Table 2. Comparison of model fit for baseline model.
| Number of free parameters | χ2 (df) | RMSEA | CFI | |
|---|---|---|---|---|
| Dutch sample | ||||
| Model 1a | Total: 37 Factor loadings: 12 Thresholds: 24 Factor covariance: 1 |
2791.368 (53)* | 0.106 (0.102–0.109)b | 0.946b |
| Model 2 | Total: 40 Model 1 + 3 factor loadings |
1189.533 (50)* | 0.070 (0.067–0.074)b | 0.978 |
| Model 3 | Total: 44 Model 2 + 4 item covariances |
677.946 (46)* | 0.055 (0.051–0.058) | 0.988 |
| Model 4 (Model for MI testing) |
Total: 45 Model 3 + 1 item covariance |
507.833 (45)* | 0.047 (0.044–0.051) | 0.991 |
| Total sample | ||||
| Model 1a | Total: 43 Factor loadings: 12 Thresholds: 30 Factor covariance: 1 |
15008.851 (53)* | 0.110 (0.109–0.112)b | 0.967b |
| Model 2 | Total: 46 Model 1 + 3 factor loadings |
12646.343 (50)* | 0.104 (0.103–0.106)b | 0.972 |
| Model 3 | Total: 50 Model 2 + 4 item covariances |
5284.281 (46)* | 0.070 (0.069–0.072)b | 0.988 |
| Model 4 (Model for MI testing) |
Total: 51 Model 3 + 1 item covariance |
2264.403 (45)* | 0.046 (0.045–0.048) | 0.995 |
* P < .001
a Model 1:6 items load on the physical factor and 6 on the mental factor; factors are allowed to co-vary. The model is identified by constraining factor variances and means at 1 and 0, respectively. Model 2: 3 items (1, 10 and 12) load on both factors; Model 3: residuals of 4 item pairs belonging to the same subscale (i.e. 2&3, 4&5, 6&7, 9&11) are allowed to co-vary; Model 4: residuals of item pair 9 & 10 are allowed to co-vary
b Poor model fit (RMSEA>0.055; or CFI<0.97)