Sensitivity of Fit Indices to Model Misspecification and Model Types

By Xitao Fan, Stephen A. Sivo. 2007

611 citations recorded by the source corpus.

Subfields: Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/6ca621ab950b6c7c81f3e0739ccc08a74d5a1e35

How researchers see this paper

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

In Management Science and Operations Research

Citations in the corpus

The Servant Leadership Survey: Development and Validation of a Multidimensional Measure (2010, methodology): “The generally accepted values of good fit are close to .95 for the CFI and the TLI, and less than .08 for the SRMR and RMSEA (Fan and Sivo 2007; Hu and Bentler 1998).”
Factors influencing early adolescents’ mathematics achievement: High-quality teaching rather than relationships (2013, methodology): “These include the standardised root-mean-square residual (SRMR), the root-mean-square error of approximation (RMSEA) and gamma hat, which have been shown to be resistant to the impact of large samples, complex models and model misspecification (Fan and Sivo 2007; Hu and Bentler 1999; Vandenberg and ”
The Relationship Between Root Mean Square Error of Approximation and Model Misspecification in Confirmatory Factor Analysis Models (2012, background): “In a related approach, Fan and Sivo (2005, 2007) suggest that the relative ‘‘size’’ of misspecification should be linked to the power of the model test statistic to reject the model.”
The dirty dozen: a concise measure of the dark triad. (2010, background): “Although some of the fit indexes could have been stronger (especially for the one-dimensional model), fit indexes do not have strict, nonarbitrary cutoff criteria (Fan & Sivo, 2007).”
Relative Performance of Categorical Diagonally Weighted Least Squares and Robust Maximum Likelihood Estimation (2014, methodology): “The severity of the model misspecifications was gauged in the manner outlined by Fan and Sivo (2007), by calculating power values as well as values of the fit function for the two model misspecifications.”
Health-specific leadership: Is there an association between leader consideration for the health of employees and their strain and well-being? (2011, background): “CFI values close to 0.95 indicate excellent fit and those] 0.90 are interpreted as good (Bentler, 2007; Chen, Curran, Bollen, Kirby, & Paxton, 2008; Fan & Sivo, 2007).”
Testing Structural Equation Models or Detection of Misspecifications? (2009, methodology): “Given the relationship between the standard test statistic T and the FIs, the same problems occur for the FIs as have been documented in detail by Fan and Sivo (2007).”
Applied Psychometrics: Sample Size and Sample Power Considerations in Factor Analysis (EFA, CFA) and SEM in General (2018, background): “Moreover, particular fit indices “react” differently in small sample sizes along with model estimators, model complexity, multivariate normality assumption and variable independence (Fan & Sivo, 2007; Saris, Satorra, & van der Veld, 2009 as cited in Byrne, 2012).”
Cognitive appraisal as a mediator in the relationship between stress and burnout (2013, background): “…greater than .90 indicate acceptable fit; NFI values greater than .95 indicate excellent fit and values ≥ .90 indicate good fit; and CFI values close to .95 indicate excellent fit and values > .90 indicate good fit (Bentler, 2007; Chen, Curran, Bollen, Kirby, & Paxton, 2008; Fan & Sivo, 2007).”
Author Guidelines for Reporting Scale Development and Validation Results in the Journal of the Society for Social Work and Research (2010, neutral): “Authors who choose to use lower cutoffs should provide a thorough rationale supported by ample citations from the recent SEM literature (e.g., Marsh, Hau, & Wen, 2004; Fan & Sivo, 2007).”
Bootstrapping Confidence Intervals for Fit Indexes in Structural Equation Modeling (2016, result): “Even though we do not study it directly, our results for the RMSEA will also apply to the less known fit index gamma (Fan & Sivo, 2007), defined in the sample as γ ¼ ppþ2ðT df Þ=ðn 1Þ ¼ ppþ2df RMSEA2 .”