Estimating Standard Errors in Exploratory Factor Analysis.

Area of Science:

• Psychometrics
• Statistical Methodology
• Quantitative Psychology

Background:

• Standard errors are crucial for interpreting the reliability of factor analysis results.
• Accurate estimation of standard errors is essential for valid conclusions in exploratory factor analysis (EFA).
• Previous research has highlighted various challenges in obtaining reliable standard error estimates in EFA.

Purpose of the Study:

• To identify and describe six key issues that influence standard errors in exploratory factor analysis.
• To provide a comprehensive review of seven distinct methods for computing standard errors for rotated factor loadings and factor correlations.
• To illustrate the application of these standard error estimation methods using real-world data from personality and intelligence studies.

Main Methods:

• Review of six factors impacting standard error estimation in EFA.
• Detailed examination of seven computational methods: augmented information, nonparametric bootstrap, infinitesimal jackknife, asymptotic distributions of unrotated loadings, sandwich, parametric bootstrap, and jackknife.
• Empirical illustration using data from a personality study (N=537) and an intelligence study (N=145).

Main Results:

• The study identifies critical factors influencing standard error accuracy in EFA.
• It systematically compares seven different approaches to standard error calculation for factor analysis.
• Demonstrates practical application and potential differences in standard error estimates across methods.

Conclusions:

• Understanding the factors influencing standard errors is vital for appropriate use of EFA.
• The choice of method for computing standard errors can impact the interpretation of factor loadings and correlations.
• Provides researchers with a comparative overview to select suitable standard error estimation techniques in their EFA applications.