The Broker's Invisible Ties : Brokerage and Balance in Signed Networks

Abstract: Brokerage is one of the most popular and well-studied concepts in the literature on social networks. It has so far been analysed, however, without taking negative ties in so-called signed networks into account. The central role of tension for brokerage motivates an integration with the hitherto not explicitly related social balance theory. Despite brokers benefitting from serving as intermediaries, in many networks, gaps remain open. We hypothesise that these gaps in some contexts are filled with conflictual relationships. Implementing balance theory, we argue that brokers face higher imbalance as they positively connect adversarial groups. This would create a trade-off between brokerage benefits and stress from unbalanced relationships. We additionally propose a novel theory on unbalanced four-cycles in which a strong desire for balance prohibits the evolution of positive ties between groups and stabilises brokerage. To answer the research questions, we used cross-sectional data from eight Hungarian work departments of sizes 16 to 43 on the relationships between colleagues. We applied descriptive statistics, non-parametric tests and linear regressions to test the hypotheses regarding the connection between brokerage and balance and their relevance for different benefits. Creating null models of the observed networks, we tested the prevalence of the specific four-cycles against the predicted number by random chance and correlated it with brokerage scores. As the last step, we created agent-based models simulating organisations with 25 agents to predict the impact of balance and brokerage mechanisms on the network evolution in connection with negative ties. Results show that brokers tend to have a higher share of negative ties in their neighbourhood but no higher imbalance. Despite heterogeneity in the networks and low statistical power, some support for the importance of balance for a broker concerning job satisfaction is found. For half of the networks, a significantly higher number of unbalanced four-cycles is observed which are predominantly connected to brokers. The simulation results show that more negative ties hamper average balance but do not foster brokerage properties. A higher motivation to obtain balance limits brokerage and vice versa. Overall, to create similar networks to the ones observed, a considerably high preference for balance must be integrated. In sum, more conflict in a network does not directly increase brokerage opportunities. Brokerage and balance both benefit actors and are not to be seen as contrary but complementary concepts. Thus, in workplaces, conflict does not need to be harmful if room exists to form balanced relationships. Central is the constellation of positive and negativeties that require the consideration of network structures at different levels.