A meta-analysis of race and sentencing research: Explaining the inconsistencies

Numerous studies have addressed the question: Are African-Americans treated more harshly than similarly situated whites? This research employs meta-analysis to synthesize this body of research. One-hundred-sixteen statistically independent contrasts were coded from 71 published and unpublished studies. Coded study and contextual features are used to explain variation in research findings. Analyses indicate that African-Americans generally are sentenced more harshly than whites; the magnitude of this race effect is statistically significant but small and highly variable. Larger estimates of unwarranted disparity are found in contrasts that examine drug offenses, imprisonment or discretionary decisions, do not pool cases from several smaller jurisdictions, utilize imprecise measures, or omit key variables. Yet, even when consideration is confined to those contrasts employing key controls and precise measures of key variables, unwarranted racial disparities persists. Further, a substantial proportion of variability in study results is explained by study factors, particularly methodological factors.