Exploratory Factor Analysis is a multivariate statistical method within psychometrics and data analysis designed to identify latent (unmeasured) structures by grouping observed variables based on their inter-correlations. The theoretical mechanism relies on the assumption that correlations among multiple observable phenomena are driven by underlying common factors, which serve as hidden constructs influencing these measured variables. This approach aims to reduce dimensionality by partitioning variance into shared components explained by specific latent factors and unique error variances, thereby simplifying complex datasets while preserving structural relationships between concepts.
Prerequisite chain context: requires Eigenvalue Decomposition in Linear Algebra.