Race, class, and criminal adjudication: Is the US criminal justice system as biased as is often assumed? A meta-analytic review

It is widely reported that the US criminal justice system is systematically biased in regard to criminal adjudication based on race and class. Specifically, there is concern that Black and Latino defendants as well as poorer defendants receive harsher sentences than Whites or Asians or wealthier defendants. We tested this in a meta-analytic review of 51 studies including 120 effect sizes. Several databases in psychology, criminal justice and medicine were searched for relevant articles. Overall results suggested that neither class nor race biases for criminal adjudications for either violent or property crimes could be reliably detected. For all crimes, effect sizes (in terms of r) for Black vs White comparisons were.054, for Latinos vs Whites, 0.057 and for Asians vs Whites −0.028. There was significant heterogeneity between studies, particularly for Asian vs White comparisons. Effect sizes were smaller than our evidentiary threshold, indicating they are indistinguishable from statistical noise. For drug crimes, evidentiary standards were met, although effect sizes were very small. Better quality studies were less likely to produce results supportive of disparities. Studies with citation bias produced higher effect sizes than did studies without citation bias suggesting that researcher expectancy effects may be driving some outcomes in this field, resulting in an overestimation of true effects. Taken together, these results do not support beliefs that the US criminal justice system is systemically biased at current. Negativity bias and the overinterpretation of statistically significant “noise” from large sample studies appear to have allowed the perception or bias to be maintained among scholars, despite a weak evidentiary base. Suggestions for improvement in this field are offered. Narratives of “systemic racism” as relates to the criminal justice system do not appear to be a constructive framework from which to understand this nuanced issue.