Online Program

Friday, February 19
CS11 Multivariate Analytic Methods Fri, Feb 19, 2:00 PM - 3:30 PM
Diamond I&II

Developing, Testing, and Comparing Theories Using Structural Equation Modeling (303184)

*Darius Singpurwalla, University of Maryland 

Keywords: Structural Equation Modeling, Modeling, Analysis, SAS

Theories of human behavior, health, and economics are complex topics that involve different patterns of relationships, variables, and groups. Thus, researchers must use statistical techniques that can account for the complexities required to model these phenomenon. Structural equation modeling (SEM) is one such tool. Researchers can use SEM to help develop and test different theories and compare competing theories. This talk will demonstrate, in the context of a theoretical model for exercise behavior, the benefits of using SEM. It will present the motivation and theoretical basis for the model and demonstrate how SEM was used to select the final model. Participants who attend the talk will benefit from the overview of SEM and how it can help identify relationships between latent and manifest variables. While the data set used to demonstrate the techniques will focus on physical activity and associated psychological constructs, SEM is applicable to all types of data sets and industries interested in understanding relationships between variables.