Inferring and Explaining

128 I, also, fnd it pretty hard to explain the Bal- dus study results as simply a “statistical fuke.” Such things are always possible, but modern statistical analysis guarantees us that they are exceedingly unlikely. Consequently, the following is also very low on my plausibility ranking: InferrIng and exPlaInIng t ″ 2 . It’s just a coincidence that victims’ race “correlated” with capital sentences in Georgia. Te only serious competitor I can imagine, therefore, is that there is some unnoticed “com- mon cause” that is independently responsible for both the race of the homicide victims and the fact that their murderers received the sen- tences they did. Te Baldus team tried to think of some of the possible factors in their original evaluation of the data. Tat’s what they were up to when they performed the statistical tests that “controlled for over two hundred nonracial vari- ables.” Even when they did this, it turned out that murderers of white victims were 4.3 times as likely to receive death sentences. Maybe something else is responsible for the correla- tion, but we have yet to see what it is. Hence I am willing to take the following seriously as a potential rival explanation. t ″ 3 . Some unidentifed nonracial factor is responsible for the correlation of victim race and death sentences. Since we have yet to even think of what this nonracial factor might be, I admit its possibility but rank it signifcantly lower than the causal explanation in t ″ 0 . Some Other Contingent Realities My argument, so far, has depended on two con- tingent realities, and so, even if the Constitu- tion permits capital punishment in the abstract, given the worldwe live in, the death penalty still remains arbitrary and capricious and, in at least some cases, racially prejudicial and is therefore unconstitutional. But there are at least two other contingent realities that make capital punish- ment even more constitutionally problematic. I believe passionately that the Baldus study, and the others surveyed in the Government Accountability Ofce (GAO) report, tells us that racial prejudice plays a huge causal role in who receives the death penalty and who is executed. But I think that there may be other causal fac- tors at work as well. Our nation does not gather data regarding socioeconomic class; we seem to believe that we are a “classless” society. Were such data readily available, I am quite certain it would show an even stronger correlation between poverty and the death penalty than the one we saw in the Baldus study. I am convinced that poor people are treated by the criminal jus- tice system as second-class murder victims, just as we have seen minorities are. But I am also convinced that the death penalty is also a “poor man’s punishment.” 7 Tose with the fnancial resources to hire frst-class criminal lawyers, and make the state’s murder trial very expen- sive, have amuch greater chance of having their charges plea bargained down to a noncapital sentence. Te last contingent reality I want to mention seems to be actually changing some people’s minds as an argument against the death penalty and changing the minds of some public ofcials such as governors. We now have a record of