Table 1 Summary of external validity analyses of ODD factor scores from best-fitting bifactor model

From: External validation of a bifactor model of oppositional defiant disorder

Dependent variable

  

Covariates

βGeneral

βIrr

βDef

R2

χ2

df

p

BIC=SEX

BICbySEX

RMSEA (95% CI)

TLI

RMSR

Inattention

 Sex, Age, Age2, Sex × Age

   

.24

.08

.27

.23

7.3

3

.062

12,015

12,027

.041 (.000−.079)

.963

.016

 

95% CIs

[.20−.30]

[−.09 to .25]

[.10−.44]

         

Hyperactivity−impulsivity

 Sex, Age

   

.24

−.06

.48

.29

9.4

3

.024

11,367

11,375

.049 (.016−.087)

.964

.026

 

95% CIs

[.19−.28]

[-.23 to .11]

[.32−.65]

         

Conduct disorder

 Sex, Age

   

.28

−.07

.53

.36

5.2

3

.157

9730

9736

.029 (.000−.070)

.978

.024

 

95% CIs

[.23−.33]

[-.30 to .15]

[.31−.76]

         

Depression/dysthymia

 Age

   

.18

.35

−.04

.17

1.7

3

.648

10,535

10,555

.000 (.000−.045)

1.011

.010

 

95% CIs

[.12−.23]

[.18−.52]

[-.20−.13]

         

Generalized anxiety disorder

 Sex, Age, Age2

   

.20

.40

−.13

.14

11.0

3

.012

11,961

11,967

.055 (.023−.092)

.898

.022

 

95% CIs

[.14-.26]

[.22-.57]

[-.30-.04]

         

Irritability not

 Equated by sex

    

.33/.45

.14/.22

1.2

2

.551

11,954

 

.000 (.000−.058)

1.015

.002

 

95% CIs

 

[.14−.51]/ [.28−.61]

          
  1. The β’s are the standardized regression coefficients for the regression of each dependent variable on each of the three ODD factors. R2 is the % of variance accounted for in each dependent variable by all three ODD factors. The fit statistics are for the comparison of a model in which the standardized regression coefficients for the three ODD factors are equated for boys and girls versus a model in which the standardized regression coefficients for the three ODD factors vary by sex. For all of the dependent variables except for GAD and hyperactivity−impulsivity, all three regression coefficients could be equated for boys and girls. For GAD, a model in which the regression coefficient for ODDIrr was larger for girls than for boys fits better, as indicated in the second row of results for GAD. For hyperactivity−impulsivity, models in which the regression coefficients for ODDIrr and ODDDef varied by sex fit better than one in which all coefficients were equated but these are not shown due to space and due to the possibility that these models fit better simply due to chance
  2. Regression coefficients in bold significantly differ from zero
Back to article page