Table 6.
Overall main effects for multinomial logistic regression of trajectory group membership on case characteristics and social context variables
| Effect | χ2 | p |
|---|---|---|
| Age | 12.74 | <.013* |
| Ethnicity | 16.45 | <.036* |
| Father arrested/jailed | 15.13 | <.004* |
| Mother arrested/jailed | 3.93 | <.416 |
| IQ | 3.64 | <.457 |
| Antisocial history | 53.33 | <.001* |
| Substance use problem | 2.81 | <.591 |
| Mood/anxiety problem | 3.98 | <.409 |
| Legal cynicism | 2.00 | <.736 |
| Legitimacy | 7.79 | <.100 |
| Moral disengagement | 0.50 | <.973 |
| Temperance | 18.32 | <.001* |
| Responsibility | 4.74 | <.316 |
| Perspective | 17.97 | <.001* |
| Neighborhood disadvantage | 3.57 | <.468 |
| Parental | ||
| Monitoring | 7.71 | <.103 |
| Knowledge | 3.22 | <.523 |
| Warmth | 4.29 | <.368 |
| Peers | ||
| Behavior | 54.48 | <.001* |
| Attitudes | 5.84 | <.212 |
| Level of alcohol use | 5.02 | <.286 |
| Frequency of substance use | 34.77 | <.001* |
Note: The chi-square values are computed as −2(LUR – LR) ~χ2, where LUR and LR are the maximized log-likelihoods of the full and restricted models, respectively. All tests have 4 degrees of freedom.
*
p ≤ .05.