Comparison of Factor Extraction Methods in Exploratory Factor Analysis and Their Impact on Model Quality
The study aims to compare three factor extraction methods in exploratory factor analysis—principal component analysis (PCA), principal axis factorization (PAF), and maximum likelihood (ML)—and to demonstrate the impact of these methods on factor structure and model quality indices. The study was conducted using survey data measuring satisfaction with teaching, consisting of ten items on a five-point Likert scale, collected from a sample of 400 students at the College of Management and Economics at the University of Kirkuk for the 2024–2025 academic year. The data were analyzed using SPSS Version 18, and the factor extraction methods were compared in terms of the number of extracted factors, the proportion of variance explained, communalities values, factor loadings, the interpretability of the factor structure, and fit indices where available. The results showed that three factors were extracted using each method, with differences in the proportion of variance explained, communalities values, and factor loadings. The principal components method achieved a higher proportion of total explained variance, while the principal axes and maximum likelihood methods focused on the variance shared among the variables. The maximum likelihood method also allowed for a statistical test of model fit. The study concluded that the choice between extraction methods should not be based on a single indicator, as the most appropriate method depends on the objective of the analysis; the principal components method is preferred when the goal is data reduction, while the principal axes and maximum likelihood methods are more suitable when seeking to uncover underlying factors. Thus, evaluating model quality requires consideration of the statistical objectives of each method and a set of integrated indicators.
Authors
- Omar Macky Rahim
- Sazan Najat Najadeen.
- Enas Salah Khorshid
Institutions
- University of Kirkuk (IQ)
Publication Details
- Journal
- Paaua Science Journal
- Published
- 2026-11-18
- DOI
- https://doi.org/10.65760/paaujs.v1i2.6
- Primary Topic
- Psychometric Methodologies and Testing
- Type
- article
- Field-Weighted Citation Impact
- 0.00