It isn’t enough to identify that some classification errors or Census undercount issues may exist. Instead it is important to try to estimate the size, the direction, and the probable impact of the possible errors—for example, in communities where there has been dramatic growth in the Hispanic or Native American populations (which may not be reflected in Census estimates) and the result is that arrest rates are calculated on a base that is artificially small. The impact will be that the arrest rate for these groups will appear to be very high and the RRI values will be correspondingly high for arrest, and perhaps for referral. Using school enrollment data as a means of ‘correcting’ the census data would be very appropriate in such a situation. However, if the concern is that Hispanic youth may be classified as white by law enforcement or other officials, the impact is likely to be the reverse. It will serve to increase the white rate of activity and decrease the rate of activity measures for Hispanic youth. If we find a significant RRI value in that situation, we can be relatively certain that the classification error means that the RRI value is an underestimate of the actual disparity in juvenile justice contacts.

Differential Behavior

Mobility Effects: Importation/Displacement

Indirect Effects

Differential Opportunities for Prevention and Treatment

Justice by Geography

Legislation, Policies, and Legal Factors With Disproportionate Impact

Accumulated Disadvantage