Advanced Multilevel Modelling using MCMC

Date:

16/09/2015 - 18/09/2015

Organised by:

NCRM, University of Southampton

Presenter:

Professor William Browne is Professor of Statistics at the Graduate School of Education, University of Bristol where he directs the Centre for Multilevel Modelling. His research interests are in statistical methodology, statistical software and application of statistics to complex datasets in many disciplines. He is one of the developers of MLwiN and Stat-JR and has worked at Bristol for 8 years. He has taught over 50 workshops on multilevel modelling and other advanced statistical modelling methods.

Dr George Leckie is a Senior Lecturer in Social Statistics at the Graduate School of Education, University of Bristol. His methodological interests are in the application and dissemination of multilevel and other latent variable models to analyse educational and social science data. His substantive interests include: school performance indicators and their associated publication in league tables; value-added models (VAMs) for measuring school and teacher effects on student achievement; and modelling rater effects on test scoring. He has taught multilevel modelling short courses across Europe, Australia and US.

Level:

Advanced (specialised prior knowledge)

Contact:

Jacqui Thorp
Training and Capacity Building Co-ordinator
University of Southampton
Email: jmh6@soton.ac.uk
Tel: 02380594069

Map:

View in Google Maps  (SO17 1BJ)

Venue:

Building 39, University of Southampton, Highfield, Southampton, Hampshire

Description:

This workshop will introduce the theory and application of Bayesian Markov chain Monte Carlo (MCMC) methods for multilevel modelling. We will focus on advanced multilevel models that benefit from MCMC estimation including: binary response models; cross-classified and multiple membership models; multivariate response models; and multiple imputation models for missing data. We will provide an overview of the main MCMC algorithms and popular MCMC diagnostics and graphical tools for assessing convergence. Theory sessions are accompanied by practical sessions using our dedicated multilevel software packages, MLwiN and Stat-JR; participants get the chance to apply what they have learned to real social, behavioural and medical datasets. Both our software packages are currently free for UK academics. We will showcase methods implemented in MLwiN to speed up MCMC estimation. We will also highlight software interoperability features within Stat-JR that allow users to call other MCMC packages such as WinBUGS, and then return the model results to Stat-JR.

The course covers:

  • MCMC methods for fitting multilevel models
  • A range of advanced multilevel models which benefit from MCMC estimation:
  • Binary response models
  • Cross-classified and multiple membership models
  • Multivariate response models
  • Multiple imputation models for missing data
  • Practical implementation of these models via MCMC methods in our multilevel software: MLwiN and Stat-JR

By the end of the course participants will:

  • Comprehend the basics of Bayesian statistics and MCMC methods for multilevel models
  • Understand a range of advanced multilevel models
  • Be able to fit and interpret multilevel models using MCMC methods in MLwiN and Stat-JR

The target audience is academic and other researchers from any discipline where multilevel modelling is required. The course is not introductory and so some previous knowledge of basic multilevel modelling is essential.

Cost:

The fee is:

• £90 for UK registered students
• £180 for staff at UK academic institutions, UK Research Councils researchers, UK public sector staff and staff at UK registered charity organisations and recognised UK research institutions.
• £660 for all other participants.

All fees include event materials, lunch, morning and afternoon tea. They do not include travel and accommodation costs.
A full refund is possible if cancelled 3 weeks prior to the course taking place. No refund is available after this.

Website and registration:

Region:

South West

Keywords:

Statistical Theory and Methods of Inference, Bayesian methods, Multilevel Modelling , Quantitative Software, MLwiN, MCMC Methods , Stat-JR

Related publications and presentations:

Statistical Theory and Methods of Inference
Bayesian methods
Multilevel Modelling
Quantitative Software
MLwiN

Back to archive...