Game-theoretic models of network formation typically assume that people create relations so as to maximize their own outcome in the network. Recent experiments on network formation suggest that the assumption of self-interest might be unwarranted and that social preferences, such as altruism and inequality aversion, play a role in the formation of social networks. We developed an experiment to systematically investigate whether people show preferences for outcomes of others during network formation. We find that such preferences play a role when network decisions degenerate to simple two-person decision tasks. In more complex environments, however, we find little evidence for social preferences as a significant decision criterion. Furthermore, we find some evidence for farsighted behavior in network formation.