Skip to main content

Table 3 Fit statistics for the proposed models of the HADS

From: A confirmatory bifactor analysis of the hospital anxiety and depression scale in an Italian community sample

  χ2 df p SB χ2 p χ2/df SB χ2/df *AIC *CFI *NNFI *RMSEA (90% CI)
1. Razavi et al. (1990) [30] 1259.54 91 <.001 1052.04 <.001 13.84 11.56 896.04 .84 .81 .09 (.08-.09)
2. Zigmond & Snaith (1983) [1] 476.03 76 <.001 397.35 <.001 6.26 5.23 245.35 .95 .94 .05 (.05-.06)
3. Moorey et al. (1991) [21] 447.93 76 <.001 373.17 <.001 5.89 4.91 221.17 .95 .94 .05 (.05-.06)
4. Dunbar et al. (2000) [7] 472.92 74 <.001 393.72 <.001 6.39 5.32 245.72 .95 .94 .05 (.05-.06)
5. Friedman et al. (2001) [10] 462.77 74 <.001 386.23 <.001 6.25 5.22 238.23 .95 .94 .05 (.05-.06)
6. Caci et al. (2003) [13] 471.71 74 <.001 394.36 <.001 6.37 5.33 246.36 .95 .94 .05 (.05-.06)
7. Bifactor 238.53 63 <.001 200.21 <.001 3.79 3.18 74.21 .98 .97 .04 (.03-.04)
  1. Note: SBχ2 = Satorra-Bentler chi-square; *= robust statistics.
  2. Recommendations and cut-off points for model fit [33, 34]: Low χ2 relative to degrees of freedom with an insignificant p value (p > 0.05); relative/normed chi-square (χ2/df) smaller than .05; smaller values of AIC; CFI and NNFI equal to or greater than .95; RMSEA equal to or less than .08; 90% confidence intervals for the RMSEA close to RMSEA.