This paper builds on Kalven and Zeisel's “liberation hypothesis” and explores the possibility that racial discrimination in sentencing is confined to less serious cases. We examined the sentences imposed on defendants convicted of violent felonies in Detroit. We found that defendant's race had a direct effect only on the decision to incarcerate but had indirect effects on both incarceration and sentence length. Further analysis revealed an interaction between the race of the defendant, the seriousness of the case, and the harshness of the sentence. Using a number of measures of the seriousness of the case, we found that race had a significant effect on incarceration only in less serious cases. Our findings provide dramatic support for the liberation hypothesis and highlight the importance of using an interactive rather than an additive model in sentencing research.
Stored only when both papers are in the corpus — 197 citations exist upstream.