suggested acceptable fit, and low CFI and TLI suggested poorer fit, one might reject both of these models.
As can be seen in Table 2, both the one- and two-factor correlated error models had improved model fit over the previous models considered, with values of RMSEA
and SRMR falling within the recommended criteria for good fit, and CFI falling within the criteria of acceptable fit.
Os resultados indicaram um melhor ajuste para este ultimo modelo, sugerindo o controle do efeito da redacao dos itens (CFI = 0,94 e RMSEA
909 competence in career with correlation between all factors (6 situations + 2 performances) Model PCFI GFI PGFI RMSEA
1--General model (6 0.
2] df CFI RMSEA
90% CI SRMR Baseline model Women 141.
Los indicadores de ajuste se mostraron adecuados en la muestra de adictos ([ji al cuadrado] = 585,65; p< 0,001; CMIN/DF = 2,2; RMSEA
= 0,07; CFI = 0,99; PCFI = 0,82) y en la de poblacion no clinica ([ji al cuadrado] = 615,97; p< 0,001; CMIN/DF = 2,3; RMSEA
= 0,06; CFI = 0,99; PCFI = 0,82).
The analysis showed a similar goodness-of-fit of the hypothetical model to the data of all 5 countries: RMSEA
was less than .
86) if compared with the original MCA model, and the RMSEA
value increased to .
assesses lack of fit based upon model misspecification and provides a measure of this discrepancy per degree of freedom (Browne and Cudeck, 1993).
In determining acceptable values for the three indices, we used the following guidelines (as cited by Hu & Bentler): For SRMR and RMSEA
, less than .
However, the fit indices did not change significantly, the RMSEA
value was well within the acceptable range, and there was a reduction in the MC value in the model with equality constraints.
So, we used fit indices, such as chi-square, TLI (Turket-Lewis Index), CFI (Comparative Fit Index), SRMR (Standardized Root-Mean-Square Residual), and RMSEA
(Root Mean-Mean-Square Error of Approximation).