Agent-based null models for examining experimental social interaction networks.

We consider the analysis of temporal data arising from online interactive social experiments, which is complicated by the fact that classical independence assumptions about the observations are not satisfied. Therefore, we propose an approach that compares the output of a fitted (linear) model from the observed interaction data to that generated by an assumed agent-based null model. This allows us to discover, for example, the extent to which the structure of social interactions differs from that of random interactions. Moreover, we provide network visualisations that identify the extent of ingroup favouritism and reciprocity as well as particular individuals whose behaviour differs markedly from the norm. We specifically consider experimental data collected via the novel Virtual Interaction APPLication (VIAPPL). We find that ingroup favouritism and reciprocity are present in social interactions observed on this platform, and that these behaviours strengthen over time. Note that, while our proposed methodology was developed with VIAPPL in mind, its potential usage extends to any type of social interaction data.

Generalized mean-field approximation for the Deffuant opinion dynamics model on networks.

When the interactions of agents on a network are assumed to follow the Deffuant opinion dynamics model, the outcomes are known to depend on the structure of the underlying network. This behavior cannot be captured by existing mean-field approximations for the Deffuant model. In this paper, a generalized mean-field approximation is derived that accounts for the effects of network topology on Deffuant dynamics through the degree distribution or community structure of the network. The accuracy of the approximation is examined by comparison with large-scale Monte Carlo simulations on both synthetic and real-world networks.