Table 3.
Fit indices for Cross-lagged Models at Baseline, Fixed across Time, and Fixed across Gender.
| S-Bχ2 (df) | p | ΔS-Bχ2 (df) | p | CFI | ΔCFI | SRMR | ΔSRMR | RMSEA [90% CI] | ΔRMSEA | |
|---|---|---|---|---|---|---|---|---|---|---|
| Model 1 | ||||||||||
| Baseline | 263.545 (16) | <0.001 | – | – | 0.977 | – | 0.021 | – | 0.085 [0.076, −0.094] | – |
| Fixed across time | 279.953 (28) | <0.001 | 12.792 (12) | 0.384 | 0.976 | 0.001 | 0.021 | 0.000 | 0.065 [0.058, −0.072] | 0.020 |
| Fixed across gender | 329.159 (68) | <0.001 | 6.104 (12) | 0.911 | 0.970 | 0.006 | 0.028 | 0.007 | 0.060 [0.053, −0.066] | 0.005 |
| Model 2 | ||||||||||
| Baseline | 272.412 (16) | <0.001 | – | – | 0.976 | – | 0.021 | – | 0.086 [0.077, −0.095] | – |
| Fixed across time | 287.206 (28) | <0.001 | 11.766 (12) | 0.465 | 0.976 | 0.000 | 0.022 | 0.001 | 0.065 [0.059, −0.072] | 0.021 |
| Fixed across gender | 340.875 (68) | <0.001 | 5.772 (12) | 0.927 | 0.969 | 0.007 | 0.029 | 0.007 | 0.061 [0.055, −0.068] | 0.004 |
| Model 3 | ||||||||||
| Baseline | 256.818 (16) | <0.001 | – | – | 0.976 | – | 0.021 | – | 0.084 [0.075, −0.093] | – |
| Fixed across time | 270.874 (28) | <0.001 | 12.127 (12) | 0.435 | 0.976 | 0.000 | 0.022 | 0.001 | 0.063 [0.057, −0.070] | 0.021 |
| Fixed across gender | 317.635 (68) | <0.001 | 10.533 (12) | 0.569 | 0.970 | 0.006 | 0.030 | 0.008 | 0.058 [0.052, −0.065] | 0.005 |
| Model 4 | ||||||||||
| Baseline | 266.017 (16) | <0.001 | – | – | 0.976 | – | 0.021 | – | 0.085 [0.076, −0.094] | – |
| Fixed across time | 279.873 (28) | <0.001 | 11.990 (12) | 0.446 | 0.976 | 0.000 | 0.022 | 0.001 | 0.065 [0.058, −0.072] | 0.020 |
| Fixed across gender | 330.659 (68) | <0.001 | 10.449 (12) | 0.684 | 0.969 | 0.007 | 0.030 | 0.008 | 0.060 [0.053, −0.066] | 0.005 |
| Model 5 | ||||||||||
| Baseline | 294.173 (16) | <0.001 | – | – | 0.973 | – | 0.025 | – | 0.090 [0.081, −0.099] | – |
| Fixed across time | 309.897 (28) | <0.001 | 10.370 (12) | 0.584 | 0.972 | 0.001 | 0.026 | 0.001 | 0.068 [0.062, −0.075] | 0.022 |
| Fixed across gender | 365.679 (68) | <0.001 | 9.217 (12) | 0.684 | 0.965 | 0.007 | 0.033 | 0.007 | 0.064 [0.057, −0.070] | 0.004 |
| Model 6 | ||||||||||
| Baseline | 304.812 (16) | <0.001 | – | – | 0.972 | – | 0.025 | – | 0.091 [0.083, −0.101] | – |
| Fixed across time | 321.220 (28) | <0.001 | 10.972 (12) | 0.531 | 0.971 | 0.001 | 0.026 | 0.001 | 0.070 [0.063, −0.077] | 0.021 |
| Fixed across gender | 365.191 (68) | <0.001 | 9.387 (12) | 0.670 | 0.964 | 0.007 | 0.031 | 0.005 | 0.072 [0.065, −0.079] | 0.002 |
S-Bχ2, Satorra-Bentler Chi-square; CFI, Comparative Fit Index; SRMR, Standardized Root Mean Square Residual; RMSEA, Root Mean Square Error of Approximation. While some changes in RSMEA (ΔRMSEA) are larger than 0.015, these changes point to an improvement of model fit in the constrained models. Increasing the degrees of freedom have indeed been found to result in smaller RMSEA values (Taasoobshirazi and Wang, 2016).