How do all of these factors fit together?

View the most recent version.

The full models in Table 1 and Table 2 include all of the factors simultaneously in order to assess their relative associations with self-reported youth delinquency. There were some reductions in the observed effects of the school-related, victimization and peer and family variables on property-related and violent delinquency; however, the results showed that most of these variables retained their independent effects after accounting for the other variables.

Variables measuring aspects of peer relationships most significantly increased the odds of reporting delinquency even after controlling for the other factors. In particular, youth who reported being part of a group of friends that considered illegal acts acceptable had odds of reporting property delinquency that were 3.7 times higher, and odds of reporting violent delinquency that were 2.9 times higher, than youth who were not part of such a group. Going out at night also continued to have a large effect on both self-reported property and violent delinquency, net of the other factors (2.6 p ≤ 0.001 and 2.8 p ≤ 0.001, respectively).

Finally, after holding constant school-related, victimization and peer and family variables, there were changes in the relationship between the demographic variables and delinquency. With respect to self-reported property delinquency (Table 1), the effects of sex and age were fully accounted for—rendered statistically non-significant—once all of the additional risk factors were added. Notably, it was the combination of all risk factors in the full model rather than any one particular category of risk factors in previous models that fully accounted for the effects of age and sex.

The compensatory effect of being a recent immigrant became statistically non-significant when we controlled for demographic and peer and family factors, and remained non-significant after controlling for all categories of risk factors in the full model. This suggests that it is the different types of relationships that recent immigrant youth have with friends and family that help account for their decreased odds of reporting property-related delinquency. The effects of living in a single-parent and step/blended family were reduced but remained significant in the full model.

In Table 2, the full model results show that youth from step/blended families were no more likely than youth from intact families to report violent delinquency. This result appears to be largely due to the contribution of school-related factors since a similar result was observed in Model 2. No factors, individually or considered together, however, mitigated the effect of sex on violent delinquency. Even after accounting for all of the other factors, the odds of male youth reporting violent delinquency were more than double those of female youth.