Comparison of Factor Retention Methods: Is Exploratory Graph Analysis Appropriate for Mixed-Format Data?

• In Event: Virtual Poster Session
  In Virtual iPresentation Only: Division D - Section 2: Quantitative Methods and Statistical Theory Virtual Poster Session

Abstract

In this study, factor retention methods' performances with mixed-datasets(a combination of dichotomous and polytomous) was outlined. With Monte Carlo simulations, the performance of EGA-TMFG and EGA-GLASSO are compared to Parallel Analysis (PA) and Minimum Average Partial Analysis (MAP) across various conditions, including different factor loadings, sample sizes, and distributions of variables. Results show that PA and MAP generally outperform EGA in both accuracy and bias, especially under conditions of normal distribution. EGA's performance diminishes with non-normal data and lower sample sizes. The study concludes that while EGA is promising, PA and MAP remain more reliable for mixed-format data. Implications of these findings are discussed.

Authors

• Tugay Kaçak, Trakya University

• Abdullah Faruk Kılıç, Trakya University

• Nuri Dogan, Hacettepe University