Structural equations modelling (SEM) is used by researchers in a wide range of fields, including biology, business and economics, education, health sciences, psychology and sociology, etc. SEM subsumes many other analytical techniques as special cases and is particularly suited to accommodating latent variables. This course provides an introduction to SEM and offers instruction in the LISREL program.
 
The course is divided into two parts. Part I provides an overview of the fundamentals of SEM, including the use of path diagrams and LISREL notation. This is accompanied by an introduction to the use of LISREL. Part II of the course focuses on three types of structural equation models, namely: path analysis models (with observed variables), one-factor measurement models and confirmatory factor analysis, and structural equations models with latent variables.
Several issues related to fitting structural equations models are addressed throughout the course. These issues include: model specification, identification and estimation; assessing model fit and interpreting goodness-of-fit criteria; assessing measure validity and reliability; treatment of missing data; modelling continuous and categorical data; recognising equivalent models; and multi-sample analyses.
 
Course participants will be given an opportunity on the final day of the course to analyse their own data. They are encouraged to bring data with them, although other data sets will be made available where required. Participants wishing to bring their own data should bring it in one of the following formats: an SPSS (*.sav) file; a text file (either fixed-width or delimited); an Excel (*.xls) file; or an Access (*.mdb) file.
 
After completing this course participants should understand the fundamentals of SEM and be competent in the use of LISREL for estimating common SEM models. They should be well prepared for learning other SEM software, including AMOS, EQS, and MPlus. Finally, the course is a useful prerequisite for advanced SEM topics, such as latent mean analysis, latent class analysis, and multi-level SEM, etc.