Because the key target of deterrence strategy is the prevention of criminal victimization, we correct the agespecific
suspect rate by a deflator for each year, reducing the suspect rate to represent the homicides committed by
each age category. This adjustment produces the pattern shown in Figure 3. There are two other corrections we do
not make in this analysis. First, we do not adjust the SHR rates upward in each year to compensate for missing agencies
that failed to properly report their data, as Cook and Laub [FN47] and Fox [FN48] have previously done. This
adjustment is not necessary for any analysis that is primarily a comparison of the two age groups because the missing
sites in any one-year period are missing for both groups. Second, we do not adjust our age-specific rates to exclude
the seventeen-year-olds in several states and sixteen-year-olds in North Carolina and Connecticut who had aged out
of the juvenile court. This might mute any differential deterrent impact that drives down the rates of the suddenly
vulnerable *67 juveniles by adding a few percentage points of young adults (at least according to those states' statutes)
into the juvenile sample. If a close to significant difference comes from this undifferentiated mixture of thirteen-
to seventeen-year- olds, we will consider the impact of the inclusion of this age group.