Table 5.
Predictive Regression of Wave 2 Problem Behaviors
| Youth Behaviors | Antisocial | Violence | Substance Use |
|---|---|---|---|
| AUC (Model 3/Model 4) | 0.800/.804 | 0.820/.835 | 0.462/.468 |
| Race/ethnic groups | |||
| White | −0.040 (0.96) | −0.133 (0.88) | −0.045 (0.076) |
| Latino | −0.129 (0.88) | 0.233 (1.26) | 0.080 (0.106) |
| Asian | −0.110 (0.90) | 0.105 (1.11) | −0.061 (0.102) |
| Control variables | |||
| Female | −0.639 (0.53)*** | −0.635 (0.53) | 0.082 (0.045) |
| Age | −0.162 (0.85)* | −0.976 (0.91) | 0.018 (0.022) |
| Single parent | 0.128 (1.14) | 0.270 (1.31) | 0.013 (0.065) |
| Low income | −0.506 (0.60) | 0.201 (1.22) | 0.049 (0.101) |
| Second generation | 0.617 (1.85) | −0.019 (0.98) | 0.049 (0.084) |
| Nonimmigrant | 0.715 (2.04)* | 0.333 (1.39) | 0.090 (0.084) |
| Prior behaviors | 1.445 (4.24)*** | 2.493 (12.10)*** | 0.414*** (0.030) |
| Peer risks | |||
| Yield to pressure | 0.754 (2.13)*** | 0.528 (1.70)* | 0.528*** (0.092) |
| Antisocial reward | 0.025 (1.03) | 0.170 (1.18) | 0.002 (0.030) |
| Substance-using friends | 0.354 (1.42)*** | 0.384 (1.47)** | 0.205***(0.038) |
| Antisocial friends | 0.152 (1.16) | −0.006 (0.99) | −0.079 (0.081) |
| White X Yield to pressure | n.s. | n.s. | n.s. |
| Latino X Yield to pressure | n.s. | n.s. | n.s. |
| Asian X Yield to pressure | n.s. | n.s. | n.s. |
| White X Antisocial reward | n.s. | n.s. | n.s. |
| Latino X Antisocial reward | n.s. | n.s. | n.s. |
| Asian X Antisocial reward | n.s. | n.s. | n.s. |
| White X Substance-using friends | n.s. | n.s. | −0.230* |
| Latino X Substance-using friends | n.s. | n.s. | n.s. |
| Asian X Substance-using friends | n.s. | n.s. | n.s. |
| White X Antisocial friends | n.s. | n.s. | n.s. |
| Latino X Antisocial friends | n.s. | n.s. | n.s. |
| Asian X Antisocial friends | n.s. | n.s. | n.s. |
***
p < .001,
**
p < .01,
*
p < .05
Note: Logistic regression was used for antisocial behaviors and violence outcomes. The numbers in parentheses after regression coefficients are odds ratios. OLS regression was used for substance use outcomes and standard errors are provided in the parentheses.