Statistical Models for Interaction Dynamics

Carter T. Butts
Departments of Sociology, Statistics, and EECS
and
Institute for Mathematical Behavioral Sciences
University of California, Irvine

Patterns of interaction among individuals, organizations, or other entities are often the result of a complex interplay of endogenous behavioral mechanisms and contextual factors.  Given observational data on such interactions, how can we identify the specific mechanisms responsible for the dynamics?  Likewise, how can we translate candidate mechanisms from behavioral theories into estimable (and ultimately testable) models for social interaction, particularly when many different mechanisms may be at work?  In this talk, I will describe one approach to addressing these questions, and discuss how this approach has been used to model social interaction in settings ranging from emergency communication and classroom interaction to email and gang violence.  I will also point to some ongoing challenges in this area, and opportunities for further applications and development.