COMPUTATIONAL AND EXPERIMENTAL STUDIES ON NETWORK FORMATION

This dissertation aims to present, discuss, and propose new developments that address three gaps in the current network formation models utilized by social scientists, including sociologists, economists, and cognitive scientists. Here, we wanted to unveil and move beyond certain hidden assumptions and necessary simplifications that lie behind existing mathematical and computational models of network formation, which in our opinion undermine our ability to study generative mechanisms across various contexts with sufficient detail. To achieve this, we have employed a combination of techniques, including empirically-informed and validated agent-based models, theoretical N-player network formation games investigated through agent-based models, and laboratory experiments. Although our strategy was not to apply methodological triangulation straightforwardly, our ambition is that these examples can testify to the explanatory advancement offered by better integration of network analysis, agent-based models, and laboratory experiments. We believe that this integration can deepen our understanding of the micro generative mechanisms that preside over the emergence of network dynamics and outcomes in various socio-economic domains.