Table 3. Model fit indices of longitudinal ME/I modelling.
| Model | χ2(df), p = | RMSEA | CFI | NNFI | Δχ2(Δdf), p = | ΔCFI |
|---|---|---|---|---|---|---|
| Parent data (n = 216) | ||||||
| Configural | 1036.759(945), 0.020 | .021 | .976 | .969 | ||
| Loading | 1162.567(1045),0.06 | .023 | .970 | .964 | 86.792(126),0.997 | .006 |
| Threshold | 1288.091(1095),0.000 | .029 | .950 | .944 | 212.387(50),0.000 | -.020 |
| Threshold*1 | 1235.593(1088),0.001 | .026 | .962 | .957 | 111.632(43),0.000 | -.008 |
| Teacher data (n = 189) | ||||||
| Configural | 1037.104(945), 0.019 | .023 | .983 | .978 | ||
| Loading | 1109.484(1045),0.081 | .018 | .988 | .986 | 100.912(100),0.456 | .005 |
| Threshold | 1241.459(1095),0.001 | .027 | .973 | .969 | 191.334(50),0.000 | .015 |
| Threshold*2 | 1232.742(1093),0.002 | .026 | .974 | .971 | 179.478(48),0.000 | .001 |
Note.
*1 free 7 threshold estimates between follow-up time.
*2 Free 2 threshold estimates between follow-up time.
RMSEA = root mean square error of approximation; CFI = comparative fit index; NNFI = non-normal fit index. N = number of participants with at least one completed baseline or follow-up measure