Table 7.
Tests of measurement invariance of the final bifactor-ESEM solution of responses to the MSFWBS.
| χ2 (df) | CFI | TLI | RMSEA | 90% CI | Δχ2 (Δdf) | ΔCFI | ΔTLI | ΔRMSEA | |
|---|---|---|---|---|---|---|---|---|---|
| Age (18–29 vs. Over 30) | |||||||||
| Configural invariance | 656.231 (316)* | 0.950 | 0.906 | 0.070 | 0.062; 0.077 | ||||
| Weak invariance | 689.030 (436)* | 0.963 | 0.949 | 0.051 | 0.044; 0.058 | 246.000 (120) | +0.013 | 0.043 | −0.019 |
| Strong invariance | 732.351 (454)* | 0.959 | 0.946 | 0.053 | 0.046; 0.060 | 46.355 (18) | −0.004 | −0.003 | +0.002 |
| Strict invariance | 782.275 (479)* | 0.956 | 0.944 | 0.053 | 0.047; 0.060 | 80.463 (25)* | −0.003 | −0.002 | 0.000 |
| Latent variance–covariance invariance | 841.122 (501)* | 0.950 | 0.940 | 0.055 | 0.049; 0.062 | 58.015 (22) | −0.006 | −0.004 | +0.002 |
| Latent means invariance | 864.168 (508)* | 0.948 | 0.938 | 0.056 | 0.050; 0.063 | 63.028 (7)* | −0.002 | −0.002 | +0.001 |
| Sex (males vs. females) | |||||||||
| Configural invariance | 670.949 (316)* | 0.949 | 0.902 | 0.070 | 0.063; 0.078 | ||||
| Weak invariance | 752.055 (436)* | 0.954 | 0.937 | 0.057 | 0.050; 0.063 | 258.639 (120) | +0.005 | +0.035 | −0.013 |
| Strong invariance | 777.579 (454)* | 0.953 | 0.938 | 0.056 | 0.049; 0.063 | 30.102 (18)* | −0.001 | +0.001 | −0.001 |
| Strict invariance | 920.369 (479)* | 0.936 | 0.920 | 0.064 | 0.057; 0.070 | 67.648 (25)* | −0.017 | −0.018 | +0.008 |
| Partial strict invariance (free uniq. MM1) | 784.457 (478)* | 0.956 | 0.944 | 0.053 | 0.046; 0.060 | 42.111 (24) | +0.003 | +0.006 | −0.003 |
| Latent variance–covariance invariance | 814.879 (500)* | 0.954 | 0.945 | 0.053 | 0.046; 0.059 | 48.939 (22) | −0.002 | +0.001 | +0.000 |
| Latent means invariance | 850.249 (507)* | 0.950 | 0.941 | 0.055 | 0.048; 0.061 | 28.988 (7)* | −0.004 | −0.004 | +0.002 |
| Personal income (<40 K annually vs. 40 K annually or more) | |||||||||
| Configural invariance | 528.982 (316)* | 0.967 | 0.938 | 0.055 | 0.046; 0.063 | ||||
| Weak invariance | 628.298 (436)* | 0.970 | 0.959 | 0.044 | 0.036; 0.052 | 219.481 (120) | +0.003 | +0.021 | −0.011 |
| Strong invariance | 675.983 (454)* | 0.966 | 0.955 | 0.047 | 0.039; 0.054 | 41.006 (18)* | −0.004 | −0.004 | +0.003 |
| Strict invariance | 757.267 (479)* | 0.957 | 0.946 | 0.051 | 0.044; 0.057 | 111.941 (25)* | −0.009 | −0.009 | +0.004 |
| Latent variance–covariance invariance | 815.240 (501)* | 0.952 | 0.942 | 0.053 | 0.046; 0.059 | 105.489 (22)* | −0.005 | −0.004 | +0.002 |
| Latent means invariance | 854.346 (508)* | 0.947 | 0.937 | 0.055 | 0.049; 0.061 | 52.327 (29) | −0.005 | −0.005 | +0.002 |
| Household income (<90 K annually vs. 90 K annually or more) | |||||||||
| Configural invariance | 508.307 (316)* | 0.970 | 0.943 | 0.052 | 0.044; 0.060 | ||||
| Weak invariance | 648.349 (436)* | 0.967 | 0.955 | 0.047 | 0.039; 0.054 | 245.508 (120) | −0.003 | +0.012 | −0.005 |
| Strong invariance | 664.708 (454)* | 0.967 | 0.957 | 0.045 | 0.038; 0.053 | 12.112 (18) | 0.000 | +0.002 | −0.002 |
| Strict invariance | 767.280 (479)* | 0.955 | 0.944 | 0.052 | 0.045; 0.058 | 147.932 (25)* | −0.012 | −0.013 | +0.007 |
| Partial strict invariance (free uniq. GS7) | 720.704 (478)* | 0.962 | 0.953 | 0.048 | 0.040; 0.054 | 103.063 (24)* | −0.005 | −0.004 | +0.003 |
| Latent variance–covariance invariance | 833.357 (500)* | 0.948 | 0.938 | 0.054 | 0.048; 0.061 | 153.31 (22)* | −0.014 | −0.015 | +0.006 |
| Latent means invariance | 780.625 (485)* | 0.954 | 0.944 | 0.052 | 0.045; 0.059 | 94.868 (8)* | −0.008 | −0.009 | +0.004 |
χ2, robust chi-square test of exact fit; df, degrees of freedom; CFI, comparative fit index; TLI, Tucker-Lewis index; RMSEA, root mean square error of approximation; 90% CI, 90% confidence interval; Δ, Change from the previously retained model.
*
p < 0.01.