ME415 Latent Variable Modelling and Structural Equation Modelling for Social Sciences Research

Applications open January 2017

This course aims to provide participants with an introduction to latent variables and structural equation models for both continuous and categorical data and their use in measurement and in modelling complex substantive hypothesis in the social sciences.

It provides a balance between methods and applications to enable participants to develop a good understanding of structural equation models and related methods.

This course is suitable for postgraduate and academic staff in applied statistics, medicine, and in social and behavioural sciences as well as government employees and people working in marketing, management, public health and banking.

Course benefits
This course provides participants with introductions to:

• modern statistical methodology for analysing multivariate continuous and categorical data
• the use of latent variables for measuring unobserved constructs such as attitudes, beliefs, health state, etc. through observed indicators.

It will also provide participants with experience in:

• the use of path models to represent complex relationships among latent and unobserved variables
• modelling relationships among latent constructs and observed covariates and providing the ability to measure direct and indirect (mediation) effects
• modelling cross-sectional data including (as special topics) the treatment of longitudinal data and multi-group data such as cross-national data
• the use of STATA and MPlus software to run real data examples. Participants will have the opportunity to run the analysis and interpret the output.

Prerequisites
Knowledge of applied regression analysis is required. No previous knowledge of latent variable analysis, structural equation modelling or of any particular software are required.

This course comprises a mixture of lectures on the theory and methodology of latent variable models and structural equation models, followed by practical sessions in which participants will have the opportunity to use STATA and MPlus to run analysis on real data sets.

The topics covered will include: an introduction to latent variables, path analysis and structural equation modelling with latent and observed variables, exploratory and confirmatory factor analysis, latent trait analysis, latent class analysis, structural equation models for continuous and categorical observed variables, cross-sectional, longitudinal data and analysis of multi-group (cross-national survey) data.

A step-by-step approach will be taken to introduce all topics so that the course will build up from introductory to more advanced material. 

Participants will receive lecture slides and material for the practical sessions that will include the STATA and MPlus commands.

Lectures take place from approx 10am-1pm each day. Computer practical classes take place in the afternoon. An optional 2-hour exam will take place on the final afternoon of the course.

Main texts
Bartholomew, D.J., Steele, F., Moustaki, I. and Galbraith, J. (2008) Analysis of Multivariate Data for Social Scientists [2nd edition]. Chapman and Hall/CRC.

Software used
STATA and MPlus

Professor Irini Moustaki

Irini Moustaki is a Professor in Social Statistics in the Department of Statistics where she was awarded her PhD in 1996. She has taught for many years a postgraduate multivariate analysis course designed for statistics and social science students and run professional short courses in UK and abroad on the same topic. Her main research interests are latent variable models, structural equation models, analysis of categorical data, methods of estimation, goodness-of-fit, detection of outliers and missing values. She has co-authored books on latent variable models and multivariate data analysis, published many methodological and applied papers and served as an associate editor and editorial member for a number of journals including Computational Statistics and Data Analysis and the Journal of Educational and Behavioural Statistics. She received an honorary doctorate from the University of Uppsala in January 2014 and elected the Executive Editor of Psychometrika in November 2014.

Please note: A full timetable will be provided at registration on Monday 14 August. The below timetable contains approximate hours only.

 

	M	Tu	W	Th	F
Morning lecture	3	3	3	3	3
Afternoon class	1.5	1.5	1.5	1.5	1.5