Co-evolution of two networks representing different social relations in NetSense.

We examine the dynamics of co-evolution of two coupled social networks. The first is a cognitive network defined by nominations based on perceived prominence collected from repeated surveys of students during their first four semesters of college while the second is built from the behavioral network representing actual interactions between these individuals based on records of their mobile calls and text messages. We address three interrelated questions. First, we ask whether the formation or dissolution of a link in one of the networks precedes or succeeds formation or dissolution of the corresponding link in the other network (temporal dependencies). Second, we explore the causes of observed temporal dependencies between the two networks. For those temporal dependencies that are confirmed, we measure the predictive capacity of such dependencies. Finally, we examine whether there are systematic differences in the dissolution rates of symmetric (undirected) versus asymmetric (directed) edges in both networks. We find strong patterns of reciprocal temporal dependencies between the two networks. In particular, the creation of an edge in the behavioral network generally precedes the formation of a corresponding edge in the cognitive network. Conversely, the decay of a link in the cognitive network generally precedes a decline in the intensity of communication in the behavioral network. Finally, asymmetric edges in the cognitive network have lower overall communication volume and more asymmetric communication flows in the behavioral network.