Skip to main content

Table 2 CFA-modelling results for latent structure models for MFQ and RCMAS data in adolescents aged 14

From: General and specific components of depression and anxiety in an adolescent population

Estimator robust WLS

Chi Squ. (DF)

df

# parameters

CFI

TLI

RMSEA

WRMR

SSABIC

SSABIC +/-

1 factor model

5345.797

1764

188

0.939

0.937

0.042

1.712

104 332

0

2 factor model aa

4654.275

1758

194

0.951

0.948

0.038

1.565

103 636

- 696

2 factor model bb

5037.228

1763

189

0.944

0.942

0.040

1.647

103 839

-493

3 factor model

4083.833

1752

200

0.960

0.958

0.034

1.424

102 753

- 1579

   - With gender as covariate

4248.097

1810

203

0.957

0.955

0.034

1.450

104 249

-83

   - Correction for differential item functioning

4135.5731

1808

205

0.959

0.957

0.033

1.425

104 073

-259

Estimator robust WLS

Chi Squ. (DF)

df

# parameters

CFI

TLI

RMSEA

WRMR

SSABIC

SSABIC +/-

Bifactor model

3839.960

1724

228

0.964

0.962

0.033

1.350

102 077

-2255

   - With gender as covariate

3951.343

1781

232

0.961

0.959

0.032

1.367

104 651c

+319

   - Correction for differential item functioning

4083.833

1752

233

0.960

0.958

0.034

1.424

  
  1. aanxiety/depression and somatic factor, following EFA.
  2. btwo-factor model with MFQ items on one factor and RCMAS items on the other factor.
  3. ccomputing the ssaBIC for the bifactor model with adjustments for DIF was computationally unmanageable with MLR.
  4. SSABIC = Sample-size adjusted Bayesian information criterion.
  5. Robust WLS is WLSMV in Mplus i.e. robust Weighted Least Squares for categorical data, mean and variance adjusted.
  6. Robust ML is MLMV in Mplus i.e. Maximum Likelihood covariance structure analysis, mean and variance adjusted.