The Henseler-Ogasawara specification of composites in structural equation modeling: A tutorial

Area of Science:

• Statistics
• Psychometrics

Background:

• Structural Equation Modeling (SEM) is a powerful statistical technique.
• Researchers face challenges integrating composite variables into SEM.
• Existing methods for handling composites in SEM have limitations.

Purpose of the Study:

• To present a novel specification for SEM that effectively incorporates composite variables.
• To demonstrate the advantages of this new approach over existing methods.
• To facilitate the use of advanced SEM features with composite variables.

Main Methods:

• The study proposes a specification for SEM based on Henseler's (2021) work.
• This approach uses synthetic variables, termed emergent and excrescent variables, to represent composites.
• The method draws on canonical correlation analysis principles.

Main Results:

• The presented SEM specification allows for the seamless integration of composite variables.
• Emergent variables are linear combinations related to other model variables.
• Excrescent variables are linear combinations unrelated to other model variables.

Conclusions:

• This new SEM specification overcomes practical challenges in modeling composite variables.
• The approach allows leveraging existing SEM developments like parameter estimation and model fit testing.
• It opens new research avenues for SEM in studying composites.