An empirical evaluation of alternative methods of estimation for confirmatory factor analysis with ordinal data.

By David B. Flora, P. Curran. Psychological methods, 2004

2,854 citations recorded by the source corpus.

Subfields: Computer Networks and Communications, Food Science, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/14e0a32ce0d079e45a54b78c5bc30735750c955f

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Networks and Communications

In Food Science

In Statistics and Probability

Citations in the corpus

Determinants of classroom engagement: a prospective test based on self-determination theory (2018, methodology): “Because the observed variables (items) were ordered categorically (Flora & Curran, 2004), weighted least squares mean and variance-adjusted (WLSMV) estimation was used.”
Undergraduate Engineering Students’ Beliefs, Coping Strategies, and Academic Performance: An Evaluation of Theoretical Models (2012, methodology): “We used this correlation matrix and estimation procedure because it is more statistically and theoretically appropriate with ordered categorical data (Flora & Curran, 2004).”
The Performance of Maximum Likelihood and Weighted Least Square Mean and Variance Adjusted Estimators in Testing Differential Item Functioning With Nonnormal Trait Distributions (2015, methodology): “We chose RWLS as a limited information estimator because this method works effectively in most situations where ordered categorical variables are used with CFA for various sample sizes (Flora & Curran, 2004).”
The Difficulties in Emotion Regulation Scale: Factor Structure and Association with Nonsuicidal Self-Injury in Adolescent Inpatients (2012, methodology): “investigation), sample sizes of 150–200 have been found to be sufficient (Brown 2006; Flora and Curran 2004).”
Confirmatory Factor Analysis of Ordinal Variables With Misspecified Models (2010, background): “A common experience with WLS both for continuous and categorical variables is the poor performance of the WLS estimators and their associated standard errors and chi-squares (see, e.g., Bentler, 1995; Dolan, 1994; Flora & Curran, 2004; Potthast, 1993; West, Finch, & Curran, 1995).”
Assessing Internet Addiction Using the Parsimonious Internet Addiction Components Model—A Preliminary Study (2013, background): “In comparison to ordinary weighted least squares analyses (WLS),WLSMVismore robust with regards to non-normality of data and less restrictive with regards to sample size, and it corrects means and variances (Flora and Curran 2004).”
Assessing Goodness of Fit: Simple Rules of Thumb Simply Do Not Work (2011, result): “Noting the contradictory results among studies examining the robustness of CFA models, Hoogland and Boomsma (1998) suggested that previous work was limited by the lack of realism in the simulated models.”
Instrumental leadership: Measurement and extension of transformational–transactional leadership theory (2014, background): “This estimator provides many advantages to those that assume that ordered items are continuous and can also incorporate continuous variables in the model (Flora & Curran, 2004; Muthén, 1984).”
An evaluation of WLSMV and Bayesian methods for confirmatory factor analysis with categorical indicators (2014, methodology): “Based on the results from previous studies (Asparouhov and Muthén, 2010a; Choi et al., 2011; Flora and Curran, 2004; Yang and Green, 2010), each method comes with merits and limitations in terms of the overall model-data fit, parameter estimates, and SEs of the parameter estimates.”
Determinants of residential satisfaction in urban China: A multi-group structural equation analysis (2017, methodology): “Therefore, we shall apply the Weighted Least Squares (WLS) estimator based on the matrix of polychoric correlations (see Flora and Curran (2004) and Jöreskog and Sörbom (1996) for details).”
Mothering versus Fathering versus Parenting: Measurement Equivalence in Parenting Measures (2007, methodology): “Some (eg, Flora & Curran, 2004 ) have recommended the use of Weighted Least Squares–MV estimation to correct this problem; unfortunately, this estimation method is not available in the SEM software used for the present study.”
Conducting Measurement Invariance Tests with Ordinal Data: A Guide for Social Work Researchers (2015, methodology): “In Mplus, the recommended robust WLS estimator is a means and variance adjusted WLS labeled as WLSMV (Flora & Curran, 2004; Muthén & Muthén, 1998–2012).”
Confirmatory Factor Analytic Procedures for the Determination of Measurement Invariance (2006, methodology): “Much of this development has focused on such issues as (a) goodness-of-fit indexes (Cheung & Rensvold, 2002), (b) analysis of ordinal data (Flora & Curran, 2004; Lubke & Muthén, 2004; Millsap & Yun-Tein, 2004), (c) appropriate practices for testing invariance (Vandenberg & Lance, 2000), (d)…”