Table 5.
Model Fit Statistics using WLSMV Estimation and Difference Testing
| Model Fit Indices | 4-Factor | 3-Factor | 2-Factor | 1-Factor | ||||
|---|---|---|---|---|---|---|---|---|
| χ2/df | 1.19 | 0.78 | 1.31 | 0.71 | ||||
| RMSEA | 0.04 | 0.05 | 0.05 | 0.06 | ||||
|
C.I. Pclose-fit H0 |
0.0–0.09 0.57 |
0.0–0.09 0.47 |
0.0–0.09 0.43 |
0.0–0.10 0.31 |
||||
| CFI | 0.96 | 0.93 | 0.92 | 0.90 | ||||
| Δ χ2 | 4-Factor and 3-Factor 05.88 (df = 2), p = 0.05 | 4-Factor and 2-Factor 09.61 (df = 4), p < 0.05 | 4-Factor and 1-Factor 14.20 (df = 5), p = 0.01 | |||||
Note. Weighted least squares-mean and variance adjusted (WLSMV), Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), 90% Confidence Interval, Probability RMSEA < = 0.05 (Pclose-fit H0). Model comparison using difference testing against the 4-factor model with the DIFFTEST option in Mplus (Δ χ2). The models met the recommended identification assumptions; the model degrees of freedom (df) was greater than zero and scaling constraints were imposed on the variances of the latent factors and loadings of the error terms. The four-factor model was identified by fixing the error term of the single indicator factor to equal 1- r (S2), where r equals reliability (Kline, 2016)