The basic methodology of former studies of the net effect of criminal justi
ce processing on racial disproportionality in America's prisons has the pot
ential to advance understanding of this issue. But aggregation of data at t
he national level and/or over offenses among other weaknesses limits the ut
ility of the previous findings. The authors employ positive aspects of the
methodology of the studies in analyzing Pennsylvania data for 1991 to 1995.
The total percentages of explained disproportionality are substantially lo
wer than other studies. The same explained disproportionalities and those f
or individual offenses are also far lower than another study from national
data, and low percentages of explained disproportionality for some offenses
suggest that high levels of racial discrimination may be operative. In par
ticular, implementation of the war on drugs seems fraught with injustice to
ward Blacks Explained disproportionality is fairly stable over time, indica
ting reliability for the conclusion that racial discrimination in Pennsylva
nia's criminal justice system seems greater than some earlier studies show.