Structural Equation Modeling: An Alternative to Predictive Modeling

• Activity Number: 141
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Social Statistics Section
• Author(s): An-Lin Cheng*+ and Patricia J. Kelly
• Companies: University of Missouri - Kansas City and University of Missouri - Kansas City
• Keywords: Structural Equation Modeling ; latent variable ; nursing

Abstract

Structural Equation Modeling (SEM) has a long history of use in social science research. This powerful analytical tool is able to model the observed and latent variables of a study. With current statistical software, SEM also has the ability to develop more complicated models, such as multilevel model, latent growth model and model with interaction effects. Some survey data sets, especially those in which latent variables are considered in the model, are more appropriately analyzed with SEM than predictive modeling. In this presentation, we will apply SEM technique to model factors affecting the retention of nurse practitioners in clinics serving low-income women, using a dataset collected from staff at these clinics.