Table 5. Fit Indices and Comparisons of CFA Invariance Models (32-countries group and country clusters).
Model | SB- χ2 [df] | CFI | TLI | RMSEA | CI90-RMSEA | SRMR | ΔCFI | ΔTLI | ΔRMSEA | ΔSRMR |
---|---|---|---|---|---|---|---|---|---|---|
Total TALIS 2013 sample | ||||||||||
Configural invariance | 30,275.3 [1,632]* | .928 | .907 | .058 | [.058, .059] | .049 | – | – | – | – |
Metric invariance | 33,222.1 [1,911]* | .922 | .913 | .056 | [.056, .057] | .058 | –.006 | +.006 | –.002 | +.009 |
Scalar invariance | 63,041.3 [2,190]* | .848 | .853 | .073 | [.073, .074] | .082 | –.074 | –.060 | +.017 | +.024 |
Strict invariance | 79,986.0 [2,562]* | .806 | .840 | .077 | [.076, .077] | .130 | –.042 | –.013 | +.004 | +.047 |
East and South-East Asian countries (Japan, Korea, Malaysia, Singapore) | ||||||||||
Configural invariance | 6,239.1 [204]* | .927 | .905 | .078 | [.076, .079] | .050 | – | – | – | – |
Metric invariance | 6,730.3 [231]* | .921 | .910 | .076 | [.074, .077] | .056 | –.006 | +.005 | –.002 | +.006 |
Partial scalar invariance | 7,556.9 [249]* | .911 | .906 | .078 | [.076, .079] | .057 | –.010 | –.004 | +.002 | +.001 |
Scalar invariance | 10,371.3 [258]* | .877 | .874 | .090 | [.088, .091] | .070 | –.034 | –.032 | +.012 | +.017 |
Strict invariance | 11,267.8 [294]* | .866 | .880 | .087 | [.086, .089] | .091 | –.011 | –.006 | –.003 | +.021 |
Anglo-Saxon countries (Australia, England, Unites States of America) | ||||||||||
Configural invariance | 2,421.5 [153]* | .927 | .906 | .065 | [.063, .067] | .053 | – | – | – | – |
Metric invariance | 2,507.7 [171]* | .925 | .913 | .063 | [.060, .065] | .054 | –.002 | +.007 | –.002 | +.001 |
Partial scalar invariance | 2,602.7 [183]* | .922 | .916 | .062 | [.059, .064] | .056 | –.003 | +.003 | –.001 | +.002 |
Scalar invariance | 2,752.2 [189]* | .918 | .914 | .062 | [.060, .064] | .059 | –.004 | –.002 | .000 | +.003 |
Strict invariance | 2,783.4 [213]* | .917 | .923 | .059 | [.057, .061] | .063 | –.001 | +.009 | –.003 | +.004 |
Nordic countries (Denmark, Finland, Norway, Sweden) | ||||||||||
Configural invariance | 5,041.0 [204]* | .927 | .905 | .059 | [.058, .061] | .051 | – | – | – | – |
Metric invariance | 5,401.5 [231]* | .922 | .910 | .058 | [.056, .059] | .057 | –.005 | –.005 | –.001 | +.006 |
Partial scalar invariance | 6,619.1 [249]* | .903 | .897 | .062 | [.061, .063] | .062 | –.019 | –.013 | +.004 | +.005 |
Scalar invariance | 9,645.8 [258]* | .857 | .854 | .074 | [.072, .075] | .076 | –.046 | –.043 | +.012 | +.015 |
Strict invariance | 10,905.8 [294]* | .839 | .855 | .073 | [.072, .075] | .105 | –.018 | +.001 | –.001 | +.029 |
Note. SB- χ2 = Satorra-Bentler corrected χ2 value. CI90-RMSEA = 90% confidence interval of the RMSEA. For the total TALIS 2013 sample, partial scalar was not tested due to a large number of possible combinations that could be used to constrain some of the item intercepts across the 32 participating countries.
* p < .01.