Table 1 Fit indices for all models and scoring methods
From: The reliability of the twelve-item general health questionnaire (GHQ-12) under realistic assumptions
Model | H0 χ2 (DF) | χ2 (DF) | NFI | CFI | RMSEA (90% CLs) | BIC | AIC | CAIC | ECVI |
|---|---|---|---|---|---|---|---|---|---|
Likert method: all items (0-1-2-3) | |||||||||
1. One dimension | 20520.61 (66) | 3161.52 (54) | 0.85 | 0.85 | 0.13 (0.12, 0.13) | 3358.74 | 3209.52 | 3382.74 | 0.87 |
2. Three dimensions | 20520.61 (66) | 1061.74 (51) | 0.95 | 0.95 | 0.07 (0.07, 0.07) | 1283.61 | 1115.74 | 1310.61 | 0.30 |
3. Response bias | 20520.61 (66) | 713.26 (39) | 0.97 | 0.97 | 0.07 (0.06, 0.07) | 1033.74 | 791.26 | 1072.74 | 0.21 |
GHQ method: all items (0-0-1-1) | |||||||||
1. One dimension | 19497.47 (66) | 1487.90 (54) | 0.92 | 0.93 | 0.09 (0.08, 0.09) | 1685.12 | 1535.90 | 1709.12 | 0.42 |
2. Three dimensions | 19497.47 (66) | 1162.49 (51) | 0.94 | 0.94 | 0.08 (0.07, 0.08) | 1384.36 | 1216.49 | 1411.36 | 0.33 |
3. Response bias | 19497.47 (66) | 869.10 (39) | 0.96 | 0.96 | 0.08 (0.07,0.08) | 1189.58 | 947.10 | 1228.58 | 0.26 |
C-GHQ method: PP items (0-0-1-1), NP items (0-1-1-1) | |||||||||
1. One dimension | 11954.89 (66) | 4006.40 (54) | 0.68 | 0.68 | 0.16 (0.15, 0.16) | 4203.62 | 5102.66 | 5275.88 | 1.38 |
2. Three dimensions | 11954.89 (66) | 916.14 (51) | 0.95 | 0.95 | 0.07 (0.06, 0.07) | 1138.01 | 917.40 | 1112.27 | 0.25 |
3. Response bias | 11954.89 (66) | 569.00 (39) | 0.97 | 0.97 | 0.05 (0.05,0.06) | 889.48 | 524.84 | 806.33 | 0.14 |