Assessing the Stability of Factor Structures in Exploratory Factor Analysis via Parametric and Nonparametric Bootstrapping

• In Event: AERA Roundtable Session Sunday 11:40 am Four Seasons Ballroom 2-3
  In Roundtable Session: Measurement Issues (Table 9)

Abstract

A primary concern in exploratory factor analysis is the stability of the factor structure. One major challenge in EFA is that traditional approaches for cross-validation such as using multiple datasets or splitting data have their limitations. The present study proposes using parametric and nonparametric bootstrapping methods to assess the stability of factor structures in EFA. Monte Carlo simulations were conducted to systematically investigate the performance of bootstrapping approaches in identifying the correct number of factors and item assignment across normal and nonnormal data distributions. Results indicate that parametric bootstrapping generally outperformed nonparametric bootstrapping under small to medium samples. In large sample sizes (n ≥ 1000) with big loadings (.7), both bootstrap approaches demonstrated reliable and accurate assessments of factor structures.

Authors

• Xinchang Zhou, University of Illinois at Urbana-Champaign

• Yan Xia, University of Illinois at Urbana-Champaign