Table 3.
Model fitting results for testing MI of the Harvard Trauma Questionnaire across five linguistic groups.
| vs. | χ2 | df | ∆χ2 | ∆df | χ2/df | CFI | ∆CFI | TLI | RMSEA | |
|---|---|---|---|---|---|---|---|---|---|---|
| 1. Configural MI: total sample | – | 1072.991 | 505 | – | – | – | .962 | – | .955 | .067 |
| 1a. Configural MI: Indo-Iranian languages | – | 220.150 | 101 | – | – | – | .950 | – | .941 | .067 |
| 1b. Configural MI: Niger-Congo languages | – | 119.752 | 101 | – | – | – | .985 | – | .982 | .038 |
| 1c. Configural MI: Semitic languages | – | 219.875 | 101 | – | – | – | .922 | – | .908 | .064 |
| 1d. Configural MI: South Slavic languages | – | 243.923 | 101 | – | – | – | .924 | – | .908 | .085 |
| 1e. Configural MI: Germanic languages | – | 290.933 | 101 | – | – | – | .976 | – | .972 | .071 |
| 2. Strong MI | 1 | 1429.417 | 673 | 414.764 | 168 | 2.469 | .949 | .013 | .955 | .067 |
| 3. Partial strong MI | 1 | 1189.434 | 631 | 188.134 | 126 | 1.493 | .963 | −.001 | .964 | .060 |
| 4. Partial strict MI | 3 | 1303.349 | 679 | 146.734 | 48 | 3.057 | .958 | .005 | .963 | .061 |
Best fitting model is printed in bold; vs. = versus; χ2, df = chi-square test statistic and degrees of freedom for model; ∆χ2, ∆df = chi-square test statistic and degrees of freedom for chi-square difference test between two nested models; χ2/df = ratio between χ2 and degrees of freedom with regard to the chi-square difference test; ∆CFI = difference in CFI value between two nested models.