Friday, November 22, 2013
(3-4pm, Yost 306)
Title: Generalized Linear and Bilinear Mixed-Effects Models for Affiliation Networks
Speaker: Kate Calder
(The Ohio State University)
An affiliation network is a special kind of two-mode social network that consists of a set of actors and a set of events where ties indicate an actor's participation in an event. Statistical methods for studying affiliation networks are less developed than methods for studying one-mode, or actor-actor, networks. One way to analyze affiliation networks is to consider one-mode network matrices which are derived from the affiliation network, an approach known as the conversion method. We illustrate this approach through an analysis activity pattern data collected as part of a large neighborhood-based survey of households in Los Angeles County . We employ a multilevel p2 network model to describe patterns of segregation in individuals' activity spaces (locations of regular activities) and to quantify the effects of this type of segregation on adolescent behavioral outcomes. Since affiliation networks are defined on subsets of actors and events, the conversion method loses important structural features of the data. In order to model both modes in an affiliation network simultaneously, we also describe a novel bilinear mixed effects models that allows for third-order dependence patterns in the interactions between actors and events.
This is joint research with Christopher R. Browning, Yanan Jia, and Samuel Bussman.