Issues and recommendations for exploratory factor analysis and principal component analysis.

Area of Science:

• Statistics
• Psychometrics
• Data Analysis

Background:

• Exploratory factor analysis (EFA), common factor model (CFM), and principal components analysis (PCA) are statistical methods for identifying underlying structures in data.
• Understanding the distinctions and appropriate applications of these methods is crucial for rigorous research.
• Previous guidance often lacks specific recommendations for different data types and research goals.

Purpose of the Study:

• To provide a concise mathematical review of EFA, CFM, and PCA.
• To offer detailed recommendations on key methodological choices in factor analysis, including measurement scales, estimation techniques, factor and item retention, and rotation.
• To guide researchers in selecting and applying appropriate factor analysis techniques for identifying latent variables.

Main Methods:

• Mathematical review of the theoretical underpinnings of EFA, CFM, and PCA.
• Comparative analysis of methodological choices relevant to factor analysis.
• Discussion of alternative data analysis techniques such as item response theory and machine learning.

Main Results:

• Exploratory factor analysis, specifically the common factor model, is identified as the appropriate analysis for researchers aiming to identify latent factors.
• For survey data utilizing Likert-type scales, weighted least squares with robust standard errors and oblique rotation are recommended.
• A checklist for researchers and reviewers is provided to ensure best practices in factor analysis.

Conclusions:

• Researchers seeking to identify latent factors should utilize exploratory factor analysis (common factor model).
• Specific methodological recommendations are provided for Likert-type scales, emphasizing weighted least squares and oblique rotation.
• The commentary aims to enhance the quality and interpretability of factor analysis in research applications.