Methodology and Statistics
International Conference, 16 - 18 September 2002
FDV, Ljubljana, Slovenia


ON THE EVOLUTION OF SIGNED SOCIAL NETWORKS

Patrick Doreian
University of Pittsburgh
USA

To the extent that social relations are signed - in contrast to being only positive - it is important to study them. In turn, the study of signed networks is helped by having sound substantive ideas, good measurement and relevant data analytical tools. At face value, we have all three. `Structural balance theory', in its various guises, provides the substantive foundation. Data have been collected for signed relations and we have a variety of flexible and useful tools for analyzing such signed network data. Some of these results are presented. Yet balance theory, measurement and signed blockmodeling tools are insufficient for establishing a general theory of balance and associated tools. In its empirical form, the core feature of structural balance theory is its focus on dynamics. Using some of the few through time data sets that exist, modest support for balance theory can be found. But there is even more contradictory evidence and this information raises a variety of important questions for the empirical study of balance dynamics. It is fruitful to pursue these questions. Some recent, and provisional, results are presented together with an agenda for the study of the evolution of signed networks.