The conflict perspective of criminology and societal reaction to crime suggests that the administration of criminal justice is determined and controlled by those segments of society which are relatively powerful. Based on this perspective, it is reasonable to expect that relatively powerful groups or categories will be far less subject to severe criminal sanctions than will those who are relatively powerless. This proposition may be tested at points in the criminal justice system where decisions are made relative to the application of criminal sanction. The findings are that the relationship between race of offender and sentence length considered both with and without selected control variables is a uniformly weak relationship. In certain categories of control variables the relationship between race of offender and sentence length does strengthen slightly, but in no case are the relationships sufficiently strong to be significant at the .05 level. Partial correlation coefficients show the relationship between race of offender and sentence length to be little affected by the control variables. Therefore, the relationship between race of offender and sentence length is in all cases considered, and by every form of analysis, quite weak. Proportional reduction in error in virtually every case considered in this study is less than 1 percent.