Learning in Social Networks: Selecting Profitable Choices among Alternatives of Uncertain Profitability in Varying Networks

Abstract

We model an individual decision problem in the context of social networks: individuals have to make choices that yield uncertain outcomes, while on average some choices are better than others in varying networks. We generate computer-simulated data to derive hypotheses about the effect of network positions and network structures on making the most profitable choices. Thereafter, we establish an experimental design in the same parameter space as our computer simulation to empirically test these hypotheses. From these experimental data we see that on the micro-level, degree and closeness centrality have a positive effect on making profitable choices, while betweenness centrality has no effect. On the macro level, we see that density has a positive effect on making profitable choices, while we did not find an effect of centralization.