The Impact Of Sample Size On Exploratory And Confirmatory Factor Analysis: A Simulation Study

Abstract

This research examines how sample size influences results from exploratory (EFA) and confirmatory factor analysis (CFA), utilizing synthetically generated datasets. Although EFA and CFA are staple tools in psychology and sociology for modeling latent constructs, consensus on optimal sample requirements remains elusive. Simulations were conducted by constructing datasets with a fixed factor architecture across five sample scales (n=100, 200, 300, 500, 1000). Key metrics evaluated covered factor extraction consistency, loading precision, and model suit signs. Outcomes discovered dwindled validity and reproducibility in smaller samples, with reliable overall performance accomplished consistently at n≥300. The consequences provide actionable guidelines for optimizing thing evaluation designs in empirical research.