Models | Fit statistics | Difference tests of relative fit | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
S-B χ2 | df | p | CFI | TLI | RSMEA | 90% CI | SMRS | Model comparison | DS-B χ2 (df)a | Dp | Scaling correction factor | |
1. Conventional cross-lagged panel model with freely estimated paths | 152.602 | 9 | < 0.001 | 0.964 | 0.867 | 0.090 | 0.078–0.103 | 0.031 |  |  |  | 1.309 |
2. Random intercept cross-lagged panel model with freely estimated paths | 4.991 | 3 | 0.172 | 1.000 | 0.995 | 0.018 | 0.000–0.046 | 0.008 | 2 vs. 1 | 140.033 (6) | < 0.001 | 1.156 |
3. Both contemporaneous and cross-lagged paths are constrained to be equal (= no change over time) | 20.696 | 15 | 0.147 | 0.999 | 0.997 | 0.014 | 0.000–0.027 | 0.015 | 3 vs. 2 | 15.737 (12) | 0.204 | 1.192 |
4 Models with auto-regressive paths are constrained to be equal | ||||||||||||
 a. Self-esteem paths | 4.766 | 4 | 0.312 | 1.000 | 0.999 | 0.010 | 0.000–0.037 | 0.008 | 4a vs. 2 | 0.006 (1) | 0.939 | 1.212 |
 b. Coping styles paths | 5.198 | 4 | 0.268 | 1.000 | 0.998 | 0.012 | 0.000–0.038 | 0.008 | 4b vs. 2 | 0.187 (1) | 0.665 | 1.151 |
 c. Anxiety symptoms paths | 5.150 | 4 | 0.272 | 1.000 | 0.998 | 0.012 | 0.000–0.038 | 0.008 | 4c vs. 2 | 0.456 (1) | 0.500 | 1.259 |
5. Select optimal model: reciprocal model with constrained auto-regressive paths and constrained cross-lagged paths | 21.960 | 18 | 0.234 | 0.999 | 0.998 | 0.011 | 0.000–0.024 | 0.016 | 5 vs. 2 | 17.061 (15) | 0.315 | 1.228 |
6. Controlling for covariates based on Model 5 | 142.532 | 78 | < 0.001 | 0.989 | 0.982 | 0.021 | 0.015–0.026 | 0.027 |  |  |  | 1.115 |
7. Gender difference | ||||||||||||
 a. Unconstrained model | 71.359 | 48 | 0.016 | 0.994 | 0.991 | 0.025 | 0.011–0.037 | 0.035 |  |  |  | 1.245 |
 b. Constrained model | 128.921 | 60 | < 0.001 | 0.983 | 0.980 | 0.038 | 0.029–0.048 | 0.074 | 7b vs. 7a | 62.900 (12) | < 0.001 | 1.209 |