Assessing Fit in Structural Equation Models: A Monte-Carlo Evaluation of RMSEA Versus SRMR Confidence Intervals and Tests of Close Fit

By Albert Maydeu-Olivares, Dexin Shi, Yves Rosseel. 2018

177 citations recorded by the source corpus.

We compare the accuracy of confidence intervals (CIs) and tests of close fit based on the root mean square error of approximation (RMSEA) with those based on the standardized root mean square residual (SRMR). Investigations used normal and nonnormal data with models ranging from p = 10 to 60 observed variables. CIs and tests of close fit based on the SRMR are generally accurate across all conditions (even at p = 60 with nonnormal data). In contrast, CIs and tests of close fit based on the RMSEA are only accurate in small models. In larger models (p ≥ 30), they incorrectly suggest that models do not fit closely, particularly if sample size is less than 500.

Subfields: Computer Networks and Communications, Experimental and Cognitive Psychology, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/ff690cb44bd0a6783a3ff0ab66c4e4a01f18986b

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

1. Chuanji Gao
2. Albert Maydeu-Olivares
3. Dexin Shi
4. Milica Miočević
5. R. van de Schoot
6. Alexander Robitzsch
7. Alex J. Bowers
8. Li Cai

In Experimental and Cognitive Psychology

1. Chuanji Gao
2. Dexin Shi
3. Albert Maydeu-Olivares
4. Milica Miočević
5. R. van de Schoot

In Management Science and Operations Research

1. Dexin Shi
2. Chuanji Gao
3. Albert Maydeu-Olivares
4. Milica Miočević
5. R. van de Schoot
6. Alexander Robitzsch
7. Daniel A. Sass
8. Holmes Finch

Citations in the corpus

• Estimating the Maximum Likelihood Root Mean Square Error of Approximation (RMSEA) with Non-normal Data: A Monte-Carlo Study (2020, methodology): “, 2012; MaydeuOlivares, 2017b; Maydeu-Olivares, Shi, & Rosseel, 2018), and inflated false-positive error rate of tests of close fit (Maydeu-Olivares, 2017b; Maydeu-Olivares et al., 2018) due to non-normality.”
• Small Sample Size Solutions (2020, background): “Although these tests were not constructed with small sample sizes in mind, they seem to work well even when n = 100 (the smallest sample size considered in Maydeu-Olivares et al., 2018) and the model is not too large.”