Application of Structural Equation Modeling in Hypothesis Testing

Abstract

This is a methodological paper illustrating the use of structural equation modeling (SEM) technique in hypothesis testing. The illustration involves a job analysis assignment in human resource management. Hypotheses are set and tested. Empirical findings are used to make inferences in support of hypotheses or otherwise. Using a private dataset collected from bus drivers for a job analysis project, the paper has three stages. First, the data are subjected to exploratory factor analysis (EFA) to identify a reasonable number of constructs to include in the model. Reliability of the constructs is tested. Second, the constructs are named, and used to specify a measurement model using confirmatory factor analysis (CFA). The third stage uses structural equation modelling (SEM) to test the hypotheses and to re-specify the model before determining the final results.

Keywords: Exploratory factor analysis, Structural equation modeling, hypothesis testing

JEL Classification: M00