A one-week course on: Latent Variable Modelling and Structural Equation Modelling for Social Science

Description:

London School of Economics and Political Science:

Methods Summer Programme

A one-week course on: Latent Variable Modelling and Structural Equation Modelling for Social Sciences Research

15-19 August 2016

Instructor: Professor Irini Moustaki

To apply and further information

http://www.lse.ac.uk/study/summerSchools/Methods/Quantitative/home.aspx

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:

a) modern statistical methodology for analysing multivariate continuous and categorical data

b) 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:

a) 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

b) modelling cross-sectional data including (as special topics) the treatment of longitudinal data and multi-group data such as cross-national data

c) 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.

Outline

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 9:30am-1pm each day. Computer practical classes take place in the afternoon. An optional project based examination is available.

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

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.

Cost:

Check the website

Website and registration:

http://www.lse.ac.uk/study/summerSchools/Methods/Quantitative/home.aspx

Region:

Greater London

Keywords:

Quantitative Data Handling and Data Analysis, Latent Variable Models, Latent trait analysis, Factor analysis, Confirmatory factor analysis, Structural equation models, Rasch models, Item response theory

Related publications and presentations:

Quantitative Data Handling and Data Analysis
Latent Variable Models
Latent trait analysis
Factor analysis
Confirmatory factor analysis
Structural equation models
Rasch models
Item response theory

Back to archive...