Approaches to Testing Interaction Effects Using Structural Equation Modeling Methodology.

Area of Science:

• Psychology
• Statistics
• Behavioral Science

Background:

• Structural Equation Modeling (SEM) is crucial for analyzing latent variables.
• Testing interactions among continuous latent variables has historically posed challenges.
• Kenny and Judd's (1984) product indicator method was foundational but difficult to implement.

Purpose of the Study:

• To review and compare simplified methods for testing latent variable interactions in SEM.
• To address the implementation difficulties of earlier interaction testing techniques.
• To provide an empirical example of interaction analysis in exercise motivation.

Main Methods:

• Review of single indicator (Joreskog & Yang, 1996) and multiple indicator (Jaccard & Wan, 1995; Ping, 1996) approaches.
• Application of these methods to an empirical dataset.
• Examination of the interactive effects of perceived competence and autonomy on exercise motivation.

Main Results:

• Simplified approaches facilitate the testing of interactions involving continuous latent variables.
• The empirical example demonstrates the practical application of these methods.
• Discussion of practical considerations for implementing interaction analyses in SEM.

Conclusions:

• Newer SEM methodologies offer practical solutions for testing latent variable interactions.
• These advancements enhance the ability to study complex relationships in psychological research.
• Researchers can now more readily investigate interactive effects on outcomes like exercise motivation.