Title: An evaluation of WLSMV and Bayesian methods for confirmatory factor analysis with categorical indicators
Authors: Xinya Liang; Yanyun Yang
Addresses: Department of Educational Psychology and Learning Systems, Florida State University, Tallahassee, FL, 32306, USA ' Department of Educational Psychology and Learning Systems, Florida State University, Tallahassee, FL, 32306, USA
Abstract: This Monte Carlo study evaluated the performance of three estimation methods in fitting confirmatory factor analysis (CFA) models for ordered categorical data, with the focus on data with non-normal underlying distributions and small sample sizes. The three methods are: weighted least squares with mean and variance adjusted (WLSMV), Bayesian with non-informative priors (BN), and Bayesian with informative priors (BI). Design factors included sample sizes, factor structures, underlying continuous distributions, and categorical distributions. Results were evaluated based on the model-data fit, point estimates, and standard errors of point estimates. Results showed that Bayesian methods encountered less convergence problems than WLSMV. Bayesian methods were robust to the non-normality of underlying continuous distributions. WLSMV tended to perform equally well or slightly better than Bayesian methods except for some conditions with small sample sizes or highly non-normal underlying distributions.
Keywords: confirmatory factor analysis; CFA; categorical indicators; weighted least squares with mean and variance adjusted; WLSMV; Bayesian estimation; non-normal distributions; small sample sizes.
International Journal of Quantitative Research in Education, 2014 Vol.2 No.1, pp.17 - 38
Available online: 23 May 2014 *Full-text access for editors Access for subscribers Purchase this article Comment on this article