Exploratory Procedure for Component-Based Structural Equation Modeling for Simple Structure by Simultaneous Rotation

Area of Science:

• Statistics
• Quantitative Psychology
• Data Analysis

Background:

• Generalized Structured Component Analysis (GSCA) is a confirmatory method for examining relationships between observed variables and components.
• Existing exploratory SEM methods lack the component-based approach of GSCA.
• There is a need for an exploratory version of GSCA to facilitate model discovery.

Purpose of the Study:

• To propose and develop an exploratory version of Generalized Structured Component Analysis (EGSCA).
• To introduce a novel rotational algorithm for EGSCA designed for simultaneous parameter matrix simplification.
• To evaluate the performance of EGSCA in recovering true parameter values and its effectiveness in model exploration.

Main Methods:

• Theoretical demonstration of orthogonal rotation indeterminacy in GSCA as justification for EGSCA.
• Development and presentation of the EGSCA procedure.
• Implementation of a specialized rotational algorithm for EGSCA.
• Conducting two numerical simulation studies to assess parameter recovery.
• Application of EGSCA to two real-world datasets.

Main Results:

• EGSCA, utilizing the proposed rotation, successfully recovered true parameter matrices in simulation studies.
• EGSCA demonstrated superior performance compared to the existing GSCA procedure in simulations.
• Analysis of real datasets indicated that EGSCA-derived models were an improvement over previously proposed models.

Conclusions:

• EGSCA provides an effective method for exploratory analysis in structural equation modeling.
• The proposed rotational algorithm enhances the simplification of parameter matrices within EGSCA.
• EGSCA demonstrates significant potential for improving model exploration and uncovering complex relationships in data.