An Evaluation of Using Correlation Residuals to Assess Model Fit in SEM

• In Event: Evaluating Model Fit in Structural Equation Models

Abstract

Methods of evaluating model fit in SEM have continuously been scrutinized with respect to their inconsistent performance. Recently, Kline (2023) has recommended the use of correlation residuals to support model fit. The purpose of the current study is to contribute to the discussion about using correlation residuals for assessing model fit and to explore correlation residuals under varying conditions. Confirmatory factor analysis models were examined under different sample sizes and misspecification conditions. Model fit information and correlation residuals were evaluated. Preliminary results indicated that correlation residuals would detect severely misspecified models when using the suggested cutoff value of .10. However, the correlation residuals would also suggest misspecification with correct models under small sample sizes. Implications of the results are discussed.

Authors

• Annemarie DeChellis, University of Texas at Austin

• Tiffany A. Whittaker, University of Texas at Austin