Race, Racial Context, and Withholding Adjudication in Drug Cases: A Multilevel Examination of Juvenile Justice

Abstract

In the Florida juvenile justice system the decision to withhold adjudication can serve as a less punitive action by the courts and could prove to be more beneficial for the youth receiving this concession. The current study examined how individual-level variables and community context, with a focus on race and racial context, may influence the decision to withhold adjudication for all juvenile drug cases adjudicated between the years 1999 and 2001. Using hierarchical generalized linear modeling, the results indicated that Black youth were less likely to have adjudication withheld compared to their White counterparts. However, contextual county-level variables did not explain variation in the overall probability of a county to withhold adjudication, nor did these variables affect the Black-adjudication withheld relationship. The implications of these results for future research on racial threat at the individual and contextual levels are discussed.

KEYTERMS:

• Juvenile justice processing
• hierarchical linear modeling
• race and ethnicity
• social contexts
• concentrated disadvantage
• racial threat

We would like to thank Michael Baglivio, Marian Borg, Charles Frazier, Chris Gibson, and Marv Krohn for their assistance on this project.

Notes

∗p < .05.

∗∗∗p < .001.

∗p < .05.

∗∗∗p < .001.

1. Although the Florida Department of Juvenile Justice data included a measurement for ethnicity, this variable had over 30% coded as unknown. Therefore, we had to include only the race measure, not the ethnicity measure. We were interested in examining differences within Hispanic youth, but lack of ethnicity data did not allow for this examination.

2. Percent urban was operationalized as the percentage of county residents that are classified as residing in urban areas. Percent urban was also included in preliminary analyses but was excluded from final models. This was due to the fact that it had multicollinearity issues with concentrated disadvantage coupled with no significant findings from these preliminary analyses.

3. Group mean centering was also used and coefficients did not significantly differ.

4. In all of the conditional models, the number of counties was reduced to 62 due to an insufficient amount of Level 1 cases in some counties.

5. Because the percentage of individuals in the entire sample classified as other race was low (each 1%), many counties had an insufficient number of individuals in these categories. This disallowed an analysis where the other race-adjudication slopes were allowed to vary. Although this limits our study in regards to assessing random effects, the fixed effects were still able to be examined.

6. In preliminary tests, county-level variables were entered into this equation separately.

7. In the HGLM analysis by Armstrong and Rodriguez (2005) fairly similar county-level results were found. In their study of preadjudication detention, urbanization, racial inequality, and crime rate were not significant predictors.