Table 2.
Fit statistics for the measurement model and measurement invariance models across grades.
| Model | k | χ 2 | RMSEA | CFI | TLI | SRMR | ΔCFIa | |||
|---|---|---|---|---|---|---|---|---|---|---|
| Value | df | p | Value | 90% C.I. | ||||||
| Measurement model | 87 | 545.1 | 188 | <0.001 | 0.042 | 0.038–0.046 | 0.972 | 0.965 | 0.057 | - |
| Measurement invariance across grades | ||||||||||
| Configural | 435 | 1,466.4 | 940 | <0.001 | 0.051 | 0.046–0.056 | 0.956 | 0.946 | 0.069 | - |
| Metric | 375 | 1,570.4 | 1,000 | <0.001 | 0.051 | 0.046–0.056 | 0.952 | 0.945 | 0.076 | −0.004 |
| Scalar | 315 | 1,679.7 | 1,060 | <0.001 | 0.052 | 0.047–0.057 | 0.948 | 0.943 | 0.078 | −0.004 |
k, number of parameters; df, degrees of freedom; RMSEA, root mean square error of approximation; C.I., confidence interval; CFI, comparative fit index; TLI, Tucker–Lewis index; and SRMR, standardized root mean square residual.
a
Only applicable on the measurement invariance models. ΔCFI describes the difference in CFI from the less restrictive measurement invariance model (i.e., the metric model compared with the configural, and the scalar model compared with the metric).