Defendant's race had a direct effect only on the decision to incarcerate, yet it had indirect effects on both incarceration and sentence length. Black defendants were significantly more likely than white defendants to be sentenced to prison. Detroit defendants who were detained before trial or were tried by a jury received harsher sentences, and blacks were more likely than whites to be detained and to be tried by a jury. An interaction was discovered between the defendant's race, the seriousness of the case, and the harshness of the sentence. With only one exception, race had a significant effect on incarceration only in less serious cases. With regard to the decision to incarcerate, the findings provide support for the "liberation hypothesis" of Kalven and Zeisel in which they suggest that judges confronted with defendants convicted of murder, robbery, or rape have relatively little latitude in deciding whether to sentence the defendant to prison. Judicial discretion is diminished similarly if the defendant has prior felony convictions, victimized a stranger, or carried a gun during the crime. 4 tables, 1 appendix, and 46 references (Author abstract modified)