Hood and colleagues studied many variables related to 2,884 sentenced defendants. Results indicated that black defendants had a 5 percent greater probability of being sentenced to prison than did white defendants who were matched on variables predictive of custody. A debate between Hood and Tamar Halevy on the validity of the study's findings appeared in a leading English law journal in 1995. The debate focused on thresholds of statistical significance and on variables excluded from the statistical analyses. Analysis of this debate indicates that the study used an adequate threshold of significance, that the exclusion of certain status variables did not invalidate the study's findings, and that the choice of which sentencing variables to control for is ultimately a normative matter. Consideration of actions, other than eliminating racial discrimination, needed to reduce racial imbalance in the prisons suggests that little is likely to be gained by seeking to make social adversity a basis for mitigating sentences. Instead, it would be appropriate to change policies relating to stages prior to sentencing; one example would be policies relating to drug law enforcement. Notes and 19 references (Author abstract modified)