Table 3.
Model comparisons for measurement invariance testing by sex for the multigroup models in each country
| Model | Description | LL | Npar | SC | Comparison | df | LRTS | p value |
|---|---|---|---|---|---|---|---|---|
| DRC | ||||||||
| 1.0 | Fully unconstrained | −9687.62 | 87 | 1.10 | – | – | – | – |
| 1.1 | Item-response probabilities constrained | −9751.44 | 47 | 1.12 | 1.1 vs. 1.0 | 40 | 119.37 | <0.001 |
| 2.1 | Self-blame (1) constrained | −9695.10 | 83 | 1.11 | 2.1 vs. 1.0 | 4 | 18.84 | 0.001 |
| 2.2 | Worry (2) constrained | −9692.35 | 83 | 1.11 | 2.2 vs. 1.0 | 4 | 11.36 | 0.023 |
| 2.3 | Can’t sleep (3) constrained | −9693.01 | 83 | 1.07 | 2.3 vs. 1.0 | 4 | 6.20 | 0.185 |
| 2.4 | Feel sad (4) constrained | −9692.26 | 83 | 1.11 | 2.4 vs. 1.0 | 4 | 10.40 | 0.034 |
| 2.5 | Think of self-harm (5) constrained | −9689.02 | 83 | 1.10 | 2.5 vs. 1.0 | 4 | 2.73 | 0.605 |
| 2.6 | Bullied/threatened (6) constrained | −9690.00 | 83 | 1.12 | 2.6 vs. 1.0 | 4 | 8.55 | 0.073 |
| 2.7 | Slapped/hit/hurt (7) constrained | −9690.39 | 83 | 1.11 | 2.7 vs. 1.0 | 4 | 7.06 | 0.132 |
| 2.8 | Been teased/called names (8) constrained | −9698.87 | 83 | 1.15 | 2.8 vs. 1.0 | 4 | 214.67 | <0.001 |
| 2.9 | Been slapped/hit/hurt (9) constrained | −9703.14 | 83 | 1.10 | 2.9 vs. 1.0 | 4 | 28.36 | <0.001 |
| 2.10 | Substance use (10) constrained | −9691.51 | 83 | 1.11 | 2.10 vs. 1.0 | 4 | 9.20 | 0.056 |
| 3.1 | Indicators 3 and 5 constrained | −9693.23 | 79 | 1.07 | 3.1 vs. 1.0 | 8 | 0.442 | 7.91 |
| 3.2 | Indicators 3, 5, and 6 constrained | −9695.95 | 75 | 1.09 | 3.2 vs. 1.0 | 12 | 14.55 | 0.267 |
| 3.3 | Indicators 3, 5, 6, and 7 constrained | −9698.19 | 71 | 1.08 | 3.3 vs. 1.0 | 16 | 17.88 | 0.331 |
| 3.4 | Indicators 3, 5, 6, 7, and 10 constrained | −9700.95 | 67 | 1.09 | 3.4 vs. 1.0 | 20 | 23.89 | 0.247 |
| 4.1 | Class prevalences constrained | −9713.69 | 64 | 1.14 | 5.1 vs. 3.4 | 3 | 148.13 | <0.001 |
| Malawi | ||||||||
| 1.0 | Fully unconstrained | −12,031.75 | 87 | 1.17 | – | – | – | – |
| 1.1 | Item-response probabilities constrained | −12,076.94 | 47 | 1.42 | 1.1 vs. 1.0 | 40 | 102.36 | <0.001 |
| 2.1 | Self-blame (1) constrained | −12,036.33 | 83 | 1.21 | 2.1 vs. 1.0 | 4 | 22.90 | <0.001 |
| 2.2 | Worry (2) constrained | −12,035.79 | 83 | 1.22 | 2.2 vs. 1.0 | 4 | 50.67 | <0.001 |
| 2.3 | Can’t sleep (3) constrained | −12,034.79 | 83 | 1.23 | 2.3 vs. 1.0 | 4 | 331.57 | <0.001 |
| 2.4 | Feel sad (4) constrained | −12,033.10 | 83 | 1.17 | 2.4 vs. 1.0 | 4 | 2.16 | 0.706 |
| 2.5 | Think of self-harm (5) constrained | −12,032.67 | 83 | 1.16 | 2.5 vs. 1.0 | 4 | 1.20 | 0.879 |
| 2.6 | Bullied/threatened (6) constrained | −12,040.52 | 83 | 1.28 | 2.6 vs. 1.0 | 4 | −16.79 | – |
| 2.7 | Slapped/hit/hurt (7) constrained | −12,035.80 | 83 | 1.24 | 2.7 vs. 1.0 | 4 | −56.41 | – |
| 2.8 | Been teased/called names (8) constrained | −12,038.80 | 83 | 1.32 | 2.8 vs. 1.0 | 4 | −7.57 | – |
| 2.9 | Been slapped/hit/hurt (9) constrained | −12,037.17 | 83 | 1.29 | 2.9 vs. 1.0 | 4 | −8.72 | – |
| 2.10 | Substance use (10) constrained | −12,051.49 | 83 | 1.41 | 2.10 vs. 1.0 | 4 | −10.59 | – |
| 3.1 | Indicators 4 and 5 constrained | −12,033.83 | 79 | 1.17 | 3.1 vs. 1.0 | 8 | 3.29 | 0.915 |
| 3.2 | Indicators 3, 4, and 5 constrained | −12,038.33 | 75 | 1.22 | 3.2 vs. 1.0 | 12 | 15.06 | 0.238 |
| 3.3 | Indicators 2, 3, 4, and 5 constrained | −12,042.70 | 71 | 1.21 | 3.3 vs. 1.0 | 16 | 21.22 | 0.170 |
| 3.4 | Indicators 1, 2, 3, 4, and 5 constrained | −12,047.10 | 67 | 1.20 | 3.4 vs. 1.0 | 20 | 28.28 | 0.103 |
| 4.1 | Class prevalences constrained | −12,052.26 | 64 | 1.26 | 5.1 vs. 3.4 | 3 | −89.10 | – |
| Indonesia | ||||||||
| 1.0 | Fully unconstrained | −23,636.94 | 87 | 1.60 | – | – | – | – |
| 1.1 | Item-response probabilities constrained | −23,871.07 | 47 | 1.93 | 1.1 vs. 1.0 | 40 | 386.82 | <0.001 |
| 2.1 | Self-blame (1) constrained | −23,640.23 | 83 | 1.61 | 2.1 vs. 1.0 | 4 | 4.71 | 0.318 |
| 2.2 | Worry (2) constrained | −23,639.40 | 83 | 1.61 | 2.2 vs. 1.0 | 4 | 3.46 | 0.485 |
| 2.3 | Can’t sleep (3) constrained | −23,653.48 | 83 | 1.65 | 2.3 vs. 1.0 | 4 | 62.36 | <0.001 |
| 2.4 | Feel sad (4) constrained | −23,640.24 | 83 | 1.62 | 2.4 vs. 1.0 | 4 | 5.88 | 0.209 |
| 2.5 | Think of self-harm (5) constrained | −23,653.61 | 83 | 1.63 | 2.5 vs. 1.0 | 4 | 33.93 | <0.001 |
| 2.6 | Bullied/threatened (6) constrained | −23,647.95 | 83 | 1.60 | 2.6 vs. 1.0 | 4 | 14.43 | 0.006 |
| 2.7 | Slapped/hit/hurt (7) constrained | −23,639.87 | 83 | 1.58 | 2.7 vs. 1.0 | 4 | 2.93 | 0.570 |
| 2.8 | Been teased/called names (8) constrained | −23,642.67 | 83 | 1.61 | 2.8 vs. 1.0 | 4 | 8.32 | 0.080 |
| 2.9 | Been slapped/hit/hurt (9) constrained | −23,658.37 | 83 | 1.62 | 2.9 vs. 1.0 | 4 | 36.72 | <0.001 |
| 2.10 | Substance use (10) constrained | −23,784.29 | 83 | 1.72 | 2.10 vs. 1.0 | 4 | −346.95 | – |
| 3.1 | Indicators 1 and 2 constrained | −23,644.50 | 79 | 1.62 | 3.1 vs. 1.0 | 8 | 11.14 | 0.194 |
| 3.2 | Indicators 1, 2, and 4 constrained | −23,648.42 | 75 | 1.65 | 3.2 vs. 1.0 | 12 | 17.99 | 0.116 |
| 3.3 | Indicators 1, 2, 4, and 7 constrained | −23,650.98 | 71 | 1.61 | 3.3 vs. 1.0 | 16 | 18.18 | 0.314 |
| 3.4 | Items 1, 2, 4, 7, and 8 constrained | −23,659.91 | 67 | 1.64 | 3.4 vs. 1.0 | 20 | 31.15 | 0.053 |
| 4.1 | Class prevalences constrained | −23,710.47 | 64 | 1.65 | 5.1 vs. 3.4 | 3 | 73.95 | <0.001 |
| China | ||||||||
| 1.0 | Fully unconstrained | −8783.53 | 65 | 1.22 | – | – | – | – |
| 1.1 | Item-response probabilities constrained | −8826.64 | 35 | 1.33 | 1.1 vs. 1.0 | 30 | 78.86 | <0.001 |
| 2.1 | Self-blame (1) constrained | −8784.12 | 62 | 1.24 | 2.1 vs. 1.0 | 3 | 1.59 | 0.661 |
| 2.2 | Worry (2) constrained | −8787.22 | 62 | 1.28 | 2.2 vs. 1.0 | 3 | −105.30 | – |
| 2.3 | Can’t sleep (3) constrained | −8784.22 | 62 | 1.26 | 2.3 vs. 1.0 | 3 | 3.24 | 0.356 |
| 2.4 | Feel sad (4) constrained | −8785.16 | 62 | 1.26 | 2.4 vs. 1.0 | 3 | 9.48 | 0.024 |
| 2.5 | Think of self-harm (5) constrained | −8791.79 | 62 | 1.29 | 2.5 vs. 1.0 | 3 | −86.04 | – |
| 2.6 | Bullied/threatened (6) constrained | −8784.19 | 62 | 1.25 | 2.6 vs. 1.0 | 3 | 1.98 | 0.576 |
| 2.7 | Slapped/hit/hurt (7) constrained | −8788.86 | 62 | 1.13 | 2.7 vs. 1.0 | 3 | 3.49 | 0.322 |
| 2.8 | Been teased/called names (8) constrained | −8790.78 | 62 | 1.24 | 2.8 vs. 1.0 | 3 | 16.74 | 0.001 |
| 2.9 | Been slapped/hit/hurt (9) constrained | −8798.40 | 62 | 1.26 | 2.9 vs. 1.0 | 3 | 80.39 | <0.001 |
| 2.10 | Substance use (10) constrained | −8784.00 | 62 | 1.17 | 2.10 vs. 1.0 | 3 | 0.43 | 0.934 |
| 3.1 | Indicators 1 and 3 constrained | −8784.77 | 59 | 1.30 | 3.1 vs. 1.0 | 6 | 5.28 | 0.509 |
| 3.2 | Indicators 1, 3, and 6 constrained | −8785.56 | 56 | 1.32 | 3.2 vs. 1.0 | 9 | 6.56 | 0.683 |
| 3.3 | Indicators 1, 3, 6, and 7 constrained | −8792.75 | 53 | 1.21 | 3.3 vs. 1.0 | 12 | 14.75 | 0.255 |
| 3.4 | Indicators 1, 3, 6, 7, and 10 constrained | −8793.17 | 50 | 1.15 | 3.4 vs. 1.0 | 15 | 13.44 | 0.569 |
| 3.5 | Indicators 1, 3, 4, 6, 7, and 10 constrained | −8794.72 | 47 | 1.20 | 3.5 vs. 1.0 | 18 | 17.54 | 0.486 |
| 3.6 | Indicators 1, 2, 3, 4, 6, 7, and 10 constrained | −8798.51 | 44 | 1.25 | 3.6 vs. 1.0 | 21 | 25.72 | 0.217 |
| 4.1 | Class prevalences constrained | −8818.20 | 42 | 1.13 | 4.1 vs. 3.6 | 2 | 10.46 | 0.005 |
Note. LL = log likelihood; Npar = number of parameters; SC = scaling correction factor; LRTS = likelihood ratio test statistic. Bold indicates best-fitting model as suggested by the LRTS. No p values reported for model comparisons with a negative LRTS.