TABLE 3.
Model | χ 2 | df | χ 2/df | Δ χ 2 | Δ df | p-Value for Δ χ 2 | RMSEA (90% CI) | CFI | NFI | IFI | AIC | ECVI |
Threshold for acceptable fit | <5 | ≥0.001 (significant level) | <0.05 | ≥0.90 | ≥0.90 | ≥0.90 | ||||||
Age | ||||||||||||
Unconstrained | 55.717 | 20 | 2.786 | 0.021 (0.015–0.028) | 0.997 | 0.995 | 0.997 | 159.717 | 0.041 | |||
Measurement weights | 56.928 | 24 | 2.372 | 1.211 | 4 | 0.876 | 0.019 (0.012–0.025) | 0.997 | 0.995 | 0.997 | 152.928 | 0.039 |
Structural covariances | 92.827 | 39 | 2.380 | 37.111 | 19 | 0.008 | 0.019 (0.014–0.024) | 0.995 | 0.991 | 0.995 | 158.827 | 0.040 |
Measurement residuals | 115.710 | 46 | 2.515 | 59.993 | 26 | < 0.001 | 0.020 (0.015–0.024) | 0.994 | 0.991 | 0.994 | 167.710 | 0.043 |
Gender | ||||||||||||
Unconstrained | 60.128 | 20 | 3.006 | 0.022 (0.016–0.029) | 0.996 | 0.994 | 0.996 | 164.128 | 0.042 | |||
Measurement weights | 76.843 | 24 | 3.202 | 16.716 | 4 | 0.002 | 0.024 (0.018–0.030) | 0.995 | 0.993 | 0.995 | 172.843 | 0.044 |
Structural covariances | 205.066 | 39 | 5.258 | 144.939 | 19 | < 0.001 | 0.033 (0.028–0.037) | 0.984 | 0.981 | 0.984 | 271.066 | 0.069 |
Measurement residuals | 298.891 | 46 | 6.498 | 238.764 | 26 | < 0.001 | 0.037 (0.033–0.041) | 0.976 | 0.972 | 0.976 | 350.891 | 0.089 |
Socioeconomic status | ||||||||||||
Unconstrained | 55.013 | 20 | 2.751 | 0.021 (0.014–0.028) | 0.997 | 0.995 | 0.997 | 159.013 | 0.040 | |||
Measurement weights | 59.237 | 24 | 2.468 | 4.223 | 4 | 0.377 | 0.019 (0.013–0.026) | 0.997 | 0.994 | 0.997 | 155.237 | 0.039 |
Structural covariances | 103.570 | 39 | 2.656 | 48.557 | 19 | < 0.001 | 0.020 (0.016–0.025) | 0.994 | 0.990 | 0.994 | 169.570 | 0.043 |
Measurement residuals | 108.863 | 46 | 2.367 | 53.850 | 26 | 0.001 | 0.019 (0.014–0.023) | 0.994 | 0.990 | 0.994 | 160.863 | 0.041 |
χ2 = chi-square; df = degree of freedom; χ2/df = The chi-squared/freedom ratio; RMSEA = Root mean square error of approximation; CFI = Comparative fit index; NFI = Normed fit index; IFI = Incremental fit index; AIC = Akaike information criteria; ECVI = Expected cross-validation index. The chi-square deference (Δχ2) tests are compared with the unconstrained model (baseline model).