Skip to main content

Table 4 Goodness-of-fit indices for the univariate latent growth curve models, the parallel process models and the latent difference score models

From: One year follow-up and mediation in cognitive behavioral therapy and acceptance and commitment therapy for adult depression

Model X2 (df) RMSEA CFI BIC
QIDS-SR
 Intercept-only 234.56 (19)*** 0.37 0.16 2451.88
 Linear change 17.03 (11) 0.08 0.98 2244.37
 Curvilinear change 14.59 (7)* 0.12 0.97 2246.94
 Multigroup LGCM ≠a 30.01 (24) 0.08 0.98 2241.28
 Multigroup LGCM =b 30.54 (25) 0.07 0.98 2240.56
DAS-17
 Intercept-only 102.37 (19)*** 0.23 0.72 2826.02
 Linear changec 29.01 (16)* 0.10 0.96 2756.41
 Curvilinear change 17.25 (12) 0.07 0.98 2749.66
 Multigroup LGCM ≠a 35.27 (27) 0.09 0.97 2741.72
 Multigroup LGCM =b 35.31 (29) 0.07 0.98 2739.26
 Parallel process modeld 102.02 (55)*** 0.10 0.93 4956.86
 Latent difference score model 55.43 (35)* 0.08 0.97 4935.32
 Latent difference score model multigroup ≠e 236.53 (70)*** 0.24 0.79 4946.48
 Latent difference score model multigroup =f 241.21 (75)*** 0.23 0.79 4944.90
EQ_D
 Intercept-only 109.26 (19)*** 0.24 0.59 2556.74
 Linear change 45.46 (11)*** 0.20 0.84 2502.95
 Curvilinear changeg 18.11 (9)* 0.11 0.96 2478.11
 Multigroup LGCM ≠a 60.20 (24)*** 0.19 0.86 2480.85
 Multigroup LGCM =b 60.30 (26)*** 0.18 0.86 2478.44
 Parallel process model 80.23 (50)*** 0.09 0.95 4667.48
 Latent difference score model 78.39 (35)*** 0.12 0.92 4684.43
 Latent difference score model multigroup ≠ 178.44 (70)*** 0.19 0.83 4665.22
 Latent difference score model multigroup = 186.54 (75)*** 0.19 0.83 4667.06
AAQ-II
 Intercept-only 91.46 (19)*** 0.22 0.63 2799.96
 Linear change 12.08 (11) 0.04 0.99 2730.59
 Curvilinear change 7.52 (7) 0.03 0.99 2731.05
 Multigroup LGCM ≠a 42.77 (24)* 0.14 0.92 2724.58
 Multigroup LGCM =b 42.78 (25)* 0.13 0.92 2723.33
 Parallel process modelh 67.66 (54) 0.06 0.98 4892.43
 Latent difference score model 93.65 (35)*** 0.14 0.90 4942.23
 Latent difference score model multigroup ≠ 172.53 (70)*** 0.19 0.84 4917.40
 Latent difference score model multigroup = 191.90 (75)*** 0.20 0.81 4930.50
  1. Note. QIDS-SR Quick Inventory for Depressive Symptomatology Self-Rated, DAS-17 Dysfunctional Attitude Scale-revised, EQ-D Decentering subscale of the Experiences Questionnaire, AAQ-II Acceptance and Action Questionnaire-II, RMSEA root mean square error of approximation, CFI Comparative Fit Index, BIC Bayesian Information Criterion, LGCM latent growth curve model. * = p < .05, *** = p < .001
  2. aIn this model a multigroup approach was applied to the best fitting univariate LGCM while freely estimating the slope factor means
  3. bIn this model a multigroup approach was applied to the best fitting univariate LGCM while an equality constraint was imposed on the slope factor means
  4. ccorrelated measurement residuals among adjacent time points were not included for these linear and curvilinear models as this provided a superior fit
  5. dthe curvilinear slope in DAS-17 scores was removed from this model, as a negative residual variance was obtained for this parameter
  6. eIn this model a multigroup approach was applied to the latent difference score model while freely estimating the regression pathway linking the latent difference score of the mediator to QIDS-SR scores at the end of the interval
  7. fIn this model a multigroup approach was applied to the latent difference score model while an equality constraint was imposed on the regression pathway linking the latent difference score of the mediator to QIDS-SR scores at the end of the interval
  8. gthe correlation between pre-treatment and session 1 scores was negative in this model and was constrained to 0
  9. h the curvilinear slope in AAQ-II scores was removed from this model, as a negative residual variance was obtained for this parameter
Back to article page

BMC Psychiatry

ISSN: 1471-244X

Contact us