Structural equation modeling (SEM) is a statistical methodology that is gaining popularity with researchers from a broad range of fields including sociology, psychology, clinical sciences, ecology, and biology. SEM provides a general framework for examining complex relationships between variables that may be either observed (manifest) or unobserved (latent). SEM can be used to investigate direct and indirect effects of manifest and latent variables, model correlated exogenous variables, compare relationships among groups, and analyze longitudinal data. Path analysis and linear models are both special cases of SEMs.

This course will cover the fundamental theory behind SEM and provide attendees with the knowledge and skills necessary to confidently apply these methods to their own research. No previous background in SEM is necessary. Researchers from all fields are welcome.

Specific topics to be covered include:

• Overview of SEM methods
• Review of linear models
• Path analysis
• Direct and indirect effects
• Exploratory factor analysis
• Confirmatory factor analysis
• Goodness of fit measures
• Reliability and validity
• Regression with latent variables
• Group analyses
• Multilevel SEM
• Modeling non-normal data