Table 2.
Coefficient tested | CFI | TLI | RMSEA | Chi-Square Test for Difference Testing |
|
---|---|---|---|---|---|
Value | DF | ||||
Research expedient | |||||
Young on academic standing | 0.989 | 0.985 | 0.031 | 3.689 | 1 |
Only on substance use | 0.986 | 0.981 | 0.034 | 9.755 * | 1 |
Middle on substance use | 0.991 | 0.988 | 0.027 | 1.126 | 1 |
Young on substance use | 0.992 | 0.989 | 0.026 | 0.514 | 1 |
Model with combined constraintsa | 0.989 | 0.986 | 0.029 | 5.935 | 3 |
Full non invariant model | 0.991 | 0.988 | 0.027 | ||
Adler’s | |||||
Second on academic standing | 0.988 | 0.984 | 0.030 | 3.104 | 1 |
Young on academic standing | 0.987 | 0.982 | 0.032 | 4.722 * | 1 |
Only on substance use | 0.987 | 0.982 | 0.031 | 4.993 * | 1 |
Second on substance use | 0.991 | 0.988 | 0.026 | 0.000 | 1 |
Middle on substance use | 0.990 | 0.987 | 0.027 | 0.534 | 1 |
Young on substance use | 0.991 | 0.988 | 0.026 | 0.024 | 1 |
Model with combined constraintsb | 0.990 | 0.987 | 0.026 | 3.233 | 4 |
Full non invariant model | 0.990 | 0.986 | 0.028 | ||
Family size | |||||
Young on academic standing | 0.985 | 0.982 | 0.028 | 3.805 | 1 |
Only on substance use | 0.986 | 0.983 | 0.028 | 3.902 * | 1 |
Second small on substance use | 0.989 | 0.987 | 0.025 | 0.457 | 1 |
Second/med/large on substance use | 0.988 | 0.986 | 0.025 | 1.060 | 1 |
Third small/med/large on substance use | 0.989 | 0.986 | 0.025 | 0.778 | 1 |
Young on substance use | 0.989 | 0.987 | 0.024 | 0.005 | 1 |
Model with combined constraintsc | 0.986 | 0.983 | 0.028 | 6.990 | 5 |
Full non invariant model | 0.988 | 0.986 | 0.025 |
p<0.05.
Constraining young on academic standing, middle, and young on substance use.
Constraining second on academic standing, second, middle and young on substance use.
Constraining all variables tested except only on substance use.
Notes. All models were nested in and compared to the full non-invariant model. All constraints in coefficients are set to be equal coefficient for male and female adolescents.