Pavel N. Krivitsky
Dr Krivitsky develops methods and models for collecting and analysing complex network data and processes with applications in epidemiology and the social sciences. He is particularly interested in unusual or indirectly observed or sampled network data. His work has a strong computational component, and I develop and maintain a number of popular R packages on CRAN.
Sessions
This workshop provides a hands-on tutorial to using exponential-family random graph models (ERGMs) for statistical analysis of social networks, using the {ergm} in R. The {ergm} package provides tools for the specification, estimation, assessment and simulation of ERGMs that incorporate the complex dependencies within networks. Topics covered in this workshop include:
* an overview of the ERGM framework;
* types of terms used in ERGMs
* defining and fitting models to empirical data;
* interpreting model coefficients;
* goodness-of-fit and model adequacy checking;
* simulation of networks using fitted ERG models;
* degeneracy assessment and avoidance.
{ergm} is a part of the Statnet suite of R packages for network analysis.
Individuals’ behaviour and attitudes are shaped by their perceptions of the social structure (i.e., cognitive social structures), yet little is known about the structural mechanisms underlying these perceptions and their interdependence. In this paper, we analyse adolescents’ perceptions of peers’ friendship networks in class to explore the structure and predictors of individuals’ social network perceptions. To do so, we extend multilayer exponential random graph models (ERGMs) to CSS data, also modelling the dependencies between perceptions. Cognitive social structures were modelled as interdependent layers in 15 Dutch high school classes (for a total of 346 students). The results show that adolescents’ perceptions exhibit patterns consistent with opportunity, balance, and homophily: friendships were more likely perceived between peers with reciprocated self-reported ties and between peers with the same gender or ethnic background. Additionally, adolescents perceived transitivity (a friend of a friend is a friend) in others' friendship networks. Contrary to expectations, cross-layer effects showed that friends did not hold more similar perceptions than non-friends. Neither did similarities in demographic characteristics (e.g., gender and ethnicity) shape similarities in perceptions. By extending multilayer ERGMs to cognitive social structures, this study provides a novel methodological approach for analysing cognitive social structures and new insights into how adolescents perceive social networks.