An identity model for network exchange

Network exchange theory has developed primarily as a static, structural theory of power and dependence in networks. I introduce a dynamic model of the exchange process in which network nodes are bused on a model of identity processes as given by identity theory. That is, I use the assumptions of identity theory to model the identity of a ''typical'' experimental subject whose primary goal is to participate in exchanges in an experimental paradigm. Computer simulations of the exchange process based on this identity model then generate predictions about the power advantage of particular nodes (actor's positions) relative to other nodes in a variety of networks. The resulting predictions correspond closely with results obtained in published experiments that have been interpreted as supporting network exchange theory. In addition, varying the characteristics of the identity model in the simulation reveal the interaction between individual (identity) and structural (network) characteristics in determining power and process in the network. I also show how a process model, like the one simulated, allows us to understand the exchange process over the long run, which can differ considerably from the usual short-term laboratory experiment.