The Relative Agreement Model of Opinion Dynamics in Populations with Complex Social Network Structure
Research in the field of Opinion Dynamics studies how the distribution of opinions in a population of agents alters over time, as a consequence of interactions among the agents and/or in response to external influencing factors. One of most widely-cited items of literature in this field is a paper by Deffuant et al. (2002) introducing the Relative Agreement (RA) model of opinion dynamics, published in the Journal of Artificial Societies and Social Simulation (JASSS). In a recent paper published in the same journal, Meadows & Cliff (2012) questioned some of the results published in Deffuant et al. (2002) and released public-domain source-code for implementations of the RA model that follow the description given by Deffuant et al., but which generate results with some significant qualitative differences. The results published by Meadows & Cliff (2012) follow the methodology of Deffuant et al. (2002), simulating a population of agents in which, in principle, any agent can influence the opinion of any other agent. That is, the ”social network” of the simulated agents is a fully connected graph. In contrast, in this paper we report on the results from a series of empirical experiments where the social network of the agent population is non-trivially structured, using the stochastic network construction algorithm introduced by Klemm & Eguiluz (2002), which allows variation of a single real-valued control parameter μ KE over the range [0,1], and which produces network topologies that have ”small world” characteristics when μ KE =0 and that have ”scale free” characteristics when μ KE =1. Using the public-domain source-code published in JASSS, we have studied the response of the dynamics of the RA model of opinion dynamics in populations of agents with complex social networks. Our primary findings are that variations in the clustering coefficient of the social networks (induced by altering μ KE ) have a significant effect on the opinion dynamics of the RA model operating on those networks, whereas the effect of variations in average shortest path length is much weaker.
Unable to display preview. Download preview PDF.
- Amblard, F., Deffuant, G.: The role of network topology on extremism propagation with the Relative Agreement opinion dynamics. Physica A 343, 725–738 (2004)Google Scholar
- Brehm, S., Brehm, J.: Psychological Reactance: a theory of freedom and control. Academic Press, New York (1981)Google Scholar
- Deffuant, G., Amblard, F., Weisbuch, G., Faure, T.: How can extremism prevail? J. Artificial Societies and Social Simulation 5(4), 1 (2002)Google Scholar
- Deffuant, G.: Comparing Extremism Propagation Patterns in Continuous Opinion Models. Journal of Artificial Societies and Social Simulation 9(3), 8 (2006)Google Scholar
- Hegselmann, R., Krause, U.: Opinion dynamics and bounded confidence: models, analysis and simulation. J. Artificial Societies and Social Simulation 5(3), 2 (2002)Google Scholar
- Klemm, K., Eguíluz, V.: Growing Scale-Free Networks with Small-World Behavior. Physical Review E 65, 057102 (2002)Google Scholar
- Krause, U.: A Discrete Nonlinear and Non-Autonomous Model of Consensus Formation. In: Proc. 4th Int. Conf. on Difference Equations, August 27-31, pp. 227–236 (1998-2000)Google Scholar
- Lorenz, J., Deffuant, G.: The role of network topology on extremism propagation with the relative agreement opinion dynamics. Physica A 343, 725–738 (2005)Google Scholar
- Meadows, M., Cliff, D.: Reexamining the Relative Agreement Model of Opinion Dynamics. Journal of Artificial Societies and Social Simulation 15(4), 4 (2012)Google Scholar
- Meadows, M., Cliff, D.: The Relative Agreement model of opinion dynamics in populations with complex social network structure (forthcoming, 2013) (manuscript) Google Scholar