A Statistical Perspective on the Racial Justice Act

Last Wednesday, House Republicans failed to garner enough votes to override Governor Perdue’s veto on the Racial Justice Act(RJA). The 2009 law, which allows North Carolina death row inmates to challenge their death sentences on the grounds of racial bias, was repealed by the General Assembly, but the governor vetoed that repeal.

RJA Supporters have a seemingly compelling argument that the probabilities of getting a death sentence increase three and a half times if the victim is white rather than a person of color. North Carolina Coalition for A Moratorium(NCCM), an enthusiastic supporter of the law, cited an empirical study on the relationship between race and death penalty to endorse the law.

The study found that there are tremendous disparities when it comes to the race of victim to that of defendant based on the data during 1993-1997 period. When non-white defendants murder white victims, the death sentencing rate is 6.4%; however, when white defendants murder non-white victims, the death sentencing rate falls by half, to 2.6%; when non-whites are both the defendant and victim, death sentences dip even more, to only 1.7% of the cases.

Despite the regression analysis and significant results, the study is still not convincing enough as evidence to support the Racial Justice Act. Proponents of the act have misunderstood some basic statistical concepts.

First, just because two variables have positive or negative relationship doesn’t necessarily mean they have a causal link. The fact that the death sentencing rate among non-white defendants is higher than the white defendants doesn’t mean non-white defendants are treated unjustly during the process of trial.  Other factors that are not taken into account might also have different degree influence on the trial. For example, if non-white defendants are generally poorer than white defendants, they might not be able to hire same quality representation as rich whites do.

Second, what statistical results show us is the way things are. Results themselves are rarely able to provide logical conclusions among variables that we are interested in. Establishing a theory by just looking at the results is like building a house on sand. For instance, suppose a study found that 85% of African American students in colleges achieved C on average of GPA, while white students reach B on average. Doesn’t that mean the faculty members in those colleges were racists? It might be the case. But one cannot deny a more straightforward possibility that African American students perform not as good as white students.  Facing these possibilities, what we should do is to conduct further and detailed research rather than jump to a conclusion recklessly.

Simply put, if the possibility that African-Americans are more likely to commit crimes than other races can’t be denied, this piece of legislation shouldn’t have been put in place until more convincing evidence is presented.


Reader Comments