Table 5.
Parameter Estimates for the Two-Group Single Factor Model.
| Item | Adolescents (N=2769) | Parents (N = 2619) | Likelihood Ratio Test χ2 (p) |
|---|---|---|---|
| Loadings (SE) | |||
| AB | .90 (.02) | .87 (.03) | 0.41 (.52) |
| CD | .74 (.03) | .66 (.04) | 1.45 (.23) |
| AD | .84 (.03) | .70 (.03) | 5.53 (.02)* |
| MD | .86 (.03) | .79 (.04) | 1.33 (.25) |
| DD | .91 (.04) | .88 (.05) | 0.13 (.72) |
| Thresholds (SE) | |||
| AB | 1.69 (.05) | 1.56 (.04) | .99 (.32) |
| CD | 1.09 (.04) | 1.28 (.03) | .33 (.57) |
| AD | 1.39 (.04) | 0.75 (.03) | 33.47 (.00)* |
| MD | 1.57 (.05) | 1.57 (.04) | 1.24 (.27) |
| DD | 2.59 (.10) | 2.38 (.08) | .37 (.54) |
Note:
Signficant by conventional levels. Parameters were estimated in a model that allowed all parameters to freely vary. The likelihood ratio test was conducted between the fully free model and a submodel where two parameters were constrained to be equal across groups (e.g., the loadings for AB in the parent and adolescents were constrained to be equal). AB = Adult Antisocial Behavior; CD = Conduct Disorder; AD = Alcohol Dependence; MD = Marijuana Dependence; DD = Drug Dependence. Loadings are regression slopes in a regression of the diagnosis onto the latent trait. Thresholds refer to the cutting score on a normal distribution required to yield the binary diagnostic measure. For example, the threshold on DD is 2.59 in the adolescent sample, which corresponds to the 99.5th percentile of a normal distribution.