Table 2

Criteria for indices in SEM modeling

IndexThreshold for good fitBrief description
Chi-squaredx2Not significant (p > 0.05) or low in relation to degrees of freedomEvaluates the overall adequacy of the model. Sensitive to sample size
Normed fit indexNFI>0.90Compare the proposed model with a null model with no relationships
Tucker-Lewis indexTLI/NNFI>0.90Adjust the model for complexity and improve if the reduction in χ2 is proportional to the increase in degrees of freedom
Comparative fit indexCFI>0.90Compares observed and predicted covariance matrices
Root mean square error of approximationRMSEA<0.05 (good); <0.08 (acceptable)Evaluates the mean error per degree of freedom; includes confidence interval
Standardized root mean square residualSRMR<0.05Indicates the standardized average discrepancy between the observed and modeled covariance matrices
Source(s): Authors

or Create an Account

Close Modal
Close Modal