).
In 2004, Mitchell and Mckenzie conducted a formal meta-analysis of 85 different studies on race and sentencing. They concluded, after taking into account legally relevant variables such as defendant criminal history and offense seriousness, that Black and Hispanic defendants were generally sentenced more harshly than white defendants. While they found that legal factors were the strongest predictors of sentencing outcomes, race continued to have a modest, but statistically significant effect. That is, racial differences persisted, and this persistence is difficult to understand especially once legally relevant sources of variation are addressed. .
Advances in research design and methodology have not put to rest nagging concerns about racial inequality in criminal justice processing. Instead, they have suggested that the connections between race and sentencing outcomes are exceedingly complex, conditioned by a diverse set of variables, including characteristics of the defendant, the victim, the crime itself and the geographical area where it took place, as well as the bureaucratic organization of local courts, the culture of courtroom workgroups, and even broader features of the social and political environments in which case processing takes place.
Numerous reforms to sentencing policy, such as the development of sentencing guidelines and "structured sentencing" models, were introduced between the late 1970s and early 1990s to neutralize the perceived effect of defendant characteristics on sentence outcomes, which were most often attributed to the enormous discretion accorded judges under indeterminate sentencing schemes (Tonry, 1995). In reaction to perceived inequalities in the federal and state sentencing systems, social and legal scholars pressed for legislation that would replace indeterminate sentencing with a system that offered greater predictability in determining proper sentence dispositions and lengths of imprisonment (Frankel, 1973; von Hirsch, 1976; Kennedy 1979).