Prerequisites
An understanding of multiple regression is essential and knowledge of exploratory factor analysis is highly desirable. It is also expected that participants have previous experience with a statistical data analysis package such as SPSS. However, the course assumes that participants have little or no prior knowledge of structural equation modelling (SEM).
Course Outline
The course is designed as an introductory, applied course in the use of SEM using the Amos18 program. SEM is used widely by researchers to find and test complex relationships amongst observed (measured) variables and latent (unobserved) variables and amongst the latent variables themselves. SEM subsumes other analytical techniques such as regression, path analysis, factor analysis, and canonical correlation.
The course is divided into three parts and follows the historical development of SEM. Part I is concerned with directly observed variables. Topics include regression and path analysis, and model testing of causal models with observed variables. Part II of the course introduces latent (unobserved) variables. Topics include confirmatory factor analysis (CFA), model comparisons, model equivalence and higher-order CFA, formative versus reflective indicators, and invariance testing via multi-group analyses. Part III combines the ideas covered in the first two parts by introducing full SEM models as path models amongst latent variables. Topics include single indicator latent variable models, and mediating, potentially moderating (interaction) and alternate models in SEM. Throughout the course issues related to fitting structural models are addressed. These include: model specification, identification and estimation, assessing model fit (goodness-of-fit criteria), and dealing with problem data including missing data, small samples, ordinal and/or dichotomous variables, non-normal data, non-positive definite matrices and inadmissible models.
Course participants will be provided with instruction and practical experience in the use of Amos to estimate parameters implied by various types of models. On the final day of the course, participants will be given the opportunity to analyse their own data and receive feedback on their data and/or their models. Participants wishing to bring their own data should bring an SPSS (*.sav) file or an Excel (*.xls) file.
While no pre-reading in SEM is required and all necessary materials will be provided for the course, interested participants may wish to consult the following reference as preliminary reading:
Kline, R. B. (2005). Principles and Practice of Structural Equation Modeling (2nd Ed.). NY: The Guilford Press.
Course Text
Cunningham, E. (2009). A practical guide to structural equation modelling using Amos™. Melbourne: Statsline.
This text of the instructor's course notes will be distributed to all course participants.
Related Courses
Participants with a good understanding of this course may progress to further courses (when offered) in SEM using either the Amos™ or Mplus™ software packages.
