Which Factor Retention Method Performs Best? A Monte Carlo Simulation for Psychological and Educational Measurement

Bhilan Rhiyu Antama (Educational Research and Evaluation, Graduate School, Yogyakarta State University, Indonesia)
Edi Istiyono (Educational Research and Evaluation, Graduate School, Yogyakarta State University, Indonesia)
Moh. Khairudin (Educational Research and Evaluation, Graduate School, Yogyakarta State University, Indonesia)
Syukrul Hamdi (Educational Research and Evaluation, Graduate School, Yogyakarta State University, Indonesia)
Heni Setiyaningsih (Educational Research and Evaluation, Graduate School, Yogyakarta State University, Indonesia)

ARMADA : Jurnal Penelitian Multidisiplin
Vol. 4 No. 9 (2026): ARMADA : Jurnal Penelitian Multidisplin, September 2026

Determining the appropriate number of factors is the most critical decision in exploratory factor analysis (EFA), but evidence-based guidance for the context of skewed ordinal data remains limited. This study compares the performance of three factor retention methods: optimal coordinates (OC), parallel analysis (PA), and exploratory graph analysis (EGA) through Monte Carlo simulations with 1000 replications using negatively skewed 5-category Likert scale data. Controlled conditions included the number of factors (2, 3, 5), sample size (300, 500, 900, 1500), factor loadings (0.4; 0.7), and inter-factor correlations (0.0; 0.3). The results show that PA achieved the highest accuracy (94.7%), followed by EGA (92.8%), while OC lagged far behind (36.8%). PA achieved perfect accuracy on large samples with high loadings, whereas OC exhibited a tendency toward under-extraction that increased drastically as the number of factors increased. EGA outperformed PA in minimizing over-extraction but was more prone to under-extraction in complex structures. These findings recommend PA as the primary method and EGA as a complementary method in Likert-scale-based research and confirm that algorithmic scree plots are not suitable for use as the sole criterion for factor retention.