Friends by choice. An actor-oriented statistical network model for friendship networks through time

It is common knowledge that once a group of individuals has the opportunity to interact for a certain period of time, friendly relation¬ships and friendships will be established and, as a result, an affective relation network will arise. How the structure of this network comes about and how it develops over time are the questions addressed in this study. To answer them, actor-oriented statistical network models are proposed. The formalization of the models depends on the context in which the networks arise and develop. Two situations are distinguished. First, the development of an affective relation network among a closed group of students, applied to a group of Dutch university freshmen. The context is relatively uncomplicated, in the sense that interaction takes place under rather homogeneous conditions. This general model includes, among others, parameters for the effects of reciprocity, similarity and the opportunity structure. Second, the development of an affective relation network within a work organization, a context in which interaction occurs under more complicated and heterogeneous circum¬stances. Such features as the hierarchical organization structure, interpersonal competition, and cohorts are incorporated in the model. To substantiate the model, it is applied to longitudinal network data collected in two departments of a Dutch hospital, namely a dialysis and a nursing department.