Analysing social influence with cross-sectional network data

Date:

31/10/2024 - 01/11/2024

Organised by:

University of Glasgow (an NCRM Centre Partner)

Presenter:

Johan Koskinen

Level:

Intermediate (some prior knowledge)

Contact:

Penny White
NCRM Centre Manager
p.c.white@southampton.ac.uk

Map:

View in Google Maps  (G12 8TB)

Venue:

School of Health and Wellbeing
University of Glasgow
Room 418a, Clarice Pears Building
90 Byres Road
Glasgow

Description:

This in-person course will introduce the analysis of individual-level outcomes for individuals that are connected through links in a social network. We will cover the basic statistical issues associated with analysing inter-dependent outcomes and then study models that specialise in inferring so-called social influence in cross-sectional network data.

We will have a particular focus on a class of models for binary outcome variables called the auto-logistic actor attribute models (ALAAMs). Emphasis will be placed on practical hands-on exercises of real datasets in R. The course will run lunch-to-lunch over two days. The first afternoon providing the necessary background and provide the participants with the skills of applying and interpreting basic models. The second morning will present more advanced topics and deeper insight into estimation and interpretation issues.

The course covers: 

  • Basic network operations in R and network regression models. Network plots and metrics; the network effects model; generalised linear mixed model

  • ALAAM fundamentals: Formatting data; estimating the independence model; estimating simple/direct contagion; tuning MCMC

  • ALAAM model fit and handling missing data: Bayesian GOF; missing values in response; more on tuning MCMC

  • More advanced model specifications for ALAAM and extensions: Indirect contagion; accounting for heterogeneity; Posterior deviance and DIC; Hierarchical ALAAM

By the end of the course participants will:

  • Have a broad overview of methods for analysing nodal outcomes in social networks
  • Have a good understanding of the basic assumptions, implementation, and interpretation of ALAAM
  • Be able to estimate ALAAM and diagnose fit of the model
  • Have a working knowledge of advanced ALAAM specifications

Audience:

This course is relevant to anyone working with social network data and in particularly researchers interested in analysing the effects of networks on health, education, and other outcomes. The course will provide the tools for analysing outcomes with ALAAM but more generally provides insights into modern statistical methods for networks and transferable skills in the areas of network analysis and Bayesian inference. Network research is not discipline specific and the tools are widely applicable. 

Pre-requisites:

Basic skills in regression modelling and knowledge of fundamental social network concepts and techniques. In addition, as the course practicals will be done exclusively in R, participants will find that basic knowledge of R is useful. Necessary R-programs can be installed from CRAN or a course-specific, designated GitHub repository for the course. Demonstration will be done in RMarkdown and you may find it useful to follow the course in RStudio.

Schedule:

This course is spread across 2 half days and equates to 1 day for payment purposes

Day 1 - Thursday 31st October - 12:30 - 17:00

Day 2 - Friday 1st November - 09:00 - 13:00

Presenter:

Johan Koskinen is a Lecturer in Statistics at Stockholm University with years of experience in developing and training people in the use of statistical network models. With colleagues at the University of Melbourne he wrote the standard reference for exponential random graph models that was published by CUP and awarded the Harrison White book prize. Johan did his PhD in Statistics at Stockholm University and after that worked at the Universities of Melbourne, Oxford, and Manchester. He has published wildly on modelling and inference for different types of network data. He has also developed methods in close collaboration with applied researchers in a range of different disciplines. The core of his research focuses on Bayesian inference and he has demonstrated its benefits in handling longitudinal data and sampled and missing network data.

Cost:

The fee per teaching day is £35 per day for students / £75 per day for staff working for academic institutions, Research Councils and other recognised research institutions, registered charity organisations and the public sector / £250 per day for all other participants.

In the event of cancellation by the delegate a full refund of the course fee is available up to two weeks prior to the course. NO refunds are available after this date. If it is no longer possible to run a course due to circumstances beyond its control, NCRM reserves the right to cancel the course at its sole discretion at any time prior to the event. In this event every effort will be made to reschedule the course. If this is not possible or the new date is inconvenient a full refund of the course fee will be given. NCRM shall not be liable for any costs, losses or expenses that may be incurred as a result of its cancellation of a course, including but not limited to any travel or accommodation costs. The University of Southampton’s Online Store T&Cs also continue to apply.


Website and registration:

Register for this course

Region:

Scotland

Keywords:

Bayesian methods, Generalized liner model (GLM), Network analysis, Network autocorrelation, Social influence


Related publications and presentations from our eprints archive:

Bayesian methods
Generalized liner model (GLM)
Network analysis

Back to the training database