Testing two theories for generating signed networks using real data

Abstract

Multiple social processes generate social network structures. We use relaxed structural balance, a generalization of classic structural balance, to facilitate a direct comparative test of two social psychological theories regarding network generation. One is structural balance theory. The other concerns differential popularity. These theories predict distinctive signed blockmodels. We use two well known empirical temporal signed data sets presenting an opportunity for comparing the two theories in terms of their predictions about blockmodel representations of these networks. The results provide strong support for differential popularity, differential disliking, and mutual disliking within a subset of actors. While there is evidence that structural balance was also operating, it seems the lesser process for the data used in these tests. We also examine the unequal distributions of receiving positive and negative ties. Both tend to become more unequal over time. Suggestions for future research are provided.