Table 2.
Model fit statistics.
| Fit indices | Cut-off criterion | Sensitive to N | Penalty for model complexity |
|---|---|---|---|
| Absolute fit indices | |||
| Chi-Square (χ2) | • Lowest comparative value between measurement models | Yes | No |
| • Non-Significant Chi-Square (p > 0.01) | |||
| • Significant difference in Chi-Square between Models | |||
| • For Model Comparison: Retain Model with Lowest Chi-Square | |||
| Approximate fit indices | |||
| Root-Means-Square Error of Approximation (RMSEA) | • 0.06–0.08 (Marginally Acceptable); 0.01–0.05 (Excellent) | No | Yes |
| • Not-significant (p > 0.01) | |||
| • 90% Confidence Interval Range should not include Zero | |||
| • 90% Confidence Interval Range should not overlap between models | |||
| • For model comparison: Retain Model where ΔRMSEA ≤ 0.015 | |||
| Standardized Root Mean Square Residual (SRMR) | • 0.06 to 0.08 (Marginally Acceptable); 0.01–0.05 (Excellent) | Yes | No |
| • For model comparison: Retain Model where ΔSRMR ≤ 0.015 | |||
| Incremental fit indices | |||
| Comparative Fit Index (CFI) | • 0.90 to 0.95 (Marginally Acceptable Fit); 0.96 to 0.99 (Excellent) | No | No |
| • For model comparison: Retain Model with Highest CFI value (ΔCFI > 0.01) | |||
| Tucker-Lewis Index (TLI) | • 0.90 to 0.95 (Marginally Acceptable Fit); 0.96 to 0.99 (Excellent) | No | Yes |
| • For model comparison: Retain Model with Highest TLI value (ΔTLI > 0.01) | |||
| Akaike Information Criterion (AIC) | • Lowest value in comparative measurement models | Yes | Yes |
| Consistent AIC (CAIC; calculated as BIC + free parameters | • Lowest value in comparative measurement models | Yes | Yes |
| Bayes Information Criterion (BIC) | • Lowest value in comparative measurement models | Yes | Yes |
| Sample-Size Adjusted BIC (aBIC) | • Lowest value in comparative measurement models | Yes | Yes |