same as Axelrod’s original model.’
• Probability of behavior being seen is a function of the number of neighbors.
• ‘Every four rounds each agent randomly selects a behavior set from the set of neighbors whose scores are at least one SD above the neighborhood average’
• ‘The payoff matrix is the same as Axelrod’s original model.’

According to Froncek’s early results, the model shows the value of subversive behavior in society, and that a highly vigilant society does not necessarily result in a moral society.

To sum up the results, Froncek said…

“You are imitating your friends that are doing well, but you are not perfect at imitating them, so there might be small differences…” — Froncek