Mitchell Centre Seminar Series, Alessandro Lomi (University of Italian Switzerland, Lugano) Modeling Social Networks with Changeable Nodes
| Dates: | 18 February 2026 |
|---|
| Times: | 16:00 - 17:30 |
|---|
| What is it: | Seminar |
|---|
| Organiser: | School of Social Sciences |
|---|
|
|
One of the defining features of social networks is that their nodes—social agents—are not fixed entities but actively reconfigure their internal structure in response to changes in their environment. When such reconfiguration occurs, widely held assumptions about mechanisms of network formation, such as homophily-based attachment, become difficult to interpret. This challenge is particularly evident when network nodes represent composite or collective actors, such as formal organizations, whose internal structure is itself subject to change. Treating node identity as fixed therefore limits our ability to understand how transformations in agents and transformations in network structure mutually shape one another over time.
This paper addresses this limitation by reframing stochastic actor-oriented models (SAOMs) for mixed-mode networks as a multinomial choice problem in which nodes select among alternative configurations of their internal structure while simultaneously constructing networks of external partners. Relaxing the assumption of stable node identity in this way preserves a coherent account of network dynamics while explicitly incorporating endogenous change in node composition. Once internal features of network nodes are allowed to change, however, standard interpretations of network autocorrelation—the tendency for ties to be patterned by shared, contextually relevant attributes—require careful reconsideration. This work introduces a new implementation of a widely adopted decomposition of network autocorrelation that explicitly accommodates shifts in the internal composition of nodes.
Using data on collaborative relations among healthcare organizations, The empirical analysis shows that this approach captures not only the observed interorganizational network structure, but also the evolving internal structures of organizations and field-level distributions of activities and resources. More broadly, the paper extends stochastic actor oriented models to settings in which node identities are endogenous, enabling the analysis of coevolving internal structures and network relations.
Travel and Contact Information
Find event
G7
Humanities Bridgeford Street
Manchester
