Workshops
The following workshops will take place on 29 November 2023, in two sessions, AM and PM.
Each workshop is $75 for regular attendees and $45 for students, to be paid at registration.
Schedule
AM
9:00 – 12:00
ERGM Analysis for Multilevel Networks with MPNet — CANCELLED
Instructor: Peng Wang (Swinburne University)
Description: In this hands-on workshop, participants will learn the fundamentals of estimating Exponential Random Graph Models (ERGMs) and Auto-Logistic Actor Attribute Models (ALAAM) with MPNet – a software developed to investigate the structural features of networks and how such structure may affect individual outcomes.
The workshop will start with a brief introduction to the overall logic of estimating (single-level) ERGMs/ALAAMs before introducing the recently developed multilevel ERGMs/ALAAMs. The latter class of models enables researchers to investigate the influence of structure at one level of analysis on structure at a different level, while taking into account the complex interdependencies that exist within and between levels. For instance, interpersonal networks between managers at the micro-level might interact with alliance networks of the organizations they are nested in. The workshop will also demonstrate how to user MPNet to model temporal network dynamics using Temporal ERGMs.
Throughout the workshop, participants will work through short exercises to get familiar with the graphical user interface and output of the MPNet software. We will discuss various case-study examples that will provide the participants with a good understanding of the possibilities that multilevel ERGMs offer for social scientists.
Prerequisites: Some basic familiarity with social network analysis will be helpful. Participants are required to bring their own laptops with MPNet installed. Note that MPNet is not compatible with Mac OS without a compatible Windows parallel.
Survey and Interview Methods in Social and Personal Network Research (Egonets)
Instructor: Malcolm Alexander (Australian Consortium for Social and Political Research Incorporated) and Rebecca Langdon (University of Queensland)
Description: Researchers in psychology, sociology and cognate disciplines use surveys and interviews to study social capital, social support and the sociology of social networks. They face crucial choices in the framing of research questions, strategies of data collection (e.g., Qualtrics), data management and analysis and strategies of publication. This workshop explores these issues and reviews the varied software options for each aspect of this research. The presenters approach this research from business/psychology, survey methods and quantitative methods (Rebecca Langdon) and from sociology, in-depth interviewing and social network analysis (Malcolm Alexander). Example datasets are drawn from both these approaches.
Prerequisites: Familiarity with setting up basic surveys in Qualtrics and knowledge of social network analyses and basic statistical procedures (e.g., ANOVA, t-test, correlation, regression) assumed.
Introduction to ERGMs with Statnet
Instructor: Pavel Krivitsky (University of New South Wales)
Description: This workshop provides a hands-on tutorial to using exponential-family random graph models (ERGMs) for statistical analysis of social networks, using the 'ergm' package in Statnet (https://statnet.org). 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.
Prerequisites: Familiarity with R and familiarity with basic concepts of network analysis.
PM
1:00 – 4:00
Collecting and Analysing Online Networks with VOSON R tools
Instructor: Robert Ackland (Australian National University)
Description: This workshop will introduce participants to open source R packages for online network collection and analysis, developed by the Virtual Observatory for the Study of Online Networks (VOSON) Lab (http://vosonlab.net) at the Australian National University. The workshop will include an introduction to (depending on workshop participant interest and available API access):
vosonSML (https://github.com/vosonlab/vosonSML) - an R package providing a suite of tools for collecting and constructing networks from social media data. It provides easy-to-use functions for collecting data across popular platforms (Twitter, Reddit, YouTube, WWW hyperlinks) and generating different types of networks for analysis.
VOSON Dashboard (https://github.com/vosonlab/VOSONDash) - an R/Shiny application providing a graphical user interface for collecting and analysing online networks and associated text data. It builds on a number of R packages, in particular igraph (for network analysis) and vosonSML.
voson.tcn (https://github.com/vosonlab/voson.tcn) – an R package for collection and analysis of Twitter conversation networks. This package uses the Twitter API v2 Early Access endpoints to collect tweets and generate networks for threaded conversations identified using the new tweet conversation identifier.
Collection and analysis of other social media data sources such as the peer-to-peer microblog site Mastodon.
Participants will be given instructions on how to install these packages prior to the workshop. Workshop materials will include R scripts, package documentation, notes on analysis of online networks, and examples of research.
Prerequisites: It is expected that participants will have some experience using R. R and RStudio will need to be installed on laptop prior to workshop. VOSON tools (R packages) installed prior to workshop (specifics of packages and versions will be given to participants in advance of the workshop).
Advanced Exponential-Family Random Graph Modelling with Statnet
Instructor: Pavel Krivitsky (University of New South Wales)
Description: This workshop will provide a tutorial of advanced usage of 'ergm' and extension packages, focusing on binary networks. Topics include specifying complex structural constraints, estimation tuning, representing complex effects with term operators, and observational (e.g., missing data) structure. Also included is using the new 'ergm.multi' package for modelling multi-layer and multi-mode networks, as well as joint models for ensembles of networks.
Prerequisites: Familiarity with R and 'ergm' required. If you are new to ERGMs, the introductory workshop on ERGMs using statnet is strongly suggested.
Estimating Auto-Logistic Actor Attribute Models in R — CANCELLED
Instructor: Johan Koskinen (Stockholm University)
Description: The auto-logistic actor attribute model (ALAAM) is a model for cross-sectional binary outcomes where dependencies through a binary network is accounted for. In its simplest form, ALAAM is a logistic regression on network summaries and other exogenous covariates. The strength of ALAAM is however the ability to model dependence of nodal outcomes through network tie, for example, are two people that are relationally tied more likely to share the same opinion or belief. As such, ALAAM allows us to model outcomes that are concordant with social influence and social contagion processes. ALAAM can be estimated in the stand-alone statistical software package MPNet but we will focus on routines for estimating ALAAM in R using a Bayesian inference framework. We will cover the entire process from reading in data into R to formatting the data for analysis and then analysis and interpretation. We will touch on common considerations such as missing data and model selection. Most of the workshop will focus on hands-on work on analysing data but some theoretical and conceptual issues will also be covered. The framework is described in detail in
Koskinen & Daraganova (2022). Bayesian Analysis of Social Influence. JRSSA https://doi.org/10.1111/rssa.12844
Prerequisites: We assume that you have a good knowledge of social network analysis and a keen interest in networks. Fundamental network concepts and theories will be assumed knowledge as will standard network techniques. Network techniques include basic graph theory, centrality, and clustering measures. It is recommended that you have some experience of using standard statistical models, such as regression models, and it is advantageous to be familiar with binary response models such as logistic regression. You will find it useful to have a working knowledge of the program R.