A Methodological Comparison of Exploratory Factor Analysis Results Between Spss and R Using The Big Five Inventory

Abstract

Objective: This study examines whether the default algorithms and configurations of SPSS and R produce different results in exploratory factor analysis (EFA) when applied to the same data. Method: The Big Five Inventory (IPIP-50, N = 1,000) was analyzed using SPSS 29 and R (psych package 2.4.3), comparing two extraction methods (Principal Components Analysis and Principal Axis Factoring) and two rotation methods (Varimax and Promax). Differences in factor loadings, communalities, explained variance, and number of factors suggested by Kaiser criterion and Parallel Analysis were documented. Results: Systematic numerical differences were found between both software packages. Factor loadings differed on average 0.08 points (range: 0.00-0.24) for PAF + Promax, with larger differences in items with cross-loadings. Kaiser criterion suggested 5 factors in both software packages, but Parallel Analysis suggested 6 factors in R and 5 in SPSS. Explained variance differed up to 2.3 percentage points.